Somewhere on a bench in San Diego, a piece of plastic the size of a credit card is doing the work of a building full of robots. Inside it: as many as 200,000 tiny wells, each holding a single living cell, each one being photographed and sequenced at the same time. This is Zafrens, and it would very much like to know what every cell in the experiment is actually doing.
Drug discovery has a dirty secret, and it is not glamorous. For all the talk of AI and breakthroughs, most of the work still happens one well at a time, in plastic trays of 96, by people moving small amounts of liquid from here to there. Scale costs money. Money buys robots. Robots buy a little more scale. The biology, meanwhile, stays mostly hidden - you can see that a cell changed, but not why, and almost never across enough cells to be sure.
Who they are now
A benchtop that thinks in millions
Zafrens is a biotechnology company built at the seam where physics, chemistry, biology and engineering meet - and it is unapologetically an instrument company first. Its platform isolates, cultures, images, runs assays on, and sequences millions of individual single cells per day. The promise is not just more data. It is a particular kind of data: the link between what a cell is (its DNA, RNA and proteins) and what a cell does (its phenotype and function), measured at single-cell resolution, at a scale that used to require a warehouse.
The company is small on paper - around 29 people - and large in ambition. It has raised $31 million, exited stealth to a round of press, and signed collaborations with pharmaceutical companies it mostly declines to name. The product is real. The bet underneath it is bigger.
The problem they saw
Biology is slow because it is invisible
Here is the tension that runs through everything Zafrens does. A drug works, or fails, because of what happens inside cells. But the standard tools force a brutal trade-off: you can look at many cells shallowly, or a few cells deeply, and almost never both at once. Functional assays - the ones that tell you whether a treatment actually does the thing - have been stubbornly low-throughput. You learn that a cell killed a tumor, but you lose the cell before you can ask what made it a killer.
So discovery teams guess. They narrow candidates early, on thin evidence, and hope the survivors hold up. Most do not. The industry has industrialized almost every step of making drugs except the step that matters most: understanding cause and effect inside a living cell, across enough cells to trust the answer.
The founders' bet
Scale changes the question you can ask
Swamy Vijayan is not a first-time founder, which is the polite way of saying he has done this before and it worked. He holds a PhD in Physics from the University of Pennsylvania and spent years moving between optics, molecular biology, nanofabrication and computation - eventually leading the Nanobiology Group at Illumina. His last two companies are why investors took his calls: Omniome, a sequencing-instrument company, was acquired by Pacific Biosciences for up to $800 million; Plexium built ultra-high-throughput cell-based screening in a 150,000-well format. Co-founder Yi Zhang rounds out the founding team.
The Zafrens bet is a first-principles one. If you could make the experiments small enough - nanoliter-scale, etched rather than pipetted - you would not need the robots at all. You could fit a screen that once filled a room onto a single chip, and because every well is also a sequencing reaction, you would never have to choose between seeing a cell and reading it. Vijayan describes the deeper target in his own vocabulary: RNA-binding proteins, the "compiler layer of the cell," the machinery that decides how RNA becomes protein. Read that layer at scale, and a lot of previously locked doors open.
The short, fast history
The product
Z-Screen, and the family it spawned
The flagship is Z-Screen: a benchtop instrument paired with a proprietary plastic chip, roughly the footprint of a credit card, carrying 50,000 to 200,000 microwells. Each well can hold a cell, capture an image of it, and run an mRNA sequencing reaction - simultaneously. The scale comes from nanofabrication, not robotics, which is the whole trick. Where a conventional lab runs experiments in the hundreds, Zafrens claims a 500x to 2,000x increase in the number a single scientist can run at one timepoint.
Z-Screen
The benchtop chip platform. Up to 200,000 wells, imaging and sequencing in each, millions of experiments per day.
Active-Seq
Maps the molecular states of cells to how they actually function - linking perturbation to genotype and phenotype.
Active-Drug-Seq
Simultaneous phenotypic and molecular high-throughput screening across efficacy, safety and mechanism of action.
What can you do with it? The use cases read like a wish list for modern therapeutics: screening CAR-T and other cell therapies for which cells are genuinely polyfunctional, discovering T-cell receptors without depending on HLA matching, profiling bispecific antibodies, and pushing huge perturbation libraries - CRISPR screens, combinatorial chemistries - through a single, cost-efficient run. The pitch to a pharma team is blunt: stop guessing early, because now you can afford to look.
The scale argument, in one picture
Wells available in a single experimental format (log-ish, illustrative)
The grey bar is the entire history of cell biology's favorite tray. The dark bar is Tuesday at Zafrens.
The proof
Money, names, and a cap table that pays attention
Zafrens exited stealth with $23 million, led by deep-tech investor Prime Movers Lab and joined by BlueYard Capital, KOFA Healthcare, Global Brain, FoundersX Ventures, Alix Ventures, Possible Ventures, Iaso Ventures and Hawktail. In February 2025 it extended its Series A, bringing total funding to $31 million, with imec.xpand - the venture arm tied to one of the world's leading nanotechnology research institutes - joining the syndicate. For a company whose edge is nanofabrication, that last name is not a coincidence.
The platform has been demonstrated through collaborations with pharmaceutical companies across oncology, immunology and neuroscience. Zafrens is characteristically quiet about which ones. The interesting signal is the investor mix: instrument-savvy and nanotech-savvy capital, the kind that tends to show up when the hard part is the physics, not the slide deck.
The mission
Unlock the full story of every cell
Strip away the throughput numbers and the mission is almost literary: to read the full story of every cell, and to make the cell's hidden regulatory layers legible to the people designing medicines. Vijayan frames the company as a 100-year foundation - infrastructure for "smart people to do important things" - rather than a single drug or a single trick. The near-term work focuses on the RNA layer, using small molecules to modulate it, with an eye toward oral drug candidates for cell and gene therapies. The long-term work is simpler to state and harder to do: make cause and effect in biology something you can measure instead of infer.
Why it matters tomorrow
If the experiment gets cheap, the science gets brave
There is an old constraint in biology: experiments are expensive, so questions are conservative. You test the hypothesis you can afford. Zafrens is trying to break that constraint at the root. When a screen costs a fraction of what it used to and returns a thousand times the cells, you can ask reckless, sprawling, beautiful questions - the kind that find drugs nobody was looking for. That is the real product. The chip is just how it ships.
Five things worth knowing
- The scale comes from etched chips, not pipetting robots. No robotic arms required.
- The whole experimental platform is about the size of a credit card - yet holds up to 200,000 wells.
- Founder Swamy Vijayan has a Physics PhD and once led the Nanobiology Group at Illumina.
- His prior company Omniome was acquired by Pacific Biosciences for up to $800M.
- He calls RNA-binding proteins "the compiler layer of the cell." Programmers, take note.
Back to that bench in San Diego. The credit-card chip is still doing the work of a roomful of robots - but now picture it copied, a thousand times over, in labs that used to ration their curiosity. Zafrens did not just shrink the experiment. It made looking cheap enough to be brave.