In a strip-mall oncology clinic somewhere off a U.S. interstate, a 64-year-old woman is being told her lung cancer has progressed. The doctor opens her chart. The chart - the boring software panel that runs most of American medicine - knows what trials she might qualify for. It knows what her last scan showed, what mutation she carries, what insurance she has. It surfaces three options before the visit ends. None of this used to happen. N-Power Medicine is the reason it does.
That, in one sentence, is the bet. Build a research layer that sits on top of community oncology - the unglamorous places where most cancer patients actually get treated - and let trials come to the patient instead of the other way around. It is not a moonshot. It is plumbing. The fact that nobody had built it is the scandal.
Most patients with cancer in this country are never offered a clinical trial. That is a problem you can fix with software, sweat and a national network.- N-Power Medicine, paraphrased from public statements
The Problem They Saw
Oncology drug development is a slow, expensive ritual built around the academic medical center. Roughly 4% of adult cancer patients in the United States ever enroll in a trial. The other 96% are not unwilling. They are unreachable. They live too far from the major centers, work jobs that cannot accommodate a four-hour drive, or simply never hear that a trial exists. The system has decided this is acceptable. N-Power Medicine has decided it is not.
The cost of this acceptance is measured in years. Trials that should take 18 months take 36. Drugs that could change a treatment guideline sit in queues. Control arms - the patients randomized to standard-of-care so the experimental arm has something to compare against - are themselves a bottleneck, and an ethically uncomfortable one for diseases where the standard is already failing.
There is, of course, a perfectly comforting industry explanation: this is just how it has always worked. Which is exactly the kind of answer a company gets started to disprove.
The Founders' Bet
Mark Lee, MD, PhD, ran Personalized Healthcare for Product Development at Genentech before this. Before that, he helped stand up the founding clinical program at GRAIL, the multi-cancer early detection company. His co-founder, Christer Svedman, MD, PhD, comes from the same world of large-scale clinical science. Neither needed a second act. They started one anyway.
Their bet was specific: that the bottleneck in oncology trials was not science, but workflow. That the community oncology practice already had everything you needed - patients, oncologists, electronic health records, an existing relationship of trust - and was missing only the operational scaffolding to run a trial without exploding the schedule of the local nurse manager. Build that scaffolding once. Run it everywhere. The drug industry will pay you for it because they have no other path to the patients they need.
The community oncologist is not an obstacle to drug development. The community oncologist is the missing distribution layer.- The thesis, more or less
The Product
N-Power Medicine's first concrete artifact is the Kaleido Registry, launched in 2022. It is a real-time, prospective registry that ingests structured oncology data from partner community practices. Read that sentence twice, because the word "prospective" is doing a lot of work. Most registries are rearview mirrors. Kaleido is a windshield.
On top of Kaleido sits an operational stack: AI-enabled tools to flag trial-eligible patients in something close to real time; a virtual, remote site-support staff that runs the trial work the local clinic cannot absorb; and an integration layer that lives inside the EHRs clinicians already use, so the workflow does not require them to learn another piece of software for the privilege of helping cure cancer.
Then there is the new piece, announced in August 2025: the industry's first prospective external control arm platform. Translation - N-Power can now build a real-time, contemporaneous cohort of standard-of-care patients to serve as the comparator in a clinical trial. Done well, this cuts the size of randomized trials, gets more patients onto the experimental arm, and shortens the path to a readout. Done poorly, regulators will eat you alive. The company is betting it can do it well.
Milestones, briefly
The Proof
A clinical-trial company has two ways to be taken seriously: regulators trust your data, and sponsors keep writing checks. N-Power has been working on both, quietly, by accumulating the unsexy assets.
In July 2024, Merck's Global Health Innovation Fund led the company's $72M Series B and Merck signed a strategic collaboration to expand oncologist and patient access to its trials. Pharma partnerships of that shape are not Twitter announcements. They are operating commitments that take quarters to negotiate. In January 2025, N-Power acquired Syapse, an established real-world oncology data company - a move that bolted years of network and data infrastructure onto a four-year-old startup.
The math, on one chart
A trial that finishes 12 months early is not a press release. It is a patient who got a drug while it could still help.- The argument for speed, in plain English
The Mission
Officially, N-Power Medicine wants to accelerate drug development by reinventing how clinical trials are run. Unofficially, the mission is more pointed: stop wasting the patient. The patient is the rate-limiting resource in oncology drug development, and the system has been treating them like an afterthought.
There is an ironic elegance to where the company chose to fight. It did not pitch a smarter molecule. It did not propose an AI that hallucinates drug candidates. It picked the part of the system everyone agrees is broken and that nobody finds interesting enough to fix: the operations of running a trial in the place where patients actually are. Unfashionable. Hard. Necessary.
Why It Matters Tomorrow
If N-Power's model holds, the second-order effects are large. Prospective external controls would let sponsors run smaller, faster trials in rare cancer subtypes that today cannot recruit a randomized cohort at all. Community-based enrollment would shift the demographic mix of trial participants closer to the population that actually gets the disease - older, more rural, more racially diverse than the current trial average. Drug labels would arrive earlier and with evidence drawn from real practice.
It also raises a quieter question for the rest of healthcare. If you can build an always-on research system inside community oncology, you can probably build one inside community cardiology, community neurology, community anything. N-Power Medicine is, in a sense, a proof of concept the industry has been waiting on for twenty years.
Back in that strip-mall clinic, the 64-year-old patient leaves with a printed page. There are three trials on it. One is across town. Two enroll patients right where she sits. She did not Google for them. Her oncologist did not call seven coordinators to find them. The chart already knew. That small, dull miracle is the entire point - and it is the kind of miracle that, once it exists, becomes very hard to remember life without.