How a Swiss Biotech Startup Is Building AI's Physical Lab

Jan de Vries · 2026-08-27

Here's something that might change how we discover new drugs and therapies. Adaptyv Bio, a startup based in Lausanne, Switzerland, just secured a massive $40 million Series A funding round. This isn't just another biotech story. It's about building the physical infrastructure for the AI revolution in biology. Think about it. AI can now design thousands of potential protein molecules in an afternoon. But then what? You have to test them in the real world. That's the bottleneck. As their lead investor, ACE Ventures, put it: "Most wet labs were built for a world where scientists tested twenty designs per target, not two thousand. Adaptyv closes this gap." ### The Automated, AI-Native Lab Founders Julian Englert and Daniel Nakhaee-Zadeh Gutierrez built something they call an "automated, AI-native lab." It's a full pipeline. You start with a digital protein sequence on a computer, and it ends with real, physical experimental data. They handle everything in-house: DNA synthesis, cell-free protein expression, and measuring how well things bind, all with built-in quality checks. The cool part? You don't need to be in the lab. Customers—which include top AI labs and big pharma companies—can submit designs through a simple web platform. Or, they can use an API. That means an AI agent itself can literally order its own experiments. It's a direct line from digital design to physical proof. ### From Seed to Scale: A Rapid Journey Let's rewind a bit. When ACE Ventures led their Seed round in late 2024, Adaptyv was a tiny team of nine with a promising idea. Fast forward to today, and the growth is staggering. - The team has tripled in size. - Revenue has grown by roughly 10 times. - They've brought DNA synthesis completely in-house. - They scaled their lab automation massively. - They launched a competition platform called Proteinbase, which attracted 680 participants and 10,000 protein designs in a single contest. They now count over 100 customers. We're talking about frontier AI research labs, companies in the top 5 pharmaceutical firms globally, and a new wave of AI-native drug discovery startups. Their public API is already integrated with major platforms like Benchling, Cradle, and Latent Labs. ### Proving the AI's Work: The Claude Benchmark Recently, they decided to put an AI model to the test in a very public way. They worked with Anthropic, sending 16 targets from their public competitions to Claude Science. The AI generated 1,320 protein designs, which Adaptyv then physically tested—anonymously, of course. The results were eye-opening. 95% of the designs were successfully expressed. Even more impressive, 354 of them actually bound to their target. That's a 26.8% hit rate. According to Adaptyv, Claude's designs performed so well they would have won 5 out of 6 of their human competitions. This test wasn't just for show. It proved Adaptyv is becoming the go-to validation partner. Google DeepMind uses a similar loop. Chai Discovery used them to validate a new antibody. Giants like Roche and Novo Nordisk are reportedly running therapeutic design cycles about four times faster using Adaptyv's system than they can with their own internal pipelines. ### What's Next? A Biological Gigafactory So, what does a startup do with $40 million? Adaptyv is planning growth on two fronts: horizontal and vertical. Horizontally, they're about to triple their lab capacity. They're expanding their home base in Lausanne and opening a brand new lab and office in London by the end of 2026. The team is set to grow from 25 to around 60 people, focusing on production, automation, and software. Vertically, they're moving into tougher biology problems. Right now, a big part of their work is measuring binding—a market worth a few hundred million dollars a year. They plan to move into: - Full biophysical characterization of molecules. - Next-generation therapies like peptides, antibody-drug conjugates (ADCs), and degraders. - Eventually, even cell-based functional testing. The founders' ultimate vision is audacious. They want to build a "biological gigafactory." A place with so much automated experimental capacity that AI can learn from physical reality as quickly and cheaply as it currently learns from text on the internet. Their goal is simple: to make lab throughput stop being the thing that holds discovery back. For anyone watching the fusion of AI and biology, this is one to watch closely.