Ginkgo Bioworks Holdings, Inc. 10-K

Item 1. Business. Unless the context otherwise requires, all references in this section to the “Company,” “Ginkgo,” “we,” “us,” or “our” refer to the business of Ginkgo Bioworks Holdings, Inc. and our subsidiaries. Overview: Our Mission is to Make Biology Easier to Engineer Our mission is to make biology easier to engineer. That has never changed. Every choice we’ve made with respect to our business model, our platform, our people, and our culture is grounded in whether it will advance our mission. Why? Because: 1. Biology is programmable. All living things run on the same DNA code. 2. Biology matters. The ability to engineer biology has had and will have a profound impact on how we develop new medicines and vaccines, grow our food, and manufacture many of the things we use every day. 3. Biology is hard. Today, it is still too difficult and too costly to engineer biology, preventing critical innovations from reaching the market. Making biology easier to engineer is a systems challenge. No single technology offers a solution. In a way, Ginkgo acts as a systems integrator, bringing together many different technologies–whether built internally, acquired, or through partners–to offer our customers a more integrated and complete solution to a multi-dimensional challenge. We offer these integrated solutions in two domains: cell engineering, where we work to solve biological R&D challenges for our customers across a range of industries, and biosecurity, where we seek to build solutions to identify, respond to, and ultimately prevent, biological threats. An overview of these two business lines is provided below. Cell engineering Our cell engineering customers work with biology to do incredible things, with transformative potential across industries: • in medicine, developing innovative new therapeutics and vaccines; • in agriculture, advancing the sustainability and security of our food systems; and • in industrial biotechnology, advancing the way we manufacture a wide range of products for better performance and lower environmental impact. We are inspired by the incredible diversity of innovations our customers are enabling. Our mission is to make it easier for them to do their critical work. Because engineering biology is incredibly hard, biotech R&D is traditionally performed by in-house labs filled with highly trained scientists running lab experiments by hand over several years in the hope of ultimately developing a working product. Many attempted cell engineering projects fail in development due to scientific challenges, and many are terminated because they are taking too long or are over budget. How many innovative new medicines or climate solutions are stuck for lack of the right tool? Ginkgo does not make products; we build enabling platform services so that our customers can bring products to market. Ginkgo provides flexible, end-to-end biotech R&D services that can offer customers better results on the dimension of cost, speed, or probability of success – and ideally on all three. In exchange for these services, Ginkgo’s customers generally pay us through a blend of fees during the development of their product and downstream value share in the form of milestones, royalties, or equity, which allows us to align our economics with the success of the products enabled by our platform. The fundamental advantage of our platform over traditional cell engineering done by hand at our customers’ labs is that our platform improves with scale while in-house cell engineering in our customers' labs largely does not. Compounding and mutually reinforcing improvements of our laboratory automation and software infrastructure—our Foundry—and our reusable data assets—our Codebase—enable us to improve our services while also enabling the flexibility to apply high throughput automation to the diversity of challenges across biological systems encountered by our customers. • Our Foundry is a highly automated, yet flexible, lab powered by proprietary automation and software to enable flexibility and scale. The Foundry automates lab workflows at high levels of abstraction, enabling users to generate potentially valuable datasets labeling broad genetic sequence design space with a wide range of 1 Table of Contents functional data through modular design-build-test-learn cycles or campaigns. Our scale economic means that the Foundry’s capacity to perform more and more diverse campaigns grows while the cost per campaign decreases. We call this scaling factor Knight’s Law. • Our Codebase is a data asset which accumulates as we operate our Foundry in service of customer projects. Our Codebase includes vast amounts of data at different levels of characterization and usability in engineering projects, including: proprietary libraries of genetic sequence data that can be used for pretraining large language models via unsupervised learning, experimental data for fine tuning task-specific generative AI models, as well as best practices for cell engineering, sequences and host cells that have been honed through dozens of programs and can be directly reusable for different applications of cell engineering. This flexible capability for large scale data generation and broad, growing data asset empowers generative AI and machine learning (“ML”) tools that enable more predictive design of cell programs and higher probability of success for our customers. As the platform scales, we have observed a virtuous cycle between our Foundry, our Codebase, and the value we deliver to customers. We believe this creates a powerful “flywheel,” where new cell programs drive improvements in our platform which in turn drives customers to outsource even more new cell programs. Figure 1: Ginkgo’s flywheel is driven by a scale economic: as we add cell programs, we aim to make programs better, faster, and cheaper – qualities of which we expect our customers will always have demand. We believe that cell programming has the potential to be as ubiquitous in the physical world as computer programming has become in the digital world and that products in the future will be grown rather than made. To enable that vision, we are building a horizontal platform to make biology easier to engineer. Our business model is aligned with this strategy and with the success of our customers, setting us on what we believe is a path towards sustainable innovation for years to come. Biosecurity As with every technological revolution, reaping its benefits to the economy and society also requires us to grapple with its risks. A critical part of making biology easier to engineer is creating biosecurity infrastructure with the goal of managing the many accelerating and diversifying sources of biological risk, whether natural or engineered, accidental or malicious. And many of the same biotechnological capabilities that we’re applying to industry can also be applied to safeguarding lives and livelihoods. 2 Table of Contents In the digital world, we’ve learned that we need to build comprehensive infrastructure to protect our personal computers and digital systems of all kinds—from financial markets to power grids—from harmful code. The modern cybersecurity industry offers tools to constantly monitor for cyberthreats, assess the damage and ongoing risk of cyber incidents in near-real-time, recover system integrity, track the sources of threats, and rapidly deploy pieces of software that reduce vulnerabilities in our systems. This is happening constantly, all around us. A physical world grown with cell programming demands the same type of widespread biosecurity infrastructure to detect, characterize, respond to, attribute, and prevent biological threats. This is a hard problem, and one we’ve been working towards for a long time. But the COVID-19 pandemic showed us—and the world—just how urgent it is to address. Our healthcare infrastructure, biomedical technology industry, and communities across the world mobilized in valiant and unprecedented ways, but still left us with losses of millions of lives and trillions of dollars. Our current systems are reactive: designed to pick up on patterns of illness, then figure out the nature of the threat, then try to quickly mobilize diagnostics, therapeutics, and vaccines. We need a fundamentally different approach to securing biology—one that starts with data. The DNA and RNA code that underlies the biological world is what allows us to program it like computers, but it’s also what allows us to understand it at a molecular level and learn to predict how it’s going to behave in the world. Ginkgo’s Biosecurity platform is built on the premise that genetic code is a high-fidelity data asset that will form the foundation for next-generation biosecurity—making step changes in our ability to rapidly and reliably respond to, prevent, and attribute biological threats. Because biosecurity is a matter of national and global security, our primary biosecurity customers are national governments. We offer them end-to-end tools and services for: • persistent and pervasive collecting of environmental samples at critical nodes—like airports and borders, agricultural settings, cities and communities; • testing and sequencing the samples’ non-human DNA and RNA and analyzing the genetic code for signatures of natural or engineered biological threats; and • putting that information in global context with insights from our network and other data sources to provide critical, early information on what’s coming and what to do about it. Like our cell engineering platform, our biosecurity platform gets better with scale. We are observing a network effect: as we partner with more countries to stand up detection nodes and build underlying local biosurveillance capacity, we have been able to achieve earlier detection and deeper insights on a global scale. We also prioritize reinvestment in our platform with the goal of serving new types of nodes and samples, detecting new targets, integrating additional sources of data, building better predictive capabilities, and eventually plugging directly back into countermeasure development, in near-real-time. We believe that with scale we can substantially strengthen the value of our platform, with global data providing insights far beyond what any one country’s data could yield alone. An introduction to synthetic biology: designing biological sequences Synthetic biology was founded on a deceptively simple question: could we program cells as easily as we program computers? Biology runs on a digital code. It’s just A, T, C, and G rather than 0 and 1. There are sequences that code for programming logic—turning genes on when certain conditions are met—and there are sequences that encode functions and behaviors—the physical structures of proteins and enzymes that create biological structures and materials or catalyze chemical reactions. Synthetic biologists build cell programs by writing new sequences combining regulatory and functional elements into a synthesized strand of DNA and booting them up in cells to perform useful tasks, usually producing a particular bioproduct such as a protein, enzyme, or chemical. Biological code programs the world of atoms, not bits. This is what makes the potential impact of synthetic biology so great, and inspires us to work to make biology easier to engineer and secure. But it also poses incredible challenges that make cell programming so hard today. Our code is a physical object with chemical properties. It folds and binds and interacts in many complex ways. It produces proteins that catalyze chemical reactions that interact in a complex web of connections. Even the simplest cell programs encounter incredible complexity within, emerging from all of the interactions of chemicals, DNA, RNA, and proteins inside of a cell. Consider a relatively straightforward program that makes a single protein product in a microbial cell, such as a biologic drug like insulin or a protein ingredient for alternative dairy: 3 Table of Contents • Different DNA sequences encode the same protein sequence, and the sequence of the DNA itself is adapted to the conditions inside of the cell it evolved in and has to be recoded to work optimally in a new host. • The DNA doesn’t do anything by itself, but must be “turned on” and transcribed into RNA by the cell’s machinery, which is in turn translated into protein by other cellular systems. Different regulatory elements impact the timing, strength, and logic with which genes are turned on and off and the amount of protein produced from the RNA. • The RNA can fold in on itself and produce structures that impact how much of the protein is produced. • The protein likely needs to be secreted or otherwise purified from the cell, and different types of cells and proteins are secreted at different levels. • The protein may need to be altered to improve its solubility, shelf stability, and how it behaves in a final formulation of a drug or food. • The protein may need to be modified in different ways by the cell in order to improve its ultimate function or prevent it from causing an immune reaction in the body. • The cell needs to be convinced to make a lot of this protein, likely shifting its metabolism from producing things more useful to its own survival. • The cell needs to be grown at large scale to manufacture the final product, which requires precise optimization of the conditions in well regulated facilities designed for the safe production of food or medicines. • Moreover, these elements interact with each other, and changing one aspect of the sequence, structure, modifications, gene regulation, host cell background, or production process will likely impact other parameters. Figure 2: The process of producing even a “simple” product of synthetic biology has many steps and contexts that impact technical and commercial feasibility. This is all for simply expressing known sequences, after the function of the protein was well characterized; designing new protein structures and functions adds many more layers of complexity based on how the protein sequence folds and functions inside of the cell. And for more complex cells and more complex programs with more parts and enzymes, the number of interactions and possibilities grows exponentially, making programming larger genetic pathways and behaviors incredibly challenging. Every single product of biotechnology that goes out into the market – a new plant trait, a new enzyme, and new therapeutic protein or new gene therapy, every chemical produced by a microbe via fermentation, required R&D teams to solve each of those layered interdependent challenges to find the right sequence, cell, and the production method that was used to manufacture it. 4 Table of Contents These sequence design and optimization problems are increasingly able to be addressed with generative AI models trained on biological data. AI models can learn to “speak” the language of DNA, RNA, and protein codes, trained on the “grammar” of sequences that have evolved over billions of years. But context and function matter. Large libraries of natural sequence data are critical for pretraining large language models, but it is labeled data mapping sequences to function in biological contexts that is critical for the task-specific models that can predict sequences that will meet particular specifications. For most applications of synthetic biology, there is simply not enough data publicly available to generate such models. Therefore, to develop new products in synthetic biology requires actually building large libraries of possible sequences and testing their function in order to: 1. identify the “hits” with better performance; and 2. iteratively train task specific models that aid in the design of improved libraries to further optimize sequences for the conditions required for commercial success of a synthetic biology product. This type of R&D requires capabilities in large scale design and synthesis of DNA libraries, high throughput screening and collection of functional data for each sequence design, and machine learning to iteratively learn from the data generated by these campaigns. We invest in this infrastructure for data generation and AI modeling so that our customers can access it as a service. Our cell programming services enable discovery, functional optimization, and efficient manufacturing of biotech products Because all organisms run on the same DNA code, general-purpose cell programming can be applied across many different markets to enable the design of new innovative products as well as improve manufacturing cost and sustainability of existing ones. Given the breadth of application areas and the potential of biology, we believe that the end markets for bioengineered products will be enormous. As we develop a greater ability to program biology and direct it towards novel and more challenging applications, the spectrum of possibilities will undoubtedly grow. Traditional biotech R&D requires large up-front investment of time and fixed capital expense in order to build laboratory infrastructure to generate data needed for the particular product application. Because of this large fixed cost, it is slow to start, difficult for companies to explore new product areas, and difficult to scale to meet the need for data required to adequately optimize to meet commercial scale requirements for functionality and manufacturability. Our services give R&D teams flexible access to automated laboratory infrastructure that provides more data per dollar without having to invest in costly capital expenditures, as well as data and AI resources that enable more predictive design for their applications, designed to help our partners better meet the needs of their markets. Today, our services cross markets and modalities to enable a wide range of biotech products, including, but not limited to, the discovery, optimization, and manufacturing systems for: • DNA sequences delivered as vaccines and gene therapies • RNA sequences for vaccines, therapeutics, and novel approaches to crop protection, including mRNA, circular RNA, and other approaches • Proteins used in food and alternative meat and dairy, protein-based materials such as silk and structural proteins such as collagen or keratin, biologic medicines and antibodies, plant traits for crop protection, AAV capsids and other delivery methods for gene therapies and vaccines • Enzymes used in industrial processing, food production and brewing, bioremediation and biomining, chemical manufacturing and biocatalysis, diagnostics, therapeutics, or RNA vaccine production • Small molecules and natural products that can be produced via pathways of multiple enzymes in engineered cells for cosmetics and food ingredients, specialty or commodity chemicals, materials, agricultural inputs, or pharmaceutical ingredients and adjuvants • Microbial cells that can provide crop nutrition or protection in agriculture, impact soil carbon sequestration to help address climate change, offer probiotic nutritional benefits, or microbiome therapeutics • Mammalian cells for manufacturing of biologics, genomic medicines, and cell therapies 5 Table of Contents A common platform across markets enables economies of scale to accrue in unexpected and powerful ways—an enzyme library generated for production of a fragrance might generate negative data that trains an AI model which in turn could empower design of an enzyme used to produce a treatment for a rare disease; models of RNA structure enable improved expression of both food proteins and novel therapeutics. A flexible platform to generate functionally relevant data from libraries of DNA sequences can be applied to an existing product and to new products and modalities developed by innovators from any market. As a horizontal platform for cell programming, we focus on developing platform services and invest in technologies that we believe will have cross-cutting impact across markets. Our customers bring incredible depth and expertise in their unique technical domains and market areas, from the underlying disease biology or plant physiology, to the performance of regulatory trials in animal studies, in the clinic, or in the field, to product formulation and functional testing, and so much more to manufacture, distribute, and market a product. We provide these innovators with services that help them access more genetic design space in order to discover and optimize functionality and develop efficient manufacturing methods for their products. Enabling customer success across markets Partners come to us with business challenges, from the need to innovate in a new product area and quickly explore different potential opportunities, to needing optimization of product functional performance, to improving COGS, sustainability, or supply chain stability. Our teams work hand in hand with our partners to understand the critical parameters for commercial success and to build programs that meet their business objectives. Pharma & Biotech There is urgent, critical need for new therapeutics and vaccines for currently intractable health conditions across the globe. There is also widespread realization within the pharmaceutical industry that research productivity must be enhanced in order to meet this need. With billions of dollars spent annually on R&D in pharma, the cost to bring a new drug to market is only increasing. At the same time, there is great promise in how AI tools may help uncover new disease biology and targets for therapeutics, as well as enable the programming of new medicines, in particular biologics and genomic medicines that are encoded in DNA and RNA sequences. Pharma R&D teams are looking for ways to generate and federate data to train these models, design and test more technical approaches and candidates at the preclinical stage to “fail fast” before costly clinical trials, and develop better leads simultaneously optimizing along multiple dimensions important for therapeutic index as well as manufacturability and cost. The pharmaceutical industry today relies heavily on outsourced R&D, both to specialized, innovative small biotech, as well as to CROs that can automate and scale specific common workflows at different stages of the R&D process for enhanced efficiency. These approaches enable access to both innovation and efficiency, but suffer from high switching costs both organizationally as well as technically. With flexible automation to drive economies of scale for data generation campaigns across modalities as well as a business model centered on innovation partnership, we work to provide our biopharma partners with the integrated platform services that can move the needle on R&D goals from discovery all the way through manufacturing. Our customers are using our platform to develop new manufacturing methods for gene therapies, biologics, and small molecule therapeutics and APIs, and to discover new natural products, RNA therapeutics, and much more. Agriculture Agriculture likewise faces urgent need for innovation to address growing pressure on growers and food systems, and similar to progress in biopharma, agricultural innovation also struggles due to long timelines, complex regulatory paths, and siloed data and capabilities. Innovators in agricultural technology need to tap into biological diversity to develop new crop protection strategies to combat resistance and provide safer, low residue options for growers that meet consumer expectations and regulatory guidelines. They need to understand mode of action and improve the performance and stability of innovative biologicals for crop nutrition and crop protection. And increasingly, they are also innovating in soil carbon sequestration and climate strategies. 6 Table of Contents Our customers in agriculture are using our platform to improve the performance and manufacturability of existing agricultural biologics, develop revolutionary new products for crop nutrition in nitrogen fixation, phosphate solubilization, or carbon sequestration, and design new insect control proteins and other crop protection products to protect food security. Industrial biotechnology Many chemicals that make up the products of our material world are accessible via biology. Living organisms can be engineered to manufacture specialty and commodity chemicals or break down pollutants and create closed loop systems. There is an enormous breadth of products—chemicals, enzymes, and proteins—already produced via biotechnology today or being actively developed by companies across markets, across food and nutritional ingredients, wellness, cosmetics, and personal care, industrial processes and chemicals, and materials innovation. Our customers leverage our services to improve the manufacturing efficiency and COGS of their existing biotechnological products, develop new production processes to replace existing extraction methods for ingredients derived from plant or animal sources to enable more sustainable and stable supply chains, replace petrochemical inputs, innovate materials with enhanced performance, develop enzymes for breaking down harmful pollutants or cells and proteins optimized for capturing rare earth elements, or valorize waste streams into feedstocks for more valuable products. An introduction to biosecurity: scaling biological intelligence for securing lives and livelihoods Addressing biosecurity starts with being clear-eyed about biological risks and threats. We hold at our core the tremendous positive potential of biology, and we know that we’re facing a biological landscape with more frequent, more severe, and more varied threats through time. Our world is increasingly interconnected through travel and trade, giving pathogens and biological agents new opportunities to spread across the globe, impacting people’s health along with the complex global supply chains that our societies depend on to function. Climate change and habitat disruption are creating the conditions for pathogens to emerge and spill over between animal populations and into humans more often and with more severe consequences. A global boom in investments into bio-laboratory capacity, designed to improve our tools to combat such pathogens, also comes with heightened risk of lab accidents—in spite of substantial efforts to improve biosafety. And unfortunately, there are those who seek to use biology for nefarious purposes, misusing its incredible potential to cause harm. These trends are intertwined with geopolitical competition and destabilization, eroding buy-in and trust in institutions, and emerging technologies in both biotechnology and AI/ML, presenting a core security challenge for nations and the world. Ginkgo Biosecurity is designed to help national and global leaders answer the dizzying questions about biological threats that keep them up at night: • What threats and outbreaks are on the horizon? • What is this new threat and how bad is it? How might it spread and evolve? Who (or what infrastructure) will be affected? • Where did it emerge and how? Is there evidence of misuse and if so, what can we learn about the perpetrators? • What can I do about it? How effective will existing countermeasures be? Should we develop new countermeasures, and if so, what should they look like? What are my ideal response options given resource constraints and mitigation goals? • How will technology change the landscape? What and how severe are the risks of biological research activities globally (including the development of AI tools like Large Language Models and Biological Design Tools)? How can we ensure the evolving technology landscape maximizes the benefits of innovations while preventing harms? These are technically difficult questions—and we’ve too often settled for flimsy answers or none at all. Having the tools to actually make these questions tractable is what will empower leaders to mount substantially earlier and more effective responses, and develop the infrastructure to prevent future threats from taking hold. While there is very important work happening in the world to coordinate global policy and strengthen existing public health infrastructure against known threats, our focus is on building the cutting-edge technology stack that countries will need to face the evolving biothreat landscape and ultimately determine how best to answer these questions. National governments are our primary customer as they seek to protect their citizens, economies, and critical infrastructures, but these technologies also have a wide variety of other applications. For instance, they can help companies predict and mitigate disruptions to their global workforces and supply chains, help health systems prepare for shocks, and help local leaders manage community health. 7 Table of Contents Figure 3: The Ginkgo Biosecurity technology stack leverages data and AI to generate actionable intelligence on biological threats. The Ginkgo Biosecurity technology stack starts with persistent, pervasive, locally-operated collection of environmental samples from strategic high-risk nodes—designed to help establish baselines across a growing array of collection and sample types. The samples are analyzed through genomic sequencing of non-human DNA to turn the environment into data. Today, we are looking for a large and growing set of known threats, and we plan to use methods that are threat-agnostic and able to pick up on totally novel genetic signatures. This is what we call bioradar. The next step is to convert that bioradar data into actionable insights that our customers can use to make more effective and timely decisions—we call this biological intelligence, or BIOINT. The bioradar data from given programs or jurisdictions are integrated with other global data sources—from our monitoring network, open-source intelligence capability, and other sources—and analyzed using a suite of AI/ML-enabled tools to help customers gain a more comprehensive picture of the threats they’re facing. Our evolving data feeds and analytics tools are bringing us closer to being able to answer the questions above, and supporting key activities like threat attribution, as well as response planning and implementation. Ginkgo’s leadership in AI/ML development, in particular, will be critical to enabling BIOINT. AI/ML already helps us do everything from detecting signatures of engineered biology, to rapidly identifying anomalies in environmental samples and digital surveillance that could signal novel threats, to forecasting how a pathogen will spread. As we advance the state of the art in AI/ML, we plan to find new ways to enhance these capabilities and build new applications for BIOINT. For instance, BIOINT can feed back into the development and optimization of vaccines, therapeutics, diagnostics, and other response tools, using data and AI/ML to help us understand how effective different countermeasures will be against novel threats or disease variants. Our commitment to caring about how our platform is developed and used Biotechnologies already touch nearly every part of society, and they will only grow in importance to our collective security and livelihoods in the future. Because of their far-reaching impacts on the world and because we are biological beings who are both dependent on and vulnerable to the capabilities we enable, we must take great care in the ways these technologies are developed and used. We are cognizant that making biology easier to engineer won’t make the world better by default, but we believe these capabilities are essential to creating a better future where we can contend with both existing and emerging threats. To succeed in our long-term mission we must avoid multiple failure modes. We must avoid creating capabilities that cause harm in ways that aren’t or can’t be mitigated. We must avoid reinforcing inequities in the uses of technologies and to 8 Table of Contents whom their benefits accrue—thereby claiming to change the world but changing not much at all. We must avoid a loss in trust in biotechnologies and the motivations of their developers that limits our ability to bring solutions to global challenges from protecting against pandemics to feeding the planet. And so we must chase, everyday, the development of capabilities and partnerships that can lead to value generation undergirded by sustained attention to the values they reflect. As our platform grows, so too does our power to enable and shape many impacts we care about. While we are proud of our direct impacts on making biology easier to engineer, most of the world cares about the impacts on the world that we indirectly enable through helping our customers with the products and services they deliver. We grow as a platform precisely because we help create more value for our customers and the world than we capture. Our position serving customers across many industries provides strategic insights into what issues need collective attention to ensure future products can deliver meaningful solutions. But as a platform we cannot anticipate and control all future uses of our technologies by our customers and those they work with. Far from abdicating responsibility, as a platform company we realize our power is to inspire and help enable others to carefully steward technologies with attention to their impacts over time. This is directly in line with our long-term value proposition, as we need our customers and the ecosystem to succeed in avoiding the failure models outlined above and build the collective biotechnology-enabled future we all can wish for. We believe that stewardship starts with our platform and the people within it. Just as we must build and inspire trust in our partners to steward our technologies with care, we build and inspire trust in all of our bioworker-owners to build our platform with care. To begin, we are investing in systems to help everyone in our company to better understand and shape its impacts. Last year in our environmental, social, and governance (“ESG”) report, Caring at Ginkgo, we described updates to our overarching philosophy as well as an internal governance experiment in ownership through a committee of elected bioworkers entrusted with helping to facilitate our caring commitments. The committee reviewed nearly every project last year in order to develop refined posture and processes. As our platform grows we are now shifting attention from case-by-case reviews to strategies that can scale and address issues and opportunities in a more upstream and integrated way. This includes better understanding where the platform has the biggest impacts for our customers along key dimensions such as sustainability. We are also assessing areas of impact and gaps in governance around impacts that we wish to avoid, and developing new ways to practically extend positive impacts and mitigate potential harms that leverage our position and capabilities. We must also pay special attention to the governance of leading capabilities for we have outsized ability to shape. At a macroscopic level building biosecurity capabilities is an example of where we assessed the need for complementary efforts that could safeguard future biotechnologies–including those developed on our platform. But this same philosophy applies across our platform including as we work to harness powerful new capabilities in AI. We believe that our platform design is a foundation for architecting security and access that can both enable positive uses while better understanding and protecting against scenarios of misuse. We see caring about how our platform is developed and used not as a net cost but as an enabler at multiple scales of impact aligned with long-term value. It builds trust and credibility not only in our capabilities but those of our customers. It motivates and enables our employee-owners to drive the platform towards the many diverse uses they co-envision with our platform (our social and motivational flywheel!). It also advances a framework to go beyond a reactive historical frame for ESG that has often positioned genetic engineering as a risk to the environment rather than a value. We recognize that platforms across other industries have lessons—many negative—on how to steward their development and use. Our high-level commitment to care also comes with the expectation of needing to regularly revisit the approaches to realizing that commitment. Our platform Ginkgo’s platform brings together the technology, data, biological assets, subject matter experts, flexible business terms, and a broader ecosystem of resources that enable our partners’ commercial success: • best in class, proprietary automation technologies that enable flexibility and scale • in house software, machine learning, and generative AI models for cell programming • massive databases of DNA sequences and labeled data on functional performance of engineered cells • reusable assets that enable faster and more predictable cell programming • flexible partnership models to meet R&D teams wherever they are in the development process 9 Table of Contents • expert scientists that leverage platform tools and data to enable partners to achieve their desired results • an ecosystem of partners with capital, assets, expertise, and capabilities beyond cell programming that enable the launch and success of synthetic biology products Our Foundry brings economies of scale to cell programming Cell programming projects involve a conceptually similar process regardless of the specific product or market. Based on customer specifications, Ginkgo’s program team develops designs of proteins, pathways and gene networks that might meet the specification, leveraging public and proprietary biological knowledge bases (see “—Our Codebase—organizing the world’s biological code”). Those conceptual designs are refined and specified into particular DNA sequences using computer-aided design tools. Those DNA sequences are then chemically synthesized and inserted into a cell to execute the new DNA code. These prototype cells are then studied and the output or performance of each is measured and compared to the customer’s desired specification. Learnings using data analytics and data science tools can inform a new round of prototypes, if needed. We refer to this engineering cycle from design to learning as a campaign and we perform campaigns both in parallel and serially until either the specification has been met or the customer decides to end the program. The likelihood of technical success increases with each iterative campaign and with the number of prototypes that are explored per campaign. However, with traditional tools for genetic engineering, each campaign can be slow, expensive and error prone. Many projects across the industry run out of budget or time. Conventional R&D teams often look to stay within budget by running rapid campaigns using largely manual tools and small numbers of prototypes per campaign. However, the inability to broadly explore the potential design space (there are more possible sequences of a 200 amino acid protein encoded in 600 DNA letters than there are stars in the observable universe) and the reliance on manual tools is a difficult handicap to overcome. Since people can only work so hard and since campaigns can’t be shortened beyond the duration of the physical steps, this approach has limited potential to improve in the future. At Ginkgo, we invest in improving the tools and technology for programming cells in order to maximize program success within the constraints of our partners’ R&D timelines and budgets. We do so by scaling the number of prototypes that can be evaluated in each campaign in an effort to reduce the number of campaigns required to meet the customer’s specification and ultimately shorten project timelines. A typical campaign for one enzyme step in a program might evaluate 1,000 to 2,000 prototypes to optimize function, of which the top 10 to 100 might be short-listed for further study. A relatively basic program for the production of a small molecule might have three to five enzymes working in concert, and so in the process of optimizing the entire pathway, thousands or tens of thousands of enzymes and pathway combinations might be designed, built, and tested in the Foundry. The methods we use to increase scale also tend to reduce the average cost per prototype, which means that more prototypes can be evaluated for a given program budget. Because diverse cell programs share similarities in process and code, many programs can be run simultaneously in a carefully designed centralized facility. This facility, where we use our investments in advanced cell programming technologies to manage diverse programs, is what we call our Foundry. We make it possible to centralize many cell programming projects in our Foundry by deconstructing programs into a set of common steps and then standardizing those steps. For each step, we have built a specialized functional team that performs that step for all programs. Those teams define a set of standardized services that can be used in concert to execute an end-to-end cell programming process. Each team has access to scientific, software, and robotic engineering resources to replace manual ad hoc operations with standardized, automated, and optimized services. In addition to enabling scale, this approach ensures standard operating procedures, know-how, and human skill become encoded in software that can be more effectively debugged, monitored, controlled, and optimized. 10 Table of Contents Figure 4: A non-exhaustive summary of the functions performed throughout the lifecycle of a program in the Foundry. At each stage, learnings are generated, driving improved designs and functional optimizations. While the engineering strategies described above have historically been relatively uncommon in the life sciences, they are obviously not our invention. Rather, we are inspired by the lessons from other engineering disciplines and seek to apply those to biology. Automotive manufacturing, semiconductor fabrication, and data centers, among many other industries, demonstrate how automation, data, economies of scale, and continuous improvement can produce compounding gains in scale, costs, and quality. Critically, routine performance of these strategies across dozens of projects gives us the data and experience needed to drive continuous improvement. As described above, a key strategy in our Foundry is to increase the scale of our operations so that we can run larger campaigns, a greater number of campaigns, and hence run more programs. This approach benefits from operational efficiencies and economies of scale across many dimensions: • Fixed Cost Amortization: Our Foundry is an inherently physical facility and as we scale and improve utilization, we are able to amortize this fixed cost across more work. • Continuous Learning and Improvement: The cumulative amount of work done as we scale leads to a better understanding about how to program cells. Much of this is then encoded in our Codebase, described below. • Purchasing Economies: By partnering with Ginkgo, our technology partners and suppliers can generate more value from a single account than they could from multiple smaller accounts, and that extra value is shared with Ginkgo. • Technology Specialization: Certain technologies that we leverage in the Foundry (such as acoustic liquid handling, automated bioreactors, and advanced mass spectrometry systems) are not easily leveraged or practical for smaller organizations. But for an engineering organization of our size, those investments can drive material improvements in cost efficiency. These efficiencies and economies of scale can be observed empirically from a relationship we refer to as “Knight’s Law,” named after Tom Knight, one of our co-founders, and loosely inspired by Moore’s Law for semiconductors. We have seen a significant increase in the output of the Foundry over time alongside a significant decline in the average cost per unit of output except during temporary lab shutdowns during the COVID-19 pandemic and reduced capacity due to social distancing. Historically, we have measured the output of the Foundry in terms of the number of strain tests. In 2023, we switched to a higher level metric of Foundry output called a campaign, which is a design-build-test-learn cycle for a given biological engineering objective. We believe campaigns are a better measure than strain tests because campaigns are a more representative and integrated “unit” of biological engineering. In 2023, we continued to see significant improvements in Foundry output and cost per unit of diverse campaigns run by the Foundry. 11 Table of Contents Knight’s Law does not provide the full story on our development, but it is a useful tool that allows us to continue to build efficiencies of scale. We believe we can continue to drive significant capacity growth in the foreseeable future, though it is dependent on the development of new technologies, which inherently carries risk, and, like Moore’s Law, we will likely hit a limit over time. This feature compares to a conventional facility, where scaling is driven predominantly by the addition of employees, an exponential increase in work would be infeasible and the cost per unit of work would decline little, if at all. We are frequently asked, and spend much time thinking about, whether it will be possible for compounding gains in output and productivity to continue for many years in the future. It is important to note that given significantly advanced tools, most steps in cell programming could be miniaturized to a point where single molecules of DNA and single cells are being manipulated and monitored. At that ultimate degree of miniaturization, the costs and timelines of cell programming could be reduced orders of magnitude from where they are today. Microfluidic and encapsulation technologies point to the reality of this future of cell programming at the single-cell level. Additionally, because many of the enabling tools of cell programming are biological in nature (e.g., polymerases and CRISPR), we are able to point the platform at itself, developing new biological tools to reduce the number of steps or the complexity of a certain operation. For example, we could develop better gene editing enzymes or novel ways to screen cells in a multiplexed format using biological sensors. It is easy to theorize about these types of developments, however they are hard to execute, we will undoubtedly run into roadblocks along the way and we will have to invest significantly in developing new technologies in order to enable the types of improvements we seek to achieve. Recent advances in machine learning, molecular simulation, and other computational techniques also hold great promise to improve our ability to program cells. We believe our Foundry is well-positioned to build the kind of large, well-structured datasets that such computational approaches need to succeed. In time, we believe computational approaches will reduce the need for certain kinds of experiments (for example, we already use machine learning to make protein and enzyme design projects more efficient). If computational approaches can replace certain sets of experiments, we expect to use the recovered Foundry capacity to work on ever-more complex cell programming challenges. The reality is that the cells that we program today accomplish relatively simple functions, such as: “produce as much of molecule X as possible.” Programming cells for complex functions, such as live-cell therapeutics, responsive building materials, multicellular organisms, etc., will require sophisticated sub-systems for environmental sensing, intracellular information processing and feedback, and a multidimensional program that responds to such environmental stimuli. Only when we can deliver such sophisticated programmed cells will we have truly unlocked the potential of biology, and we see the Foundry as being an integral part of the platform for doing so. Our Codebase—organizing the world’s biological code Codebase is a familiar term to software developers but is a new concept in biology. Modern software firms develop their own (typically proprietary) codebase of source code and code libraries that can be leveraged by their software developers to more easily create new applications than they could starting from scratch. Additionally, vast repositories of debugged code are shared publicly so that programmers across application areas can leverage prior art in order to innovate faster. This allows software developers to focus their time and effort on developing new features rather than recreating existing logic. Engineering biology is complex (see “Introduction to synthetic biology”)—one of the reasons that Foundry scale is important is that it remains highly difficult to predict the performance of a biological “part” in a given context from a DNA sequence alone. The genomics revolution has outpaced biologists’ ability to test the functionality of each DNA sequence as it was discovered, particularly because most of the community is still performing biological experiments by hand without the benefit of automation. Each program performed at Ginkgo involves testing thousands to millions of DNA sequences; with a small fraction of those ending up in our final engineered cells. This reflects the advancement of Ginkgo’s Codebase and the three levels of biological assets and data that it indexes and organizes: 1. Sequence data from public databases and proprietary gene sequences from a range of unique sources. Our proprietary data asset of metagenomic gene sequences contains over 2.7 billion unique protein sequences that we leverage in the training of large language models and the design of sequence libraries for many campaigns. 2. Vast datasets mapping genotype to phenotype from every strain test within a campaign (approximately 513 campaign starts in 2023). These datasets, including all the sequences of “losing” designs and “negative” data are incredibly valuable for fine tuning task-specific AI models for generative design of RNA, DNA, or protein to particular functional performance specifications. 3. High performance, reusable sequences that can be thought of as a “parts catalog” that can be drawn from in the design of a new program. For example, we developed novel synthetic promoters (DNA sequences that can turn on 12 Table of Contents the expression of a gene of interest) that allowed us to increase production of proteins in yeast. Initially, we tested tens of thousands of designs to arrive at a select number of promoters with high performance. Now those high-performing promoters can be reused in any program that involves producing a protein in yeast; they are a modular piece of genetic code. Ginkgo’s Codebase allows our customers to draw from a broader set of biological assets than any single company would develop for a given application, as well as diverse host organisms that are optimized for the production of a range of different bioproducts. The scale and diversity of our programs have allowed us to develop a large Codebase that grows with the addition of each new program and can be opened to the broad swath of partners and cell programmers using our platform. The combination of well characterized, reusable parts, host organisms, foundry workflows, and design tools for the production of classes of products is referred to as a Cell Development Kit (CDK), inspired by the Software Development Kits (SDKs) used in the software industry to enable rapid development of complex applications. Our Foundry and Codebase are inextricably linked. Our Foundry scale allows us to generate unparalleled Codebase assets. These Codebase assets help us improve our designs and provide reusable parts and chassis strains that improve the efficiency and probability of success of our cell programming efforts in the Foundry. As the capabilities of the platform improve, it drives further demand, which increases the rate of learning in our Codebase. The continuous learning and improvements inherent in this relationship is one of the key features of our platform. We continue to invest in both the efficiency and reusability of our platform services and assets in areas where we have significant codebase and expertise, as well as expand into new domains and modalities. As we execute more programs, generate more data, and validate more CDK parts and hosts, we aim to get better and deliver value to our customers in less time, for less budget, and with lower technical risk. Figure 5: Our Codebase incorporates both biological assets from nature as well as engineered assets and data from our Foundry experiments. Because the Foundry enables us to test many thousands of prototype enzymes, pathways, and strains in individual engineering cycles, we are able to quickly expand the range of characterized biological assets in our Codebase. An ecosystem to support cell programmers Ginkgo has long recognized that it is critical to build a true ecosystem around our technical platform. We have been inspired by the leading horizontal platforms in information technology, such as Microsoft Windows and Amazon Web Services (“AWS”), which built real developer communities and provided a range of value-added services on top of their core technology. Like these pathbreakers, who set the stage for a generation of computer developers, we too are trying to 13 Table of Contents ensure that the cell programmers who build applications on our platform have the tools they need to succeed beyond the lab. Figure 6: Ginkgo strives to create an ecosystem to ensure that cell programmers have the tools they need to succeed Access to capital As in the early days of computer programming, it is still extremely expensive to program biology. For that reason, it can be easier for larger companies to make investments in innovation around this space. But Ginkgo’s platform gives small companies and innovators access to the same horsepower as larger players and obviates the need to invest in fixed laboratory assets, providing an even greater strategic benefit. To help address this discrepancy, Ginkgo has assisted in launching new companies (such as Motif and Arcaea) by bringing together strategic and financial investors to secure funding for these early stage companies. While we maintain a conservative approach to cash management, we are able to leverage our capacity and partner with investors to enable companies at all stages to benefit from our platform. We believe that, as Ginkgo’s customers demonstrate increasing success, there will be an explosion of capital for cell programming applications and a recognition of Ginkgo’s platform as setting the industry standard and providing the backbone for these development efforts. In a challenging capital markets environment, access to capital becomes an even bigger challenge for emerging companies. While we remain thoughtful around ensuring a healthy mix of large and small customers, our value proposition to emerging companies has continued to expand significantly. Manufacturing support Our job is to ensure that our cell programs can be executed at scale and we support our customers to ensure successful commercial scale manufacturing. We have built relationships with a number of leading contract manufacturing organizations and have demonstrated that we can transfer our lab-developed protocols to commercial scale (e.g., 50,000+ L fermentation tanks) with predictable performance. We have an in-house deployment team dedicated to supporting our customers’ scale-up and downstream processing needs. We have even helped certain customers, such as Cronos acquire and build out their own in-house manufacturing capabilities and certain programs, such as our work with Moderna, focus on manufacturing process optimization. In 2022, we acquired Bayer’s West Sacramento agricultural biologicals R&D facility, which included robust pilot manufacturing infrastructure for microbial strains, with room to grow. We plan to continue to invest in this capability, helping bridge the gap for our customers between R&D and commercial production. 14 Table of Contents Intellectual property protection and regulatory support Ginkgo takes responsibility for the cell engineering intellectual property generated through customer collaborations. Our scientific team collaborates with our customers and with Ginkgo’s intellectual property team to file patent applications and monitor collaboration deliverables for freedom to operate. We are also active in the evolving regulatory landscape for biological engineering. While our customers are responsible for handling their own regulatory procedures on a product-by-product basis, our broader view can help build understanding of and support for novel product classes. Building a community of cell programmers We launched Ferment, our annual conference, in 2018. The conference highlights developments and thought leadership in the field and brings together scientists, entrepreneurs, investors, and suppliers, and we look forward to hosting our next Ferment in April 2024. Even prior to launching Ginkgo, our founders focused on building community within the emerging field of cell programming. Tom Knight, one of our founders, was among the professors who launched the iGEM Competition in 2004, which has now had over 70,000 participants from over 40 countries take part in the competition (including dozens of Ginkgo employees and all five founders!). Facilitating partnerships within our community Because Ginkgo serves both large market incumbents and smaller startups, our community also serves to facilitate introductions between innovators and those looking to invest in innovation. We believe that investors and large strategic companies have come to recognize Ginkgo’s platform as a key enabler of innovation and are keen to get to know the companies that are building with us. Those relationships can be the source of funding and go-to-market support for the earlier stage companies building on the platform, increasing the odds that they develop successful products. We invest in building trust and credibility for the entire industry The most powerful technologies require the most care. Biology is too powerful for us to not care about how our platform is used. We have and will continue to invest heavily to build and maintain trust in bioengineering as a technology platform across all layers of the industry. At the platform layer, we have focused on building robust biosecurity measures. At the application layer, we are proud to enable a diverse set of programs that drive towards environmental sustainability. We are committed to ESG practices and broad stakeholder engagement at a corporate level. We are also engaged in deep conversations around the implications and ethics of biotechnologies through many forums, helping shape our platform and our ecosystem to promote sustainability in our global community. Our Biosecurity platform—building the end-to-end infrastructure for countering biological threats Our biosecurity platform is designed to provide end-to-end support for gathering, organizing, analyzing, and sharing data and insights on biological threats to inform decision-making. We combine: • the technologies and logistics required for collection of physical samples, • continually advancing biodetection capabilities, leveraging the expertise of our Foundry and external partners, • state-of-the-art tools and infrastructure for data and analytics, increasingly enabled by AI, • a team of scientific experts who work directly with customers to shed light on trends in biological risk, and • a long-term partnership mindset focused on meeting our partners’ sustained biosecurity needs. We lead with care across our platform. Our bioradar approach places privacy at the center, using anonymized and aggregated samples to glean insights on pathogens without requiring the collection of any personally identifiable information. As developers of AI/ML tools and biosecurity experts, we are actively helping policymakers to understand potential risks at the nexus of AI and biotechnology and develop guardrails and tools to manage them. We're building a global bioradar network We believe that every person—and therefore every country—on the planet should have access to biosecurity tools. Biology doesn’t respect borders, so we can’t achieve local biosecurity goals without plugging in gaps in global biosecurity, and vice versa. We are working to establish long-term partnerships with national governments, taking a comprehensive approach to empowering their biosecurity goals. We partner with local operators on the ground to equip and train them to conduct sample collection, in order to be minimally disruptive to ongoing operations in airports, municipalities, and beyond. We 15 Table of Contents work with in-country laboratories and public health institutions—for instance, the Rwanda Biomedical Centre—to build local capacity for sample testing and sequencing through training, supplies and equipment, quality assurance, and ongoing technical support. This year, we joined forces with Illumina, Inc., a global leader in sequencing and array-based technologies, to further expand biosecurity capabilities, particularly next-generation sequencing tools, across the international market. Taken together, these aspects of our model allow us to rapidly scale to new bioradar nodes in our partner countries, where we establish persistent monitoring programs for biological threats. Many of these nodes are international airports, which multiply the reach of our network by tracking incoming pathogens from over 100 countries of origin. These programs add immediate local value by providing early warning for pathogens entering or emerging within a country, while also contributing to a unified data asset that gives our partners access to broader and deeper global insights than what any country alone is collecting. Figure 7: Our global network now includes 14 countries and several multilateral organizations who are actively building biosecurity infrastructure on our platform, with 10 international airports running active pathogen monitoring programs. As we’ve stood up distributed nodes, Ginkgo’s headquarters in Boston has served as a hub for analyzing data from our global network throughout the development of our Biosecurity business. In the future, our network will increasingly be bolstered by regional hubs, known as Centers for Unified Biosecurity Excellence (CUBEs). We recently announced the establishment of CUBE-D, a first-of-its-kind international biosecurity analytics center in Doha, Qatar. When completed, this facility will support analysis of data from bioradar nodes across the region to generate valuable insights. We push the frontiers of biodetection to new nodes, modalities, and targets We are developing our tools and techniques to collect new types of samples from new varieties of nodes, partnering with public, private, and academic institutions for research and development. Early in our development, we were focused on collecting nasal swab and saliva samples in schools, senior living communities, correctional facilities, and other congregate settings. As we have established airport-based monitoring in partnership with CDC, we have innovated new methods to collect wastewater from aircraft, lavatory trucks, and the airports themselves. We plan to continue evaluating new sampling approaches for addition to our platform, such as air monitoring. We are also expanding to differentiated nodes. While airports and aircraft represent the largest share of our network, we’ve also expanded our model to diverse settings, from municipal conflict areas in Ukraine, to a dairy farm in Australia, to the deer populations of Texas. In the future, these development efforts can help us to strategically place nodes in settings where 16 Table of Contents we think pathogens are most likely to emerge or spread, with the aim of enhancing our capabilities to provide early warning. Figure 8: Our collection technologies are expanding, and we aim to eventually be able to sample from a wide variety of nodes where pathogens emerge or spread, such as those pictured above. (Note: nodes are illustrative and not necessarily indicative of past programs.) In the past year, we’ve rapidly expanded the capabilities across our network to mature from the SARS-CoV-2 monitoring programs of the emergency pandemic response to over 30 new viruses, bacteria, and antimicrobial resistance targets. Where today our sequencing platform identifies and characterizes pre-determined sets of targets that are amplified or enriched within samples, we look forward to incorporating customized detection tools and metagenomic (or fully agnostic) sequencing technologies that can afford us greater flexibility to detect a wider variety of known and as-yet unknown threats. Generating a new form of intelligence The goal of collecting genetic sequencing data on pathogens and biological threats is ultimately to generate actionable biological intelligence that can help decision makers plan timely, effective, and resource-efficient responses, guide the development and deployment of novel countermeasures (such as vaccines and therapeutics), and attribute the origins of a given threat. We are building towards this goal with state-of-the-art data integration, analysis, and reporting capabilities that are increasingly enabled by AI/ML and develop symbiotically with Ginkgo’s Foundry and Codebase. 17 Table of Contents Figure 9: Ginkgo’s biosecurity and cell engineering platforms accelerate each other’s development through feedback loops of data, analytics, and intelligence on Ginkgo’s AI/ML platform Our bioinformaticians ingest bioradar data and assemble biothreat genomes to rapidly detect anomalies (threats emerging or surging in unexpected ways) and identify new variants of pathogens that could have a negative impact. These pathogen genomic insights have provided early warning, such as in the case of SARS-CoV-2 Omicron sub-variants BA.2 and BA.3, which were identified through our traveler-based program with CDC 7 and 43 days, respectively, before clinical detection in the U.S., and fill in gaps in genomic surveillance of pathogens with little to no data available globally. We plan to continue developing our capabilities in genomic epidemiology to better understand pathogens’ evolution through time and space and help predict the emergence of new variants. Pathogen genetic data also forms the basis for our novel engineering detection system, known as ENDAR (Engineered Nucleotide Detection and Ranking). ENDAR is a computational platform that uses a series of algorithms, reference databases of genomic data, and expert analysis to identify and characterize genetic engineering in a sample of interest. The platform is designed to be compatible with a variety of real-world applications and samples—ranging from clinical specimens to complex, multispecies samples such as those collected from wastewater or the environment. By systematically identifying known or novel strains and various signatures of engineering in a given sample, ENDAR can help us determine not only whether a sample was engineered, but where in the genome, how, and potentially why it was engineered. We supplement the pathogen genetic data from bioradar programs with a collection of open-source intelligence through digital surveillance. We track hundreds of different open-source epidemiological data feeds—ranging from official public health reports to unofficial sources like media and social media—on an ongoing basis. We use natural language processing and machine learning techniques, along with careful expert review, to aggregate, structure, and validate these insights into a unified data feed that identifies and tracks the progress of infectious disease outbreaks all around the world for early warning and situational awareness. In addition to this near-real-time monitoring, we have curated a database with information on thousands of outbreaks from the past 60 years. We use all of this information to model how diseases spread and estimate the risk that they pose. For instance, we: • advance global experts’ understanding of the likelihood and distribution of future epidemics and pandemics to inform policy conversations, • help regional leaders understand the risk profile of their jurisdiction to inform preparedness investments, and • iteratively design our bioradar sampling strategies for more efficient capture of high-risk targets. We’re currently working with a consortium of partners funded by the CDC Center for Forecasting and Analytics, known as EPISTORM, to advance capabilities for forecasting and predictive analytics, such as finding new ways to tell when a disease is about to spike and what measures should be taken against it. Taken together, these efforts are bringing us closer to generating a novel form of intelligence on biological threats. 18 Table of Contents Our Business Model The key input into our unit economics is a cell program. For each of these cell programs, we generate economic value in two primary ways. First, we charge service fees for Foundry services, in much the same way that cloud computing companies charge usage fees for utilization of computing capacity or CROs charge for services. Additionally, we negotiate a value share with our customers (typically in the form of royalties, milestones, and/or equity interests) in order to align our economics with the success of the programs enabled by our platform. As we add new programs, our portfolio of programs with this “downstream” value potential grows. Because we typically do not incur material downstream costs (e.g., manufacturing or product development, which our customers manage), these value share payments flow through with approximately 100% contribution margin. This flexible business model allows for more predictable near-term revenue in up-front research fees and technical milestones without sacrificing our ability to create long-term value with asymmetric upside through downstream value share (typically in the form of a royalty stream, milestone, and/or equity share). Foundry (or Cell Engineering) Revenue Illustrative Program Economics Figure 10: Ginkgo generates economics from programs in multiple ways that help calculate a program’s NPV. First, customers generally pay upfront fees to cover initial R&D costs for a program. Ginkgo also receives revenue from the technical milestones that the program achieves throughout the R&D process. Ginkgo also shares in the downstream value of a given program, typically in the form of commercial milestones and royalties generated by a given program. Cell Engineering Service Fees The first stage of a cell program consists of R&D work being performed on Ginkgo’s platform, leveraging our Foundry and Codebase. R&D is inherently risky and our customers recognize that this is a cost they will incur regardless of success and whether they are working on the program in-house or with a partner. Ginkgo can provide a much more efficient platform to conduct this R&D work, encouraging companies to build on or adopt our platform. We estimate that the unit costs of our Foundry cell engineering services are several times less expensive on average than the status quo (a customer doing equivalent R&D in-house, by-hand) and we expect that cost advantage to grow over time. We typically earn service fees tied to the units of work that we perform on behalf of our customers’ programs and as our platform matures, we would expect our growing cost advantage to enable us to fully cover our direct costs, eventually enabling us to earn a modest margin. Service fees provide a strong foundation of predictable revenue that is independent of any commercialization efforts by our partners. 19 Table of Contents As we continue to scale the Foundry and build Codebase, we expect to drive further efficiencies and decrease our average unit costs. This presents us with a strategic choice going forward. We could retain these efficiencies and increase our margins or we could pass these efficiencies on to our customers, increasing the number of shots on goal and, therefore, the likelihood of program success given a fixed budget. We believe the right choice for long-term value creation is to pass the savings to our customers, reducing the barriers to adoption and driving increased demand for our platform. Our service fees are thus impacted by a number of drivers: • Number of active programs: We hope to dramatically increase the number of programs working on our platform over time, and if we are successful, we believe this will drive increasing service fees. • Units of work per program per year: If our Foundry becomes more efficient as we scale, we have the opportunity to run more experiments with the same budget. At the same time, technical advancements, such as our investments in AI, may allow us to achieve program goals with less work. • Average price per unit of work: If we bring on innovative technologies or step change improvements in existing Foundry services, we plan to pass capability and cost improvements on to our customers. If these new technologies or services are adopted across programs, we believe the average price per unit of work will fall over time. • Number of years per program: If our platform improves, we expect program duration to decrease over time. Some programs may still be charting new territories and take several years, but programs that are able to leverage substantial pre-existing Codebase (e.g., our Nth program in bulk protein production) should have shorter duration and, in general, greater Foundry capabilities should shorten program durations. The multi-year nature of an average cell programming project means that our service fees are recurring in nature. Additionally, given the lead times inherent in developing technical plans as part of a sales process, we have visibility into new service fee bookings. This provides a strong foundation for the business and allows us to be patient while we wait for downstream economics. Downstream Value Share As the key enabling technology for our customers’ products, we are able to earn a share of the value of the products that are created using our platform, an important component of the financial potential of most cell programs. We are quite flexible and have structured a variety of value sharing mechanisms, including royalties, lump-sum milestones, and equity payments. As Ginkgo has matured, we have seen a shift in our downstream value towards milestone payments and commercial royalties rather than equity. Because Ginkgo typically will have completed the program (and received associated service fees) prior to realizing downstream value, cash flows from the downstream value capture component generally fall straight to the bottom line as we incur minimal to no ongoing support or delivery costs once the strain is commercialized. This dynamic creates opportunities for outsized returns as our clients successfully commercialize products built on our platform. As we add more programs to the platform over time, we expect downstream value share to contribute income, and therefore we believe our overall margins and cash flow profile will grow significantly. The realization of potential revenue related to downstream value in the form of potential future milestone payments and royalties and/or equity consideration is dependent upon a number of factors, including our ability to successfully develop engineered cells, bioprocesses, data packages, or other deliverables, and the product development and commercialization success of our customers. Biosecurity Revenue Since the end of the COVID-19 public health emergency in May 2023, Ginkgo has transitioned its Biosecurity business to focus its efforts on building out scalable biosecurity infrastructure. As of February 2024, Ginkgo has partnered with 14 countries to form memorandums of understanding, active or pilot programs. Through these partnerships, Ginkgo works to operate programs for collections, testing, sequencing, and insights delivery on pathogen samples in different countries. Ginkgo is also investing in building our BIOINT offering and wraparound technical assistance services in 2024, in consultation with our existing network and additional public and private partners, as we think it has the potential to significantly drive revenue in the future. Our revenue flows are expected to become more recurring as we increasingly incorporate longer-term contracts with recurring monthly fee models for data, analytics, and services. Our Sustainable Advantage We have defined a unique business model over the past 16 years. The biotechnology industry has been product-centric for decades, with early horizontal platforms in life sciences frequently vertically integrating upon the development of the first 20 Table of Contents successful product on their platform. As Ginkgo has embarked on this journey, we have studied and learned from innovators and established platform companies in other industries as we built our platform and business. We now benefit from significant historical investments, a virtuous cycle that grows with scale, and a strong business model that is aligned with our customers’ outcomes. These establish a strong sustainable advantage that we believe will help establish Ginkgo as a true industry standard. Decade-plus head start in creating an industry standard platform Hardware, software and biological tools need to be tightly integrated to replicate our platform. We have spent over 15 years building the software, automation and data science to best support a high throughput, generalized platform and expect to continue investing in this area. Our software, automation and data infrastructure cannot be easily replicated without bringing together a number of rare, specialized skill sets. In addition, without the scale and demand to stress test a high throughput platform, we expect any newly developed platform would be suboptimal. We estimate that it took us over eight years of investment and iteration to reach cost parity with “by hand” cell programming. We believe competitors will find it difficult to justify the investment in the software, automation and data science needed for high throughput operations before they acquire matching high demand. Scale economics provide a structural cost advantage As the only scaled horizontal platform in this space, we have the broadest number of programs that can be run on our platform, providing the highest potential for scale economics. Other companies choose to target specific markets and vertically integrate into products with high expected value. This has a tendency to overfit the capabilities of their R&D team to their targets. As discussed above, our continued scaling and investment in flexible tools that can apply to a broad range of end markets helps us drive efficiencies in the Foundry and Codebase across our diverse programs. Furthermore, as we scale, we are able to leverage advanced technologies that are only practical at scale and also may obtain preferred pricing with a number of suppliers. Competitors may be unable to source equivalent technology or negotiate similar pricing without first achieving scale, a feat that is difficult to do with a narrowly focused R&D platform. Strong network and learning effects In addition to a raw scale economic, we also accumulate knowledge and reusable Codebase from each program that runs on the platform. Every program benefits from the programs that came before and generates benefits for other current and future programs. These learnings and reusable assets are cumulative, extremely hard to replicate, and increasingly valuable to our customers. Because our learnings are generated by the work we execute in our Foundry, the scaling in our Foundry drives a scaling in our rate of learning. Thus, there is a recursive element to our platform: as the platform gets better, it also improves faster—we are excited to make this advantage of our platform available to our ecosystem of cell programmers. As Ginkgo drives scalability through our models, we have heavily invested in the use and creation of AI foundational and fine-tuned models. Efficient use of AI is only possible with the use of massive amounts of data. Because of our access to large amounts of data, Ginkgo has the ability to build superior foundational models and from there, build fine-tuned models designed to cater to our customer needs. We believe this is and will continue to be a major asset to our current and future customers. Ginkgo’s value creation is aligned closely with customer success Our platform drives value for customers along two dimensions: reducing the cost of laboratory work via automation and increasing the probability of technical success due to cumulative data and learnings. Our financial model is aligned with those factors. As we gain efficiency, we drive further demand for cell programming, which drives our Cell Engineering revenue up. As both demand and probability of success increase, our risk-adjusted value share also increases. Our model only requires we share in a small fraction of the downstream value created by our programs, providing our customers the opportunity to generate and retain significant value. Ultimately, this encourages broader adoption of our platform across industries. Furthermore, we seek to maintain close relationships with our customers, supporting their work, and earning their loyalty and satisfaction. The breadth and highly integrated nature of our platform makes it inefficient for a customer to simultaneously work with Ginkgo and any theoretical competitor. As there is not yet a standard interface for cell programming, it requires an upfront investment to learn how to choose and design programs to make the best use of our platform. These substantial switching costs are expected to be a long-term driver of customer retention. We are uniquely positioned to attract the top cell programmers 21 Table of Contents Just as the top software programmers want to work with the latest technologies, we believe the top cell programmers will be attracted to our industry leading platform and access to its unique capabilities. Our ability to hire and retain the best cell programmers as internal users and developers of our platform pushes us to continually improve and also builds a base of Ginkgo-trained experts. If these Ginkgo trained cell programmers move on to roles and opportunities in product-specific companies, we expect they will become ambassadors for the Ginkgo approach in their next role, expanding our reach into potential customers. History of investing in credibility and trust Let’s face it, GMOs have an image problem. This image problem has led to activities by the first generation of genetic engineering companies that backfired: lobbying against transparency in labeling laws, trying to “rebrand” GMOs with different terminology, and other efforts that have failed to build trust and engagement with stakeholders. We have taken a different approach. Rather than avoid the term, we’ve championed transparent labeling, sought to engage and build trust through open dialog, and enthusiastically embraced the potential for GMOs to do great things. We don’t seek to make GMOs acceptable through branding; we aim to make GMOs that people love. Figure 11: Ginkgo seeks to make GMOs that people love. Doing so requires care and attention to both the technical and social aspects of our platform and its impacts. This means investing in biosecurity and embedding it into our platform and how we operate (see below). This also means engaging with the social complexities of science and technology with a diverse group of people. We strive for a company culture based on a foundation of Diversity, Equity and Inclusion (see also the sections titled “—The Impact of Cell Programming—ESG is in our DNA” and “—Our People & Culture”), and aim to engage different perspectives through our creative residency and through our magazine, Grow. Through both our internal and external efforts, we seek to engage with the realities of what has made genetic engineering an ESG risk historically, and work towards equitable and positive impact. Early investments in biosecurity Too often, security infrastructure is built long after a new technological area has been built and commercialized, after security implications have already reared their heads and governments and corporations alike are left scrambling. We take a different approach—proactively acknowledging the risks inherent in biotechnological advances and addressing potential vulnerabilities. We aim to build biosecurity in tandem with biotechnology and the bioeconomy, through strong, early collaborations across the public and private sectors. For instance, we started working on technology to screen DNA sequences for potential threats through an Intelligence Advanced Research Projects Activity (IARPA) program in 2017—a security need that became a critical global policy priority with the emergence of AI chatbots in 2023. We’ve built on that work with IARPA through additional projects to 22 Table of Contents identify whether a DNA sequence is engineered (yielding our ENDAR platform, see above) and create a cellular “flight recorder” to support attribution. For years, we have also contributed to and continue to learn from committed stakeholders and communities, including, for example, within the International Gene Synthesis Consortium and at the National Academies of Sciences, Engineering, and Medicine. The COVID-19 pandemic drove our Biosecurity ambitions even higher, to formally establish a business unit that actually builds the global infrastructure needed to predict, detect, investigate, neutralize, and attribute biological threats, whether from Mother Nature, bioerror, or bioterror (see above). We choose to take a systemic approach—scoping biosecurity along the lines of modern cybersecurity infrastructure for protecting the digital economy—because it will be a critical piece of enabling the bioeconomy to grow to a similar scale. We’re working with national governments, scientists, and other partners around the world to move the needle on biosecurity vulnerabilities, from the ground up. We’re also routinely helping policymakers to understand the evolving threat surface, shedding light on everything from the increasing trends in likelihood and severity of pandemics of natural origin to the biorisk of new generative AI models. We believe that advancing biosecurity across the whole ecosystem and leading its development will allow the bioeconomy to advance, and our platform and customers to lead within it. Our Growth Strategy We are seeking to usher in a new paradigm for cell programming. It took us over eight years of basic research and investment in software, automation, data science and scale to reach parity with the status quo of individual scientists conducting experiments by hand at a lab bench. It took us several more years to demonstrate business model maturity: delivering a platform with enough value-add to customers that we could cover the cost of cell engineering R&D programs while building Codebase and sharing in the downstream value of our programs. We believe that we are now at an inflection point where we have the opportunity to become the industry standard. We see several drivers of this evolution and growth. Scale our platform and continue to drive efficiencies and improvements As discussed above, our platform improves with scale and to date we have observed a positive feedback loop between our Foundry and Codebase. As we scale capacity and demand on the Foundry, we expect our average unit costs to fall, creating a better value proposition for our customers as their program budgets stretch further and drive more demand. Similarly, Foundry output also grows our Codebase, which supports better program execution and helps with building out our AI models, creating a better value proposition for our customers as well. We occupy over 325,000 square feet at our headquarters and maintain state-of-the-art machinery and laboratory equipment. We have built more than 50 custom integrated work cells, consisting of robotic automation systems, mass spectrometry, fermenters, sequencers, and more. We have the capabilities to engineer dozens of species of organisms from bacteria to fungi to mammalian cells. We have worked on enabling products as varied as polymers, bacterial therapeutics, bulk protein production, novel antibiotics, fine chemicals, and more. We have been able to work on a diversity of programs while consistently driving efficiencies in the Foundry with scale. We expect to accelerate growth in capacity by integrating new technologies across our existing footprint, building new Foundry space, and investing in software, automation and data to increase utilization. Leverage our proof points to grow within all industries We have now established proof points of success in a diverse set of end markets, in several cases far exceeding our customers’ specifications. When engaging with existing customers or potential new customers in similar or adjacent industry verticals, we can point to these case studies of success to demonstrate the value of our platform. This reduces the barriers to adoption, helps us grow our customer base, and increases the number of new programs under contract. Importantly, the reusable Codebase we generate from these new programs enables us to stay ahead of vertically focused competitors. Grow with existing customers Once we establish a relationship with a customer, there is significant room to expand the scope of our program engagements. We are able to grow with our customers and/or expand into other existing pockets of R&D spending. We have seen customers expand from one early program to five or ten programs a few years later and each new logo we add has the potential to become a true platform partner. 23 Table of Contents When we work with companies from their inception (or at least from the inception of their biotech investments), we enable them to avoid significant fixed cost investments and benefit from our economies of scale. Our relationship with these customers is extremely strong, as we are the core technology powering their R&D efforts. As a result, when these customers scale, their usage of our platform typically scales commensurately. For companies with existing, established biological capabilities, as we demonstrate the value of our flexible platform, we are able to grow our relationships to complement their core capabilities and increase the probability of success. Reduce barriers to adoption by integrating with external R&D teams It can be easy to fall into the trap of assuming that new disruptive technologies must subsume existing ways of working. When hosted servers and SaaS started rising in prominence, corporate IT teams had to wrestle with changing integrations and demands. Some information technology departments were resistant to moving “off-prem” because they felt they were effectively outsourcing their jobs. In response, the leaders in this field, such as Dell, would sometimes hire their customers’ information technology departments and find them jobs within Dell simply to get past this internal resistance. The reality was that these technologies were ushering in a much more substantial era for information technology, which dramatically increased the demand for this type of talent. This centralization of the model (from every company having large information technology departments building customized code to a broader array of specialized software vendors) didn’t come at the expense of information technology and digital technologies, but enabled its flourishing across all industries. We see something similar happening in biotechnology today. Internal R&D teams are typically both very excited to learn about the power of our platform but are also understandably nervous about what “outsourcing” work to Ginkgo might mean for the future of their teams. We have the opportunity to help them see the benefit in a true partnership with Ginkgo. The vast majority of programs being run on the platform today are being run and managed by Ginkgo program teams—in-house scientists and engineers who are managing the R&D project to meet a customer’s specifications. Over time, we would like to build in enough standardized interfaces that a distributed network of scientists could access the platform directly through a well-defined integration and self-service layer. This transition will allow our program teams to devote more of their efforts to developing Codebase assets, enabling more rapid scaling, and reducing the barriers to adoption by our customers. There are significant technical hurdles for us to overcome in developing this technology, but it is on our roadmap and we are constantly thinking about how to “productize” individual workflows on the platform. As an example, we are developing CDKs that standardize common cell engineering workflows and assets, capturing best practices that we’ve identified. Building long-term global partnerships for advanced infrastructure in biosecurity In Biosecurity, we apply this mindset to our relationships with national governments. We work to build capacity in local and regional biosecurity institutions and integrate them into our operations as a means to add local value in each node while also strengthening the global network as a whole. We seek out partnerships that will allow us to have immediate local impact, grow our centralized Biosecurity data asset by providing unique insights, and further enable us to bring more countries and customers onto our platform. We often start our partnerships with well-established, off-the-shelf offerings, like our airport-based bioradar product, to lay the foundation for broader and longer-term engagements focused on growing and securing national bioeconomies. And we work with partners who enable us to accelerate our growth, such as Illumina: we co-market their next-generation sequencing technologies with our detection and analysis service offerings to expand our international reach. Our People & Culture A company is made of people. We have sought to bring together a diverse and multidisciplinary group of people who share our mission to make biology easier to engineer. Today, our extensive cross-functional team is collaborating to build our ecosystem, from organism designers to automation engineers, software developers the people team, business development to facilities management, finance to molecular biology. A culture built on care We’ve strived to grow a culture based on care. As engineers, it is easy to fall into the trap of thinking of ourselves simply as tool builders. Tools can be used in many different ways, both good and bad, and engineers often discuss their tools as value neutral. But tools reflect the social beliefs and biases of the people who make them: today this is becoming increasingly apparent, with more and more evidence of algorithmic bias being built into AI systems, facial recognition, and much more. As designers of the largest horizontal platform for cell programming, we are keenly aware of the need to care about how our platform is used. More significant than the impacts we have seen from digital platforms on our social world, biology is 24 Table of Contents our health, our bodies, our food, and our environment. As we build the tools for programming biology, we must also care how those tools are used, and ensure that the risks and benefits are transparently and equitably shared. A diverse, world-class team As of December 31, 2023 we had 1,218 employees. Building a horizontal platform for cell engineering and a biosecurity and public health unit requires collaboration between diverse skills and functions. It also requires deep technical expertise. Our employees are dedicated to the following functions: • Platform functions including organism engineering, design, DNA synthesis and assembly, genome engineering, protein engineering and characterization, transformation and transfection, next generation sequencing, assay development, ultra high throughput screening, analytical chemistry, synthetic chemistry, directed evolution, and fermentation. • Platform infrastructure functions including automation, software, development operations (“DevOps”), product management, data engineering, data analysis, and data science. • Deployment functions including upstream and downstream process engineering, project engineering, quality assurance and quality control. • Commercial functions including marketing, business development, alliance management, and corporate development. • Operational functions including bioinformatics, lab network management, delivery logistics and customer support. • Shared enabling functions including legal, people, operations, finance, information technology, information security, facilities, environmental health and safety, procurement, shipping and receiving, inventory management, laboratory operations, media preparation, and transformations. In addition to our employees, our success would not be possible without the collaboration and support of the broad network of partners, contractors, contingent workers and temporary staff who make up the Ginkgo team. Technologies reflect the values of the people who build them. Diversity, Equity, and Inclusion are valuable and necessary in their own right, but we believe that it is essential to build a diverse team where people from different backgrounds are included and empowered to speak up and shape the growth of this technology. We are committed to growing a diverse team and continuing to empower an inclusive culture with strong employee ownership and engagement. The full breadth of Ginkgo’s diversity and inclusion cannot be captured in demographic statistics, just as demographic categories cannot capture the full spectrum of diversity of human experience; however, we collect and report these numbers for transparency and as a lagging indicator of our efforts. As of December 31, 2023, 42.1% of our U.S. employees self-identify as an underrepresented gender (not cis male) and 14.1% self-identify as coming from an underrepresented racial or ethnic group in science and engineering (Black or African American, Hispanic or Latino, American Indian or Alaska Native, and Native Hawaiian and other Pacific Islander). We are not yet satisfied with these numbers and all teams have objectives around increasing diversity and building a culture of inclusion to ensure that diverse perspectives thrive. Laying the groundwork for strong employee engagement in the future As a founder-led company we have been able to infuse the organization with long-term strategic thinking from the start. The long-term engagement and mentality of our employees can be seen in our turnover: voluntary attrition is well below the industry average. The individuals who work at Ginkgo and build our platform care deeply about how that platform is used and the impact our company will have in the world. We hope to maintain the long-term mentality we have benefited from as a founder-led public company. We believe a workforce with strong equity ownership will make the wise decisions needed to build long-term value for our company and build a company whose long-term impacts make them proud. That is why we have implemented a multi-class stock structure that permits all employees (current and future), not just founders, to hold high-vote (10 votes per share) common stock. We believe that our multi-class stock structure will help maintain this long-term mentality and encourage long-term equity ownership by our employees, thereby resulting in increasing employee ownership over time. For more information, see “Risk Factors—Risks Related to Ginkgo’s Business—Risks Related to Our Organizational Structure and Governance—Only our employees and directors are entitled to hold shares of Class B common stock (including shares of Class B common stock granted or otherwise issued to our employees and directors in the future), which shares have ten votes per share. This limits or precludes other stockholders’ ability to influence the outcome of matters submitted to stockholders for approval, including the election of directors, the approval of certain employee compensation plans, the adoption of certain amendments to our organizational documents and the approval of 25 Table of Contents any merger, consolidation, sale of all or substantially all of our assets, or other major corporate transaction requiring stockholder approval.” Competition To our knowledge, there are currently no other cell engineering companies that serve all industries covered by our horizontal cell programming platform. The solutions and applications offered by potential competitors vary in size, breadth, and scope, and given our broad set of application areas, we could face competition in many different forms. We face competition from customers’ internal R&D departments and other research solution providers that largely conduct genetic engineering by-hand. We also compete against companies that seek to utilize synthetic biology technologies to develop specific products or target certain end markets. Additionally, competing platforms may emerge from various sources, including from joint ventures and partnerships between well-capitalized technology and life sciences companies. We identify the following three groups as our principal set of competitors: The Status Quo: “on prem” cell programming efforts The main source of competition we encounter is from potential customers choosing to build or maintain in-house cell engineering teams and capabilities. This status quo includes building out laboratory space and then hiring a team of highly trained scientists to conduct research, largely “by-hand” and with limited scale efficiencies. Some internal R&D operations maintain a full suite of capabilities and can design, build and test relatively complex pathways while others may have certain internal capabilities and need to outsource other elements to CROs. We believe this is far less efficient for the customer and likely to yield worse outcomes as customers get fewer shots on goal for a given program budget. That said, it can still be very difficult for companies to choose to trust Ginkgo with their R&D efforts versus building more traditional “on prem” labs. Smaller companies may feel like they’re “betting the farm” on Ginkgo, while larger companies may be sensitive to displacing existing R&D teams. As such, a key focus area for us is reducing the barriers to adoption for the platform by de-risking the upfront investment for earlier-stage companies and by helping larger companies integrate their scientists closely into our workflows and empower their scientists to manage requests directly so we feel more like a resource and partner than a fully outsourced provider. Investing in these areas is a key focus area for us going forward. Examples of traditional “synthetic biology” companies that have been vertically integrated from their founding with a focus on building products using synthetic biology include Amyris, Inc. (“Amyris”), Genomatica, Novozymes, DuPont, and DSM. Additionally, the vast majority of therapeutics companies that are leveraging genetic engineering have in-house capabilities, including Biogen, Novo Nordisk, Vertex, Regeneron, Bayer, and many others. These companies may be viewed as competitors to Ginkgo because they are creating products, using cell programming, that may compete with the products Ginkgo is enabling for our customers. However, as a horizontal platform, we view these companies not as competitors but as potential customers and focus not on “beating” them but rather on demonstrating our value proposition. Verticalized cell engineering platforms Within certain end markets, Ginkgo may compete against vertically-focused biotechnology companies providing cell engineering R&D capabilities to customers within a narrow set of end markets. While we believe the siloed nature of these companies limits their long-term potential, in the near-term, we may have a harder time penetrating those end markets given the incumbent vertical specialists in that space. The vast majority of these companies exist within therapeutic end markets given the history of cell engineering in that field. In theory, the expertise and learnings they develop from work in one field could be leveraged into neighboring end markets if these companies decided to adopt (and invest in) a more horizontal strategy. Examples of these vertically-focused platforms include AbCellera (antibody discovery), Codexis (enzymes), Senti Bio (cell therapy for oncology applications) and WuXi biologics (therapeutics). Other possible entrants We may also face competition from new entrants in the market, including well-capitalized technology companies with possible strategic interests in synthetic biology and its capabilities. Such companies may emerge as competitors given their access to capital, capacity to create multi-disciplinary teams across biology, chemistry, computer science and engineering, and flexibility to enter strategic ventures with life sciences companies. Biosecurity competition We’re unique in the global biosecurity market because our approach is global and comprehensively covers end-to-end biosecurity needs. We face competition from a small number of companies who operate in single biosecurity verticals, such as wastewater monitoring (e.g., Verily and Biobot, both primarily in the US) and digital biosurveillance and modeling (e.g., 26 Table of Contents BlueDot, Airfinity, and the Public Health Company), as well as internationally from BGI, China’s national champion for sequencing and diagnostics. As we partner with national governments, we also face competition from homegrown public solutions to particular vertical challenges, especially among high-income countries and large multilaterals with little history of engagement with the private sector. We have several important attributes that contribute to our competitive advantage: • Ginkgo’s cell engineering platform, which allows countries to partner with us across biosecurity and bioeconomy needs (e.g., in Serbia), sets us up for future biosecurity partnerships with medical countermeasure developers across the biopharma industry who work with the Foundry already, and accelerates our technical development through access to proprietary Codebase and AI; • unique technological tools, like our ENDAR platform for engineering detection, that to our knowledge has no equivalent capabilities in the world; • a comprehensive offering that allows customers to come to a single platform for multimodal physical and digital surveillance and integrated global insights, rather than fragmented approaches; • a foundation of partnerships with 14 countries and key multilaterals such as Africa CDC, African Risk Capacity, and the International Livestock Research Institute; and • global leadership in the airport-based pathogen monitoring space. Intellectual Property Overview: Foundry and Codebase As discussed above, Ginkgo’s two core platform assets include: • Ginkgo’s Foundry, which enables high-throughput cell programming; and • Ginkgo’s Codebase, which includes reusable biological assets that can be used to accelerate cell programs. Ginkgo protects each of these core assets—the Foundry and the Codebase—through a combination of patents and trade secret protections. Patents Our general policy has been to seek patent protection for those inventions likely to be incorporated into our offerings. Many of our collaboration agreements also provide a limited exclusive patent license to our collaboration partners relating to new technology developed in the collaboration. We typically retain the right to outlicense patents developed in connection with collaborations to third parties outside the scope of the exclusive license granted to our collaboration partner. Our worldwide patent portfolio includes patents acquired in transactions over time, including, most significantly, our acquisitions of Gen9 in 2017; Novogy in 2020; and Zymergen Inc. (“Zymergen”) in October 2022. Because these acquisitions more than doubled the size of our patent portfolio, and because the strategic priorities of the companies we acquired often differed from Ginkgo’s priorities, we may decide that it is in our interest to abandon, sell, or otherwise dispose of certain patents or patent applications from these acquisitions or that we determine are no longer relevant to our business. Patents generally have a term of twenty years from the date they are filed. As our patent portfolio has been built over time, the remaining terms of the individual patents across our patent portfolio vary. No single patent or patent family is essential to Ginkgo as a whole or to any of Ginkgo’s subsidiaries. In addition to developing our patent portfolio, we license patents from third parties. We intend to pursue additional patent protection to the extent that we believe that it would be beneficial and cost-effective. We cannot provide any assurance that any of our current or future patent applications will result in the issuance of patents. We also cannot assure the scope of any of our future issued patents or warrant that any of our patents will prevent others from commercializing infringing products or technology. Trade secrets Ginkgo’s technology-related intellectual property that is not patent-protected is maintained as trade secrets. We employ a variety of safeguards to protect our information and trade secrets, including contractual arrangements with our employees, consultants, contractors and other advisors that impose obligations of confidentiality, assignment of inventions, and security; digital security measures; and physical security precautions. 27 Table of Contents We require confidentiality and material transfer agreements from third parties that receive our confidential data or materials, and we also incorporate confidentiality and material transfer precautions into our collaboration agreements. For example, in the course of a cell program, we might transfer samples of intermediate strains to the customer for testing and scale-up work and then transfer a final commercial strain upon completion of our work. To protect both intermediate and final strains, we use strain transfer agreements that document the contractual restrictions and controls we have put into place, typically including, in the case of intermediate strains, covenants requiring the customer to return or destroy all strain samples after testing. Trademarks and domain names Although our business is directed at sophisticated corporate customers rather than end consumers, we have trademark rights and registrations in our name, logo, and other brand indicia in the United States and other jurisdictions around the world. We also have registered domain names for websites that we use in our business, such as www.ginkgobioworks.com. Intellectual property transaction structure We earn revenue from collaboration agreements with customers under which we perform cell programming activities. Through our cell programs, we develop cells that produce or are products for our customers, which they market in their verticals. With respect to intellectual property, we have relatively standard transaction structures that apply to cell programs for a customer. In this situation, our collaboration agreements typically provide that Ginkgo will own all collaboration-related intellectual property (“Foreground IP”) concerning cell programming. To protect our collaboration partners’ investment in the collaboration and to provide them with a competitive advantage from working with Ginkgo, Ginkgo provides a limited exclusive license to patents within the Foreground IP that cover the product, usually within a specified field. However, our terms may vary. We typically do not provide exclusive licenses to unpatented Foreground IP (i.e., trade secrets and other know-how) that results from a collaboration. In our typical deal structure, we also do not provide exclusive licenses to our “background” intellectual property—i.e., the intellectual property, whether patented or unpatented, that we developed before entering into a collaboration or develop independently from our work in the collaboration. We believe that our transaction structures allow us to maximize the reuse of Codebase across programs and ensure that technology we develop does not lie fallow. In-License Agreements In addition to our proprietary methods and technologies, we also non-exclusively in-license certain intellectual property assets from third parties. Amyris Partnership Agreement On October 20, 2017, we entered into a partnership agreement (the “Partnership Agreement”) with Amyris, which, as amended from time to time, terminated all prior agreements between Ginkgo and Amyris. In the Partnership Agreement, Amyris, among other things, granted us a non-exclusive license effective as of June 28, 2016 (the date of an earlier agreement between the parties) under all of Amyris’s rights in and to certain specified microbial strains, and under all patents and applications associated with such microbial strains, to make, have made, use, sell, offer to sell and import any products other than farnesene and/or farnesene derivatives that are chemically produced from farnesene. The license is subject to any previous exclusive licenses provided to third parties and is royalty-free, fully paid-up, sublicensable, non-exclusive and perpetual (i.e., it survives termination or expiration of the Partnership Agreement except in the case of our insolvency). Strateos Collaboration Agreement On October 2, 2017, we entered into a collaboration agreement with Strateos, Inc. f/k/a Transcriptic, Inc. (“Strateos”), which was amended and restated on April 20, 2021 (the “Strateos Collaboration Agreement”). Under the Strateos Collaboration Agreement, Strateos granted us a non-exclusive, perpetual, irrevocable, fully paid-up, royalty-free license under certain intellectual property rights to use its software platform in a range of activities relating to our business, including, among other things, developing and commercializing cell lines, developing data packages, providing foundry and analytical services and performing diagnostic testing. The Strateos Collaboration Agreement expired in 2022 and we retain a license to use Strateos’ software. 28 Table of Contents Suppliers Ginkgo’s suppliers for cell programming operations comprise primarily manufacturers and distributors of life science tools, consumables and equipment as well as certain specific providers of contract research, development and manufacturing services. We will sometimes enter into long-term, strategic partnerships with innovative suppliers. Because of the significant scale of our Foundry’s operations, we believe we are often an early adopter and the largest customer at scale of certain new life science tools and technologies. We will also occasionally acquire technology or Codebase assets for strategic reasons and because we can integrate the technology effectively into our platform — Zymergen, Altar, and Circularis Biotechnologies, Inc., (“Circularis”) are recent examples. Our software, automation, data, information technology, DevOps and information security functions utilize various third party software and information technology service providers, including AWS, for data storage and processing. We also routinely engage a variety of third parties for professional services, contract employment services and consulting services. Government Contracts We have entered into agreements with governmental entities and contractors in the past to serve as a U.S. government contractor or subcontractor and may do so again in the future. See “ Risk Factors—Risks Related to Governmental Regulation and Litigation—We have pursued in the past and may pursue additional U.S. Government contracting and subcontracting opportunities in the future and as a U.S. Government prime contractor and subcontractor, we are subject to a number of procurement rules and regulations. ” Government Regulations Our business, or the business of our customers, may be regulated by the FDA and other federal authorities in the United States, including the U.S. Federal Trade Commission (“FTC”), U.S. Department of Agriculture (“USDA”), U.S. Drug Enforcement Administration (“DEA”) and U.S. Environmental Protection Agency ("EPA"), as well as comparable authorities in foreign jurisdictions and various state and local authorities in the United States. Failure to comply with applicable regulations may result in enforcement actions, civil or criminal sanctions, and adverse publicity. FDA regulation We provide cell engineering and product discovery services to customers engaged in the manufacture of foods, cosmetics and pharmaceutical products. The FDA regulates the research, development, testing, quality control, import, export, safety, effectiveness, storage, recordkeeping, premarket review, approval or licensure, processing, formulation, manufacturing, packaging, labeling, advertising, promotion, marketing, distribution, sale, post-market monitoring and reporting of our customers’ pharmaceuticals, cosmetics and food products, and the FTC also regulates the advertising and promotion of these products. We have acted as a systems integrator and authorized distributor of certain COVID-19 over-the counter diagnostic tests manufactured by independent third parties. We worked with laboratory partners that provide surveillance testing services as part of the COVID-19 and other pathogen surveillance testing services we offer, and these tests and test kits may be subject to regulation by the FDA. In particular, the tests and test kits used in our testing services may be subject to regulation by the FDA as medical devices, and may be required to comply with the requirement that such products have obtained clearance, approval, or other marketing authorizations, before they can be commercialized, as well as post-market requirements such as adverse event reporting and restrictions on labeling, marketing, and distribution. Laboratories must seek FDA marketing authorization and otherwise comply with FDA device regulations when marketing COVID-19 Laboratory Developed Tests (“LDTs”). An LDT is an in vitro diagnostic test that is intended for clinical use and is designed, manufactured, and used within a single laboratory. LDTs are classified as medical devices, but the FDA has historically exercised enforcement discretion and has generally not enforced FDA requirements, including premarket review, with respect to laboratories that offer LDTs. However, FDA intends to phase out its enforcement discretion for LDTs. While HHS and FDA have announced their intention to require premarket review of COVID-19 LDTs, either agency may change its position in the future. Medical products, including COVID-19 tests, that are granted a clearance, Emergency Use Authorization (“EUA”), or other marketing authorization must comply fully with the terms and conditions provided in the clearance, EUA, or other marketing authorization. For example, EUAs for COVID-19 tests may include conditions of authorization applicable to the EUA holder, authorized distributors and authorized laboratories. Noncompliance with applicable requirements could result in negative consequences, including adverse publicity, judicial or administrative enforcement, warning letters or untitled letters from the FDA, mandated corrective promotional materials, advertising or communications with doctors, and civil or criminal penalties, among others. The FDA can also withdraw marketing authorization for the applicable product, and in 29 Table of Contents the case of a product subject to an EUA, the FDA may require EUA holders to transition to permanent marketing authorization which could impact some of the tests in our supply chain. DEA regulation We are engaged in the research, development, and export of certain products that may be regulated as controlled substances, including microbes designed to generate precursors to cannabinoids or other chemical intermediates. The Controlled Substances Act of 1970, as amended from time to time, establishes registration, security, recordkeeping, reporting, storage, distribution and other requirements administered by the DEA. The DEA is concerned with the control of handlers of controlled substances, and with the equipment and raw materials used in their manufacture and packaging, in order to prevent loss and diversion into illicit channels of commerce. The DEA regulates controlled substances as Schedule I, II, III, IV or V substances. Schedule I substances by definition have no established medicinal use, and may not be marketed or sold in the United States. Schedule I substances are considered to present the highest risk of abuse, and Schedule V substances the lowest relative risk of abuse among controlled substances. Marijuana is classified as a Schedule I controlled substance. However, the term does not include “hemp,” which means the cannabis plant and any part of that plant, including the seeds and all derivatives, extracts, cannabinoids, isomers, acids, salts, and salts of isomers, whether growing or not, with a delta-9 THC concentration of not more than 0.3% on a dry weight basis. Annual registration is required for any facility that manufactures, distributes, dispenses, imports or exports any controlled substance. The registration is specific to the particular location, business activity and controlled substance schedule. For example, separate registrations are needed for import and manufacturing, and each registration will specify which controlled substance schedule is authorized for that activity. The DEA typically inspects a facility to review its security measures prior to issuing a registration. The DEA requires “effective controls and procedures” to guard against theft and diversion of controlled substances. Security requirements vary by controlled substance schedule (with the most stringent requirements applying to Schedule I and Schedule II substances), type of business activity conducted, quantity of substances handled, and a variety of other factors. Required security measures include background checks on employees and physical control of inventory. While the specific means by which effective controls and procedures are achieved may vary, security practices may include use of cages, surveillance cameras and inventory reconciliations. Records must be maintained for the handling of all controlled substances, and, in certain scenarios, periodic reports made to the DEA. Reports must also be made for thefts or losses of any controlled substance, and disposal of controlled substances must adhere to various methods authorized by the regulations. In addition, special authorization and notification requirements apply to imports and exports. Failure by registered establishments to maintain compliance with applicable requirements, particularly as manifested in loss or diversion, can result in enforcement action. The DEA may seek civil penalties, refuse to renew necessary registrations, or initiate proceedings to revoke those registrations. In certain circumstances, violations could eventuate in criminal proceedings. Individual states also regulate controlled substances. Laboratory Licensing and Certification Requirements The clinical laboratories we partnered with for our COVID-19 testing program are subject to federal oversight under the Clinical Laboratory Improvement Amendment of 1988 ("CLIA"), which requires all clinical laboratories to meet certain quality assurance, quality control and personnel standards. Laboratories also must undergo proficiency testing and are subject to inspections. Standards for testing under CLIA are based on the complexity of the tests performed by the laboratory, with tests classified as “high complexity,” “moderate complexity,” or “waived.” Laboratories performing high complexity testing are required to meet more stringent requirements than moderate complexity laboratories. Certain of our partner laboratories must undergo on-site surveys at least every two years, which may be conducted by the Centers for Medicare and Medicaid Services (“CMS”) under the CLIA program or by a private CMS-approved accrediting agency. In addition, we hold CLIA Certificates of Waiver and may perform certain CLIA-waived tests on behalf of our clients, which subjects us to certain CLIA requirements. The sanction for failure to comply with CLIA requirements may be suspension, revocation or limitation of a laboratory’s CLIA certificate, which is necessary to conduct business, as well as significant fines and criminal penalties. The operations of our partner laboratories and our laboratories holding CLIA Certificates of Waiver are also subject to state and local laboratory regulation. CLIA provides that a state may adopt laboratory regulations different from or more stringent than those under federal law, and a number of states have implemented their own laboratory regulatory requirements. State laws may require that laboratory personnel meet certain qualifications, specify certain quality controls, or require maintenance of certain records. No assurances can be given that we or our partner laboratories will pass all future licensure or certification inspections. 30 Table of Contents Our facilities and laboratories hold local, state and federal permits, licenses and registrations necessary for compliance in specific work and operations, including from the Massachusetts Water Resource Authority, Boston Fire Department, Massachusetts Department of Environmental Protection, Boston Public Health Commission, Cambridge Biosafety Committee, Massachusetts Department of Public Health, USDA and DEA. Federal Select Agent Regulations Our research facilities that synthesize DNA sequences or perform other activities could become subject to the FSAP, which involves rules administered by the CDC and the USDA Animal and Plant Health Inspection Service (“APHIS”). The FSAP regulates the possession, use, and transfer of biological select agents and toxins that have the potential to pose a severe threat to public health, animal or plant health, or animal or plant products. FSAP regulatory requirements include: (i) registration with the CDC and/or APHIS for research facilities that deal with the select agents and toxins; (ii) submission to periodic biosafety and security inspections; and (iii) reporting of theft, loss or release of select agents. Federal agency enforcement actions for violations of FSAP regulations can include the initiation of corrective actions, complete or partial suspension or revocation of select agent registrations or civil or criminal liability. Genetically Modified Materials Regulations Our technologies and the technologies of our customers involve the use of genetically modified cells, organisms and biomaterials, including, without limitation, GMOs and genetically modified microorganisms ("GMMs"), and their respective products. In the United States, the FDA, the USDA through its APHIS, and the EPA are the primary agencies that regulate the use of GMOs, GMMs and potential products derived from GMOs or GMMs or Genetically Modified Materials, pursuant to the Coordinated Framework for the Regulation of Biotechnology. The FDA reviews the safety of food consumed by humans and of feed consumed by animals under the Federal Food, Drug and Cosmetic Act (“FDCA”). Under the FDCA, food and feed manufacturers are responsible for ensuring that the products they market, including those developed through genetic engineering, are safe and properly labeled. In addition, the FDA must approve the use of any food additives, including GMOs, before marketing. USDA's APHIS examines whether a plant itself presents a “plant pest” risk under the Plant Protection Act (“PPA”). Specifically, APHIS is responsible for regulating the introduction (i.e., importation, interstate movement or release into the environment) of certain GMOs and plants under the plant pest provisions in the PPA to ensure that they do not pose a plant pest risk. APHIS finalized changes to the PPA’s implementing regulations with respect to certain GMOs in May 2020. A person or organization may request a regulatory status review from APHIS to determine whether a GMO is unlikely to pose a plant pest risk and, therefore, is not regulated under the plant pest provisions of the PPA or the regulations codified at 7 C.F.R. Part 340; requesting a regulatory status review tends to assume the GMO at issue does not otherwise fall within a regulatory exemption. If the GMO does not qualify for an exemption or if the APHIS regulatory status review process finds that the plant poses a plausible plant pest risk, then the GMO may require an APHIS permit, i.e., be a regulated article under Part 340. A regulated article may be subject to APHIS for the environmental release, importation, or interstate movement of the GMO or its progeny. EPA regulates, under the Federal Insecticide, Fungicide and Rodenticide Act (“FIFRA”), the pesticides (including plant incorporated protectants) that are used with crops, including GMO herbicide-tolerant crops. FIFRA generally requires all pesticides to be registered before distribution or sale, unless they are exempted. Under FIFRA, a pesticide registrant must demonstrate that the pesticide at issue, when used pursuant to its specifications, “will not generally cause unreasonable adverse effects on the environment” to secure a registration. EPA must approve each distinct pesticide product, each distinct use pattern, and each distinct use site. In addition to EPA’s FIFRA authority, EPA also regulates potential human health impacts from pesticides under the FDCA. EPA does so by establishing “tolerance levels” (i.e., “the amount of pesticide that may remain on food products”) under the FDCA. Certain genetically modified microorganisms that are not otherwise regulated under FIFRA and FDCA may be subject to EPA regulation under the Toxic Substances Control Act (“TSCA”). New microorganisms that are formed by combining genetic material from organisms in different genera (known as intergeneric microorganisms) may be subject to reporting requirements prior to production or distribution in commerce (Microbial Activity Commercial Activity Notice), or use in research and development (TSCA Experimental Release Application), unless the entity can meet all required criteria to obtain an exemption under TSCA. Federal and state data privacy and security regulations Numerous state, federal and foreign laws, including consumer protection laws and regulations, govern the collection, dissemination, use, access to, confidentiality and security of personal information, including health-related information. In the United States, numerous federal and state laws and regulations, including data breach notification laws, health 31 Table of Contents information privacy and security laws, including HIPAA, and federal and state consumer protection laws and regulations (e.g., Section 5 of the FTC Act), that govern the collection, use, disclosure, and protection of health-related and other personal information could apply to our operations or the operations of our partners. HIPAA, and its respective implementing regulations, imposes obligations on “covered entities,” including certain health care providers, health plans, and health care clearinghouses, and their respective “business associates” that create, receive, maintain or transmit individually identifiable health information for or on behalf of a covered entity, as well as their covered subcontractors with respect to safeguarding the privacy, security and transmission of individually identifiable health information. Violations of the HIPAA privacy and security regulations may result in civil and criminal penalties. HHS is required to conduct periodic compliance audits of covered entities and their business associates. HIPAA also authorizes state attorneys general to bring civil actions seeking either an injunction or damages in response to violations of HIPAA privacy and security regulations. In addition, certain state laws, such as the California Confidentiality of Medical Information Act, govern the privacy and security of health-related information in certain circumstances, some of which are more stringent than HIPAA and many of which differ from each other in significant ways and may not have the same effect, thus complicating compliance efforts. States including California, Virginia, Colorado, Connecticut and Utah have also enacted comprehensive privacy laws that are currently in effect, and similar laws have been passed or are being considered in several other states, as well as at the federal and local levels. Failure to comply with these laws, where applicable, can result in the imposition of significant civil and/or criminal penalties and private litigation. Privacy and security laws, regulations, and other obligations are constantly evolving, may conflict with each other (thus complicating compliance efforts), and can result in investigations, proceedings, or actions that lead to significant civil or criminal penalties and restrictions on data processing. Ginkgo Corporate Information Ginkgo’s principal executive office is located at 27 Drydock Avenue, Boston, Massachusetts 02210, and Ginkgo’s telephone number is (877) 422-5362. Ginkgo’s corporate website address is www.ginkgobioworks.com. We make available on the Investor Relations section of our website, free of charge, our annual reports on Form 10-K, quarterly reports on Form 10-Q, current reports on Form 8-K, proxy statements, and Forms 3, 4 and 5, and amendments to those reports as soon as reasonably practicable after filing such documents with, or furnishing such documents to, the U.S. Securities and Exchange Commission (the “SEC”). The SEC maintains a website ( www.sec.gov ) that contains reports, proxy and information statements and other information regarding issuers that file electronically with the SEC. The information contained on, or accessible through, our corporate website is not incorporated into this Annual Report and should not be considered part of this Annual Report. The inclusion of the corporate website address is an inactive textual reference only.