Curi Bio

Right now, a heart is beating in a plastic well

It is the size of a grain of rice. It was grown from human stem cells. And in a Seattle lab, a sensor is quietly counting every twitch - no dye, no label, no scientist squinting through a microscope at 2 a.m. That well is one of twenty-four on a single plate, and the plate is one of many. This is an ordinary Tuesday at Curi Bio.

Curi Bio builds the equipment, the cells, and the software that let drug developers watch human tissue do the one thing animal models and flat layers of cells never quite manage: behave like a human. The company sits at the intersection of stem-cell biology, mechanical engineering, and machine learning - three disciplines that historically did not share a coffee machine, let alone a lab bench.

Cells in a dish should resemble cells in a body. Most of biotech has spent decades hoping that was close enough.

In December 2025 the company closed a $10 million Series B led by DreamCIS, a Seoul-based contract research organization. It was the kind of round that doesn't make national headlines but tells you a quiet thing: serious people who run drug trials for a living decided this was worth backing.

The drug that works in a mouse and fails in a human

Here is the uncomfortable arithmetic of modern medicine. A compound passes its animal studies. It looks promising in a dish of flat-grown cells. It advances to human trials - and then most of them fail. The gap between "worked in preclinical" and "worked in people" swallows years and billions of dollars, and patients wait on the far side of it.

The reason is unromantic: a mouse is not a person, and a single layer of cells stuck to plastic is not a tissue. Real muscle pulls. Real heart cells beat in coordinated rhythm. Real biology is three-dimensional, mechanically active, and stubbornly human. The cheaper the model, the further it drifts from the thing you actually want to predict.

Sources vary on the exact failure rate; the direction of the problem does not. (Approximate figures.)

Preclinical animal and 2D cell models have failed to be good predictors of human response.Michael Cho, Chief Strategy Officer

Start with the surface, then the whole tissue

The company began in 2015 as NanoSurface Biomedical, a spinout from the University of Washington incubated at CoMotion Labs in Fluke Hall. The early insight was almost architectural: cells take their cues from the surface they grow on. Give them a nanopatterned surface that mimics the body's own extracellular matrix, and they organize, mature, and start behaving like the real thing.

That nanopatterned plate was the wedge. From it grew a broader bet, led by CEO Nicholas A. Geisse (a physicist-turned-biologist who started as the company's chief scientist) and co-founder and Chief Business Officer Elliot Fisher: don't just sell a better surface - sell the entire workflow. Cells, tissue, the instruments to stimulate and measure them, and the software to make sense of the data.

They didn't pitch a gadget. They pitched a way of seeing human biology that happens to require their gadgets.

The rebrand to Curi Bio signaled the shift - from a surface company to a human-data company. The name is easier to say at a conference, which, in biotech, counts for more than anyone admits.

Grow it, move it, watch it, measure it

Curi Bio's catalog reads like a small aquarium - Mantarray, Stingray, Nautilus - which is fitting for a company obsessed with things that move. Each tool handles one stage of the same loop: build mature human tissue, stimulate it, image it, and turn the wiggle into numbers a drug team can act on.

Hardware, wetware, and software in one loop. Remove any one and you're back to hoping the mouse was right.

Who's buying, and what the money says

A platform is only as credible as the people who pay for it. Curi Bio's reported customer list leans toward the kind of names that have entire buildings devoted to caution - large biopharma R&D teams that adopt new methods slowly and for good reason.

Customers as reported in press coverage of the Series B. Treat as indicative, not exhaustive.

Drug developers are demanding human-relevant functional data earlier in the pipeline.Elliot Fisher, Co-founder & Chief Business Officer

Close the gap, before the clinic

Strip away the instrument names and the mission is plain: give drug developers a more honest preview of how a therapy will behave in a human, while it's still cheap to change course. Curi Bio's focus diseases are the ones where that preview matters most - cardiac indications, muscle metabolism and wasting, and neuromuscular conditions like ALS.

The Series B adds a second front: AI and machine learning layered on top of the functional data, so the signal coming off all those twitching tissues gets read faster and more consistently. Hardware generates the data; the software is increasingly where the argument is won.

This investment enables us to aggressively scale our integrated platforms, providing our biopharma partners with predictive, functional data on human systems across devastating diseases like muscle wasting, ALS, and cardiac indications.Nicholas A. Geisse, CEO

The well that beats on a Tuesday

Regulators are warming to human-relevant methods. Biopharma budgets are tightening. Both trends point the same direction: toward data that predicts human outcomes earlier, and away from expensive surprises in late-stage trials. A company that turns lab-grown human tissue into a reliable measurement sits squarely in that current.

There's still distance to travel. "Human-relevant" is a spectrum, not a finish line, and engineered tissue is not yet a person. But the bet Curi Bio made in 2015 - that the surface, then the tissue, then the data would matter - looks less like a gamble each year.

So return to that plastic well. The heart inside it is still beating, still counted, still human enough to teach us something. A decade ago that was a lab curiosity. Now it's a line item in a drug program - and, increasingly, the reason a bad molecule gets stopped before it ever reaches a patient.

Find Curi Bio