Home›Companies›b12 Labs
b12 Labs

Making hard molecules for Pharma and Biotech

b12 makes high-value molecules that are otherwise hard to make. We help pharma companies plan and make new drugs, accelerating early-stage drug discovery from years to months. Andres built the first AI agent that can autonomously think and make molecules in a robotic lab, published in Nature Machine Intelligence (1000+ citations), and won the best paper award at the NeurIPS AI for Science conference. Zlatko is a 3x National Chemistry Olympiad champion, competed at IChO level, and medicinal chemist who has worked at Roche and leading lab robotics companies. Combined, we have an extensive network of hundreds of chemists, automation scientists, and AI 4 Science researchers. We run paid pilots with big pharma companies, demonstrating that our platform planned a set of experiments that achieved full conversion in a single attempt, eliminating the iteration cycle.
Active Founders
Zlatko Jončev
Zlatko Jončev
Founder/CEO
at b-12 we are developing AI agents for automated workflows in chemistry and pharma R&D, as a way of manifestating chemical super intelligence in physical world. our long-term goal is to create autonomous labs that can perform complex, multistep syntheses with minimal human intervention.
Andres Bran
Andres Bran
Founder/CTO
Before b-12, Andres built the first AI agent to plan and make molecules in a robotic lab (750+ cites in Nature Machine Intelligence, and Best Paper at NeurIPS AI for Science), along with several other works in agents and LLMs for chemistry (1100+ citations). He was an Intern at Max Planck (Math in Science), and PhD at EPFL (AI+Chemistry). Automation is the future of chemistry, but integration and human operability is the hidden demon nobody talks about. We're fixing that at b-12.
Company Launches
b12 Foundry - We Make the Molecules AI Designs
See original launch post

uploaded image


TL;DR:

AI can design a molecule in seconds, but making it can still take a chemist weeks. b12 Foundry turns a molecule on a screen into a compound in a vial, in a fraction of the time it would take a chemist. A combination of state-of-the-art AI agents plans the route, designs the experiments and runs them on robots in our partner lab, so you get your molecules, and clean data including failed reactions, without needing a lab of your own. Last year we launched here as a chemistry copilot. Today we work with several large companies, across pharma, flavor and fragrance, energy and biotech industries and universities, and the Foundry is actively making molecules for clients.

❌ The problem:

AI designs molecules faster than anyone can make them:

  • AI models can now design a molecule in seconds, but making it can still take a chemist weeks.
  • The most interesting designs sit outside the catalogues vendors build from simple, reliable reactions, so someone has to make each one by hand.
  • In practice, teams end up testing an easier cousin of the molecule their model wanted, or nothing at all.

Good chemistry data is hard to find:

  • Published data mostly shows the reactions that worked, so the failures that models need to learn from rarely get recorded.
  • Running clean, consistent experiments at scale takes a lab, robots and people that most AI teams don't have.

If you're building AI for drug discovery or materials, synthesis is quickly becoming your bottleneck.

✅ Our Solution:

https://youtu.be/HV_etgiCdkM?si=ypo5-kO_EwbztgpO

b12 Foundry combines state-of-the-art AI agents with robots in our partner lab, so we can plan and run complex chemistry for you, end to end.

How it works:

  • Send us a molecule or a workflow: Share a structure your model designed, or an experiment you want run.
  • Our agents plan the chemistry: They work out how to make the molecule and design the experiments for the hardest steps.
  • Robots run it in our partner lab: The agents run every experiment on the robots, including chemistry that is hard to do by hand.
  • You get compounds and data back: You receive the molecules plus every result, failed reactions included, and everything we run for you belongs to you.

👥 Our Team:

We are Zlatko and Andres, and in the past, we have:

  1. Worked with big pharma (Roche), Chemical Robotics (Chemspeed), and big chemical data (MaxPlanck);
  2. Pioneered AI agents for chemistry and retrosynthesis (at EPFL);
  3. Invented new chemistry.

When we launched here last year, we promised a full stack from planning to robotic execution. Since then, our customer base has grown well beyond pharma. Alongside three large pharma companies, including a top-3 global pharma that came back as a repeat customer, we now work with a flavor and fragrance leader, energy companies, biotech companies, and universities around the world. An AI drug discovery platform has also built our synthesis agent into its own pipeline, and this summer we opened the Foundry, which is actively making molecules for clients.

The team has grown with it. Adriana Faraone, our Research Scientist, runs the experiments in the Foundry, and Victor Sabanza Gil joins us from his PhD at EPFL, where he worked on Bayesian optimization for molecules and materials, the kind of AI that decides which experiment to run next. Together, we give anyone with a molecule on a screen extra brainpower and an additional pair of hands.

🙏 Our Ask:

Do you know a biotech or AI-for-science team whose molecules are hard to make, or who needs experimental chemistry data? We would love a warm intro!

If you've set up a wet lab in the Bay Area, we'd love 20 minutes of your advice, because our own lab here is next.

Send us an email at: founders@b12-labs.com or send them to our website: b12-labs.com

Previous Launches
Full Stack Chemistry Platform - from planning to robotic execution of experiments
YC Photos
Hear from the founders

How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?)

Andres and Zlatko met at EPFL, while working in the Lab of Artificial Chemical Intelligence. Zlatko was doing his postdoc and Andres was doing his PhD.

How did you decide to apply to Y Combinator? What was your experience applying, going through the batch, and fundraising at demo day?

We decided to apply after we had a YC partners visit our campus at EPFL. Experience was very intense, a lot of work and great advice on how to build a startup.

Jobs at b12 Labs
San Francisco, CA, US
$120K - $160K
0.50% - 2.00%
3+ years
San Francisco, CA, US
$130K - $180K
0.50% - 1.50%
Any (new grads ok)
b12 Labs
Founded:2025
Batch:Summer 2025
Team Size:3
Status:
Active
Location:San Francisco
Primary Partner:Diana Hu