For sixty years, getting a drug approved in America has followed the same grinding playbook: run a trial, collect data, package millions of pages of submissions, send it to the FDA, and wait. The average journey from Phase 1 to market takes 10 to 12 years. Nearly half of that, according to FDA Commissioner Marty Makary, is “dead time” — paperwork, administrative lag, and data sitting in limbo.
On Tuesday, the FDA said it’s done with dead time.
The agency unveiled its first-ever real-time clinical trial initiative — a system that uses AI and cloud computing to let FDA scientists monitor trial data as it happens, not months or years after the fact. Two cancer drug trials are already live. A broader pilot launches this summer.
This isn’t incremental. If it works, it rewrites how drugs reach patients.
What “Real-Time” Actually Means Here
Traditional clinical trials work like a relay race where every handoff takes months. Data flows from trial sites to the pharma company, which analyzes it, compiles regulatory submissions, and eventually hands it to the FDA. Each transition is a bottleneck.
The new system gives the FDA a direct data feed via a cloud platform built by health tech firm Paradigm Health. When a patient develops a fever, when a tumor shrinks, when an adverse event fires — regulators see it through predefined clinical endpoints, in real time.
“When a patient develops a fever, or a tumor shrinks, FDA regulators can see in the cloud, in real-time, exactly what is happening,” Makary told reporters at FDA headquarters in Silver Spring.
Important nuance: the FDA isn’t accessing raw patient records. Chief AI Officer Jeremy Walsh emphasized they receive only aggregated signals — adverse event rates, tumor response percentages, safety endpoints. Patient data stays with the trial sponsor. No privacy minefield.
Two Trials Are Already Running
This isn’t theoretical. Two major pharma companies are feeding data to the FDA right now:
AstraZeneca’s TRAVERSE is a Phase 2 multi-site trial testing combination therapy for treatment-naïve mantle cell lymphoma — an aggressive blood cancer. It’s running at MD Anderson and the University of Pennsylvania. The FDA has already received and validated real-time signals, confirming the technical framework works.
Amgen’s STREAM-SCLC is a Phase 1b trial for limited-stage small cell lung carcinoma. Site selection is ongoing, but the infrastructure is live.
The 40% Number
The FDA estimates this could cut 40% off clinical trial timelines — not by lowering safety standards, but by eliminating the dead zones between trial phases.
Here’s the math that matters: most drug development happens in discrete phases with gaps between each. After Phase 1 wraps, there’s a lag before Phase 2. Another gap before Phase 3. Each involves data compilation, regulatory review, and protocol development. Real-time monitoring could compress or eliminate these gaps entirely, enabling what the FDA calls “continuous trials.”
If the average drug takes 10-12 years and 45% of that is administrative lag, we’re talking about saving 4-5 years per drug. In oncology, where patients die waiting for treatments to clear regulatory hurdles, that’s not a statistic. It’s the difference between life and death.
Why Now? China Is Eating America’s Lunch
Makary didn’t hide the geopolitical motivation. China surpassed the U.S. in Phase 1 clinical trials around 2021, with “exponential” growth since. Faster enrollment, lower costs, a massive patient population — Beijing has spent decades and billions of yuan becoming a clinical trials superpower.
If the U.S. approval process stays stuck in a 1960s workflow while China modernizes, the center of pharmaceutical innovation shifts eastward. The real-time trial initiative is partly the FDA’s competitive answer — maintaining America’s edge without sacrificing the safety standards that make FDA approval the global gold standard.
The Legitimate Concerns
Speed-versus-rigor is a real tension, and critics aren’t wrong to flag it.
Bentley University professor Fred Ledley has pointed out that pressure to act on incomplete information grows when regulators can watch data stream in. There’s also a thorny technical question: when AI systems update themselves during trials, do sponsors disclose every model retrain (risking regulatory delays) or treat updates as routine maintenance (risking data integrity challenges)?
The regulatory framework for AI-in-the-loop trials simply doesn’t exist yet.
Clinical researchers have also noted that individual trial results don’t tell the full story. Real-time data on a single study shouldn’t be the sole basis for regulatory decisions — the risk/benefit profile requires the full picture across an entire development program.
Walsh’s response was direct: “The goal here is to get to a regulatory decision in a faster timeline, without compromising any safety. The goal here is to raise the bar for what can be done.”
The Quiet Infrastructure Overhaul Behind the Headlines
The real-time trial announcement sits on top of a much larger — and underreported — tech modernization at the FDA.
Over the past year, the agency consolidated 40 separate application intake systems into one. It merged three data monitoring systems and seven adverse event reporting systems into single platforms. It eliminated duplicate software licenses across its centers. All without additional staff — in fact, after significant downsizing.
Makary claims these efforts save at least $120 million annually. Whether that number holds up, the direction is clear: the FDA is transforming from a paper-shuffling bureaucracy into a data-driven regulatory agency. Real-time trials are the flagship, but the engine room got rebuilt too.
The agency has published a Request for Information with comments due May 29. Pilot selections should be complete by August.
What This Actually Changes
The FDA’s move matters beyond just timelines. It signals that the agency is ready to treat AI as core infrastructure rather than a novelty.
If the pilot succeeds, the cascade effects are significant. Smaller biotech companies — the ones most crushed by the time and cost of traditional trials — could benefit enormously. Insurance companies could factor faster approvals into coverage models. And patients with serious diseases could see promising treatments years sooner.
The risk is real: the FDA’s track record with accelerated approvals is mixed. Some were life-saving. Others were embarrassments. Real-time trials need to prove they can be both fast and thorough.
But here’s what’s undeniable: a process designed in the 1960s is not adequate for a world where AI can analyze clinical endpoints in real time. The FDA is finally acknowledging that — and backing it up with two live trials and a summer pilot.
The question isn’t whether this is the right direction. It’s whether they can execute without stumbling.