Over the past decade, we’ve watched AI reshape software, tools, and infrastructure across the tech ecosystem. Now, that same transformative wave is hitting biotech, and it is not just hype.
Many modern biotech companies are not just layering machine learning on top of traditional drug discovery; they are utilizing AI to fundamentally reinvent how therapeutics are designed, validated, and even conceived. Meaning, we are now seeing drugs generated that would have been fundamentally impossible without AI. The intersection of AI and the life sciences is creating a launchpad for incredible innovation and inspiring a class of founders to focus deeply on advancing human health.
At our 2025 Annual Meeting, we brought together three leading biotech founders building at the frontier of scientific progress: Archon Bio’s James Lazarovits, Nosis Bio’s Jim Martineau, and Synthesize Bio’s Jeff Leek. These aren’t companies simply using AI as a tool. They are natively embedding AI capabilities from day one in all of their workflows, rethinking the boundary between platform and product, and using models, data, and automation to define entirely new classes of medicines.
We are at a crucial moment in the advancement of biotech, with more competition and challenges than ever before. AI-native biotech companies are no longer theoretical and will define the next generation of biopharma. The leading players are securing pharma partnerships, generating revenue, putting drugs in the clinic, and building next-generation therapies that couldn’t have existed five years ago.
If you’re a founder, investor, or operator in this space, the choices made now, about data strategy, AI development, product focus, and team structure, will define the next decade of medicine.
Proprietary Data is Crucial, Not a Nice-to-Have
In the tech world, you can often bootstrap models using public data. In biotech? Not so much. All three founders emphasized that off-the-shelf, public datasets don’t deliver signal, and in some cases will actually degrade model performance. The world of public data in biotech is insufficient, and while we’ve seen amazing advances in protein structure, like the Rosetta family of models and AlphaFold, leverage public datasets, the models that truly unlock drug design and product will emerge from the companies with the best data.
Synthesize Bio, for example, uses foundation models to generate synthetic wet-lab data that ultimately will lead to a leap forward in experiment generation and prioritization. Meanwhile, both Archon and Nosis actively design their own data generation loops to improve models and generate high-throughput data, with a focus on therapeutic properties. The consensus: Public data might help you get started, but only purpose-built proprietary data will allow you to win.
AI Won’t Replace Scientists, But Will Require a New Skill Stack
Forget the “AI will replace researchers” trope. What’s actually happening is more nuanced. AI is enabling and accelerating scientific workflows, freeing up human talent to spend more time in the lab and more time generating killer insights and therapeutic directions. Nosis Bio’s team can now generate experimental hypotheses in a massively parallel way and focus the majority of their time on interrogating the data outputs.
Companies like Archon are pushing the opposite edge: researchers who don’t know how to code or interface with AI models are increasingly bottlenecked. Their scientists are working seamlessly with protein generation models and AI tools to generate and refine their molecule designs. The emerging standard for scientific talent is AI-native and scientifically exceptional.
Product-Led Platform, Not Platform vs. Product
The term “platform” shows up in almost every AI-led biotech pitch deck. But as Jim put it, “The only measure of success is taking a medicine all the way to the patient.” His company, Nosis Bio, focuses on building repeatable systems that are deeply in service of specific therapeutic products.
That same mindset drives Archon’s team. James described Archon’s approach as a “product-led platform” that utilizes therapeutic progress against specific targets to validate and sharpen the system. The best platforms, they argued, won’t be built in isolation, they emerge from repeated success at the product level and from a precise focus on solving real therapeutic problems.
Nosis has secured major pharma partnerships by utilizing its platform to enable external innovation. The key, Jim explained, is only partnering where the fit is perfect and the opportunity cost is low. Nosis uses its proprietary models and data to work with partners who are aligned with its approach and capabilities, while the scale of the underlying technology still allows the team to largely focus on progressing their internal pipeline.
AI Is Now a Mandatory Tool for Biotechs To Stay Competitive
Biotech has always been an industry focused on innovation. Life-changing drugs come from pushing the boundaries of science and exploring new areas of biology. But the industry has long been bottlenecked by constraints on designs, experimentation, and data at scale.
AI is beginning to loosen some of these constraints. Whether you’re designing fundamentally novel proteins (Archon), optimizing RNA medicines for specificity (Nosis), or simulating entire experimental systems (Synthesize), AI opens up design spaces nature never explored.
Still, acceleration isn’t limitless. As Jeff explained, there’s an irreducible amount of time between designing a molecule and validating it in patients. The opportunity? Use AI to compress every loop you can between the idea and the outcome.
The Bottom Line
AI is not the next phase of biotech. It is already here, and the best companies are using it every day, from large pharma all the way to seed-stage startups. AI is the force multiplier allowing the next generation of therapies to be built, but we are still in the early days and need to build a proprietary foundation. The shift from “AI as a tool” to “AI as fundamental biotech infrastructure” changes everything: the teams you hire, the data you prioritize, and the milestones that matter.
If you’re building in this space, now is the time to define your data strategy, sharpen your product focus, and assemble the multidisciplinary team that can turn m
Related Insights
