ARPA-H Announced $100M for Four Clinical Trial Programs. Three Were Already Awarded to Two Companies Before the Press Release.

Read the ARPA-H clinical trial announcement the way a grant seeker reads it — looking for the door — and the arithmetic turns hostile fast. The package is worth more than $100 million across four programs. Three of those four arrived with a prime contractor already named, a dollar ceiling already negotiated, and a period of performance that had already begun in September. The fourth, SURPASS, is the only one you can apply to.

That is not a complaint about ARPA-H's authorities. It is the shape of the opportunity, and it determines whether your next three weeks are spent writing a solution summary or writing a teaming email.

The $91.08 Million That Went to One Company

Evidence Health Inc. is the prime awardee on two of the three pre-awarded programs, with the University of Texas at Austin's Dell Medical School collaborating on both.

STACK — Standardized Trial Activation through Composable Knowledge — carries a ceiling of up to $41.18 million. Its job is to use AI to compress clinical site activation and convert research-naïve sites into functioning trial sites, expanding domestic trial capacity so patients can enroll closer to where they live.

COMMONS carries a ceiling of up to $49.9 million. It is the data layer: a nationwide infrastructure for regulatory-grade clinical data, built around what ARPA-H describes as privacy-by-design architecture for consent at national scale. The stated target is reach across more than 40 million individuals and 3 million linked records.

Those two ceilings total $91.08 million to a single prime.

CINCH — Cancer Intelligence for Navigation, Care and Health — went to San Diego-based Courage Health Inc. at up to $8.95 million. CINCH is the patient-facing end: a path for patients to contribute real-world data and surface trials they qualify for, aimed at rural and underserved populations who live far from oncology expertise.

Start dates are where the timeline tells on itself. Reported accounts place the program starts in September 2026 — one puts STACK's start on September 2, another places all four efforts between September 21 and 29. Either way, the work was underway before the first week of October, when HHS put the package in front of the public alongside the SURPASS solicitation. No competed solicitation for STACK, COMMONS, or CINCH appears on ARPA-H's open funding page.

One more number is missing from the whole announcement: SURPASS's own budget was not disclosed. The "more than $100 million" figure attached to the package is carried almost entirely by the three awards that are already spoken for.

What SURPASS Actually Wants, and When

SURPASS — Simulation-augmented, Real-time Platform Adaptive Seamless Trials — went live September 30, 2026. It is the competed program, and it is genuinely ambitious: ARPA-H is asking for cross-disciplinary teams from statistics, AI, and clinical trial design to collapse the phase structure of drug development.

Three technical areas, and the solicitation treats them as one system rather than a menu:

1. A phaseless design engine using digital twins and predictive models to simulate trial outcomes before a trial launches, with the goal of reaching an answer using fewer patients.
2. A continuous inference engine that analyzes accumulating trial data in real time and supports rapid adaptation without surrendering statistical rigor.
3. An agentic operations layer that automates trial startup activities and data collection.

Program manager Daria Fedyukina has framed the technical core plainly: SURPASS is developing the technology and biostatistical frameworks needed to analyze trial data as it accumulates. HHS Secretary Robert F. Kennedy Jr. framed the policy goal as building a modern clinical trial system that generates rigorous evidence faster while reducing patient burden.

The calendar is tight and front-loaded:

• October 12, 2026 — registration closes for the informational webinar
• October 15, 2026 — informational webinar
• October 28, 2026 — registration closes for Proposers' Day
• November 6, 2026 — Proposers' Day
• November 30, 2026 — solution summaries due
• January 22, 2027 — full proposals due

If you are reading this on publication day, the webinar registration deadline is four days out and the solution summary deadline is 53 days out. The ordering matters: ARPA-H uses the solution summary as a gate, and teams that skip Proposers' Day forfeit the only structured chance to find the statistician, the trial operations partner, or the sponsor they are missing. Our deeper treatment of the solicitation's structure — including the mandatory-bundle problem across the three technical areas — is in the SURPASS solicitation analysis.

Why Three Programs Skipped the Competition

ARPA-H is not NIH. It operates with other-transaction authority and standing, rolling solicitations rather than fixed-deadline program announcements for much of its portfolio. Its four mission offices each publish an open Innovative Solutions Opening that accepts proposals on a rolling basis: Health Science Futures (ARPA-H-SOL-24-104), Proactive Health (ARPA-H-SOL-24-106), Resilient Systems (ARPA-H-SOL-24-103), and Scalable Solutions (ARPA-H-SOL-24-105).

That architecture means an award can be negotiated and started without ever appearing as a discrete, competed program solicitation — and then be announced later as a named program. For a funder trying to move at the speed of a technology window, that is a feature. For an applicant watching the funding page for a door, it means the door opened and closed somewhere you were not looking.

Two practical consequences follow.

First, the "up to" in those ceilings is load-bearing. ARPA-H structures milestone-gated agreements; $49.9 million is an authorization, not a deposit. Programs get descoped, and the gap between ceiling and obligated dollars on these efforts is where subcontract capacity will actually be decided over the next 18 months.

Second, the rolling ISOs are the only self-service route into ARPA-H outside a named program competition. If your idea is adjacent to COMMONS or STACK but not inside them, the ISO for the relevant mission office — Scalable Solutions for infrastructure and data-layer work, Resilient Systems for health system capacity — is where it goes, and it does not wait for a deadline.

The Realistic Entry Points

For most organizations reading this, the honest strategy has three lanes, and only one of them is a proposal.

Compete for SURPASS. This is the open lane, and it rewards an unusual team shape. ARPA-H asked for statistics plus AI plus trial design in one proposal. Pure AI shops and pure academic biostatistics groups are both incomplete here. Clinical research organizations with adaptive-platform experience, academic statistical coordinating centers, and companies building trial-operations automation each hold roughly a third of what the solicitation wants — which is precisely why Proposers' Day is worth the travel.

Subcontract into the awarded three. STACK needs sites. Its entire premise is converting research-naïve sites into trial-capable ones, which means community health systems, FQHCs, rural hospital networks, and regional practice groups are the raw material of the program, not spectators to it. COMMONS needs data contributors and consent infrastructure partners to reach 40 million individuals — health systems, EHR vendors, registries, and IRBs. CINCH needs oncology navigation and rural patient-reach partners. Approach Evidence Health, Dell Medical School, and Courage Health directly, as a capability, with your covered-lives count and your data governance posture in the first paragraph.

Position for the policy wave behind it. These four programs sit inside Operation TrialBlazer, the department-wide clinical research initiative HHS launched June 22, 2026 across FDA, NIH, and ARPA-H. The FDA side includes an Expedited Investigational New Drug pilot that moved through a Federal Register request for information with comments due August 24, 2026, and a real-time monitoring pilot from April 2026 that HHS has associated with trial timeline reductions of up to 40 percent. Nothing in that stack is finished. More solicitations will come out of it, and they will favor organizations that already have the site network, the consent architecture, or the statistical method documented and operating.

The uncomfortable read on this announcement is that federal AI-for-health infrastructure is consolidating into a small number of named primes faster than competed solicitations can keep up. The useful read is that consolidation creates subcontract demand — $100 million of it, already obligated, already needing partners — while the one competed program on the page still has its solution summary window open.

Granted's ARPA-H funding tracker and our deadline tools keep the open ARPA-H solicitations and their summary gates in one place, which is the difference between finding a window and reading about it after it shut.