The Cambridge startup trying to compress weeks of biological analysis into a single working day - and make every scientist their own supercomputer.
It is a Tuesday in a lab outside Boston. A biologist uploads a messy pile of sequencing files - the kind that used to mean a week of waiting and a polite email to whoever ran the cluster. She picks a workflow. She presses go. By the time her coffee is cold, the variants are called, the charts are drawn, and a bioinformatician three time zones away is already poking at the same dataset. No tickets. No "the pipeline broke overnight." No translation layer between the person who knows the biology and the person who knows the code.
That Tuesday is the entire argument for Watershed Bio. The company - legally Watershed Informatics - sells a cloud platform called Omics Bench, and the pitch fits on a sticky note: go from sample to therapeutic insight in a single day. It sounds like marketing. It is also, awkwardly for the skeptics, the most honest description of what the product is trying to do.
For two decades the life sciences have been drowning in their own success. Sequencing got cheap. Instruments got fast. A single study can now spit out genomics, transcriptomics, epigenomics, proteomics, and metabolomics - the field's beloved "multi-omics" - in volumes that would have been science fiction in 2005.
And then everyone went home and analyzed it on a laptop, a borrowed cluster, and a stack of scripts held together by hope. The biologists who understood the questions couldn't run the code. The bioinformaticians who could run the code became a permanent bottleneck. Somewhere in that handoff, discoveries quietly died of bureaucracy.
Watershed's founders looked at this and noticed something inconvenient: the limiting reagent in modern drug discovery wasn't the science. It was the software. There was, as CEO Jonathan Wang put it, a stubborn lack of centralized, collaborative platforms that served biologists and bioinformaticians at the same time. The two tribes had no shared workbench.
Jonathan Wang and Mark Kalinich met as undergraduates at MIT, which is the sort of origin story venture capitalists collect like baseball cards. What makes theirs slightly more entertaining is the detour. Before he ever touched a genome, Wang founded Domeyard, a high-frequency trading firm - a business that lives and dies in microseconds. He learned, expensively, what happens when infrastructure is the difference between an insight and a missed one.
So when he and Kalinich (now Chief Scientific Officer) started Watershed in 2019, the bet was almost finance-flavored: treat biological computation like a latency problem. Give scientists elastic supercomputing that spins up on demand, wrap it in workflows anyone can run, and the analysis stops being a multi-week negotiation. It becomes a button.
Co-Founder & CEO. MIT alum. Previously founded the high-frequency trading firm Domeyard - he traded microseconds before he traded base pairs.
Co-Founder & Chief Scientific Officer. Met Wang at MIT. The biology conscience to Wang's infrastructure instincts.
Canvas Ventures led the Series A; Bessemer and Accomplice were already in. Canvas's Paul Hsiao took a board seat.
Wang and Kalinich found Watershed Informatics, convinced the gap between biologists and bioinformaticians is a product, not a personality clash.
The team builds a cloud platform that unifies multi-omics tools into one ecosystem, aiming squarely at the space between "high-code" pipelines and "no-code" toys.
Canvas Ventures leads; Bessemer Venture Partners and Accomplice participate. Total raised reaches roughly $18.5M. Paul Hsiao joins the board.
A lean Cambridge team keeps pushing the same claim: sample in, insight out, before the day is done.
Strip away the brochure and Omics Bench does three unglamorous, essential things. It stores and harmonizes messy biological data in a secure environment. It runs analyses on every data type through customizable workflows - including ready-made AI tools - while tracking exactly where every result came from. And it hands those workflows a supercomputer on demand, so a job that once took weeks finishes in minutes.
The clever part is the range. The same platform handles whole-genome sequencing, single-cell and spatial transcriptomics, epigenomics, proteomics, metabolomics, microbial sequencing, and protein folding - each with its own visualizations. Most teams stitch that together from a dozen tools and a prayer. Watershed's answer is one login.
Access, manage, and harmonize complex datasets across an organization - with provenance tracked at every step.
Run and track analyses on any data type, including ready-to-use AI/ML tools, without rebuilding the pipeline each time.
Elastic, GPU-backed infrastructure finishes large analyses in minutes instead of weeks. The cluster wrangling is someone else's problem.
The headline number is time. Watershed describes turning analyses that once ran for weeks into runs measured in minutes - the same compression that high-frequency traders chase, pointed at the genome instead of the order book. Here is the shape of the argument.
Bars are scaled for readability, not to a single shared axis. Revenue is a third-party estimate, not a company figure - read it with one eyebrow raised.
Watershed states its mission without much hedging: empower every scientist to make breakthrough discoveries and accelerate the development of life-saving therapies. It is a big sentence for a small team, and they know it. The vision underneath is narrower and more believable - a single, unified platform where the data analysis stops being the part of research that everyone dreads.
There is an honest tension here worth naming. Watershed is a lean operation selling against deep-pocketed incumbents and the gravitational pull of "we'll just build it in-house." Its bet is that biologists are tired of being held hostage by their own tooling, and that drug-discovery teams will pay to get their afternoons back. The investors agree enough to have written eight figures of checks.
Drug discovery is a numbers game played against the clock. Every week shaved off an analysis is a week closer to a therapy that works - or proof that one doesn't, which is its own kind of progress. If Watershed is right that the bottleneck has quietly moved from the bench to the browser, then the company that fixes the browser fixes a surprising amount of the bench.
Back to that Tuesday. The biologist's coffee has gone cold, the variants are called, and the bioinformatician three time zones away has already left a comment. A week ago this would have been a ticket in a queue. Now it is a footnote in someone's afternoon. Watershed didn't make the science easier. It made the waiting disappear - and in this field, the waiting was always the expensive part.
Watershed Bio has not published an official YouTube channel or product-demo video that could be verified at the time of writing. When they post a walkthrough of Omics Bench, this is where it belongs - until then, the platform tour lives on their website.