AI has taken the world by storm. On the scientific front, the ability to mine huge datasets and make informed, actionable predictions has accelerated the pace of innovation. So it makes sense that startups at UC Berkeley’s Bakar Bio Labs, the university’s leading biotech incubator, would embed AI into their work to discover new medicines, redesign how they are delivered, and even create entirely new biological tools.
Bakar Bio Labs’ tenant companies are redefining the world of biotech, and they’re doing it with out-of-the-box solutions. For three of the tenants — Insamo, Aikium, and Profluent — AI is core to their technology and strategy.
Insamo
Solving problems of both cost and delivery using machine learning to engineer a pill
Insamo uses AI to reinvent how patients receive life-saving antibody therapies. “Everyone wants a pill, and they want it not to cost $30,000,” says CEO and founder Timothy Craven. Costly injectables like Humira and Dupixent are often priced at thousands of dollars a month and require cold-chain storage. Such medications are not widely accessible. Looking to solve the twin problems of cost and delivery, Insamo applies machine learning to design trillions of peptide molecules that mimic antibodies, but are engineered to be taken in pill form. By embedding AI directly into its drug discovery engine, the company can rapidly generate and screen massive libraries of candidates, narrowing in on those most likely to succeed in the clinic.
What began as a graduate school idea—with the founders investing just $20,000 of their own money—has grown into a team split between Berkeley and Sydney, Australia, with top biotech investors backing the effort. Craven describes Bakar Bio Labs as giving them an advantage, from flexible lab space to impromptu encounters with pharmaceutical executives and VCs at networking events that have led to major deals.
AIkium
Applying AI to identify drug targets that traditional technologies cannot reach
Aikium, also utilizes AI for its research and development, but takes a different approach, employing it to identify drug targets that traditional technologies can’t reach. CEO Eswar Iyer explains that most AI models in protein engineering are limited because they’re trained on only “point one percent of the proteome,” static snapshots that fail to capture how proteins actually move. To overcome this, Aikium has built what it calls a “yotta-scale” drug discovery engine, with 10^24 possibilities, that fuses AI with wet-lab screening to generate massive, high-quality datasets and design binders for previously inaccessible receptors.
“AI is here to stay. It’s not going away. It’s not a fad. Everybody knows it. Everybody accepts it.” [Eswar Iyer, CEO Aikium]
“AI is here to stay. It’s not going away. It’s not a fad. Everybody knows it. Everybody accepts it,” says Iyer, describing why building better data at scale is the only way to push the field forward. At Bakar Bio Labs, Aikium has found not only the facilities to expand on demand but also a collaborative culture and network that help transform its ambitions in AI.
Profluent
Applying large language models to protein sequences
Profluent, a Bakar Labs alum, treats DNA and proteins as a language, using large language models to learn that “grammar” and write new biology. It’s a bold vision, and it has already yielded breakthroughs, such as OpenCRISPR-1, the world’s first AI designed gene editor. In July 2025, the design and performance of OpenCRISPR-1 were published in Nature, which also ran independent coverage highlighting how new gene editing tools are created.
Beyond OpenCRISPR-1, Profluent unveiled ProGen3, a family of AI models for protein generation. Trained on a proprietary dataset called Protein Atlas v1 (PPA-1), which contains 80 billion protein sequences, Profluent reports their models can scale up to 46 billion parameters.
That momentum is spilling into partnerships. This month, Revvity announced a strategic collaboration with Profluent. This brings together Profluent’s range of novel AI-engineered enzymes with Revvity’s established Pin-pointTM base editing platform. The result is simplified access for customers to a therapeutically relevant base editing toolkit.
AI is redefining the future of healthcare – Join Bakar Bio Labs to be part of it
Together, Aikium, Insamo, and Profluent show how AI is redefining the future of healthcare—and they’re doing it at UC Berkeley’s top biotech incubator. Startups are offered state-of-the-art infrastructure, the flexibility to scale as they grow, and constant exposure to investors, pharmaceutical leaders, and the university’s extraordinary talent pool. For students, researchers, and entrepreneurs, the message is clear: the convergence of AI and biotechnology is happening now, and Bakar Bio Labs is the place to be if you want to be part of it.
About Bakar Bio Labs/QB3
Bakar Bio Labs is the flagship biotechnology incubator based at UC Berkeley, run in partnership with QB3. QB3 is a multicampus UC institute spanning Berkeley, UCSF, and UC Santa Cruz. Its mission is to bridge quantitative disciplines (physics, chemistry, engineering, computation) with biology and to accelerate the translation of research into real-world applications.
Bakar Bio Labs provides wet-lab and office space, shared core equipment (e.g., freezers, centrifuges, cell culture rooms), and operational support to early-stage biotech startups. The incubator can support as many as 40 early-stage companies from around the world focused on translating life science-based innovations that promise to improve human health. No UC affiliation is required to join. For information about how to join or form a partnership, visit the Bakar Bio Labs website.
Author
Media Intern
Bakar Bio Labs
Cover image caption: The Profluent Team
