Machine learning meets the wet lab to design enzymes drugmakers couldn't build before.
Somewhere in the flat sprawl of the San Francisco peninsula, a screen is running an experiment that no cell is alive to witness. There's no petri dish, no incubator humming with living things. Instead there is a cell-free reaction - biology's machinery pulled out of the cell and put to work in the open - and a machine learning model reading the results by the thousand, deciding which enzyme to try next.
This is Ridge Biotechnologies on an ordinary Tuesday. The company is small enough to fit in a decent conference room and quiet enough that until September 2025 most of the industry had never heard its name. Then it stepped out of stealth with $25 million and a claim that is either audacious or obvious depending on how much biology has disappointed you: that the reason so many promising drugs fail is not bad luck, but bad aim - and aim is an engineering problem.
Ridge Bio's founders don't talk about curing diseases in the abstract. They talk about linkers and payloads and site-specific modification - the plumbing of modern medicine. It is unglamorous work with enormous stakes. Get the plumbing right and a drug lands where it should. Get it wrong and it poisons the wrong tissue on the way to the target. Ridge decided the plumbing was worth rebuilding from the enzyme up.
The old bargain in drug development was cruel. The most powerful medicines - the ones that could actually kill a cancer cell - were often the hardest to aim, and the ones easiest to aim were rarely powerful enough to matter. Ridge Bio's whole thesis is that this bargain was never a law of nature. It was a limit of the tools. So they built new tools.
Designer biocatalysts that site-specifically modify therapeutic proteins - attaching payloads, extending half-lives, adding targeting - without changing the protein sequence, cell lines, or upstream manufacturing. The biotech equivalent of a firmware update. NativeLink-AXC gives antibody-drug conjugates a homogeneous, clean attachment.
Precision linkers that conditionally activate or release their payload only inside a target tissue. Designed to shrink the off-tumor toxicity that has quietly killed more promising drugs than bad chemistry ever did - and to unlock targets that were previously off-limits.
A new class of highly specific, enzyme-based therapeutics with differentiated mechanisms - built to reach the hard-to-reach and drug the difficult-to-drug. The pun is intentional: the company that designs catalysts wants to catalyze a new category of medicine.
Plenty of companies say "AI for drug discovery." Fewer actually wire the machine learning to a lab that can run experiments fast enough to feed it. Ridge Bio's advantage is the loop: proprietary models trained on massive cell-free datasets propose enzymes, high-throughput cell-free experimentation tests them by the thousand, and the results sharpen the next proposal. What used to be a years-long guessing game becomes weeks of directed design.
A serial entrepreneur and pioneer in cell-free synthesis and enzyme design. He previously co-founded SwiftScale Biologics - a cell-free manufacturing company acquired by Resilience - and was an Entrepreneur-in-Residence at Sutter Hill Ventures before building Ridge.
A leading figure in cell-free biology whose academic work underpins the platform's speed. His line - years compressed into weeks - has become the company's unofficial thesis statement.
The tagline - "Catalyzing the Next Generation of Precision Medicines" - is a quiet pun. The company literally designs catalysts.
Roughly 12 people raised $25M. That's about $2M of seed capital per employee - the economics of a very concentrated bet.
NativeLink can re-arm an antibody without changing its sequence, cell line, or a single upstream manufacturing step.
CEO Weston Kightlinger's last company, SwiftScale Biologics, was acquired by Resilience. Ridge is the second act.
Return to that quiet Tuesday screen. A year ago the enzyme it's proposing didn't exist. Now it does - designed, tested a thousand ways, and pointed at a target the industry had written off. Ridge Bio didn't leave stealth to make an existing drug slightly better. It left to build the ones no one else could. The plumbing, rebuilt from the enzyme up.