ARPA-H's $400 Billion Sleep Problem: Inside the REST Program and Its August 12 Solution Summary Deadline

Sleep is the rare health problem where almost everyone is a patient and almost no one is being effectively treated. ARPA-H puts hard numbers on it: roughly 150 million Americans struggle with poor sleep, an estimated 86 million adults experience insomnia, and current treatments fail for more than 70% of patients. Stack on the downstream links to depression, dementia, cardiovascular disease, and diabetes, and the agency arrives at an eye-widening figure — an estimated $400 billion annual economic burden.

The agency's response is REST — Restorative and health-Enhancing Sleep Time (solicitation ARPA-H-SOL-26-159), run out of ARPA-H's Proactive Health Office. It is a 66-month, three-phase program funded through Other Transaction (OT) agreements, ARPA-H's flexible non-FAR mechanism. And it has a near-term gate that matters more than its long-term ambition: a mandatory Solution Summary due August 12, 2026 at 12:00 PM ET, preceded by a Proposers' Day on July 13, 2026. Full proposals — in Fall 2026 — are invitation-only, meaning the solution summary is not a formality. It is the filter that decides who competes at all.

The core reframe: sleep quality, not sleep quantity

The intellectual premise of REST is a critique of the entire consumer-sleep-tech category. Today's wearables and apps are very good at telling you how long you slept and which stages you cycled through. What none of them can tell you is whether that sleep actually did you any good — whether it was restorative in a way that predicts long-term health outcomes.

REST wants to close that gap. ARPA-H frames the first technical goal as defining sleep quality by identifying the physiological features that predict health outcomes, not just the ones that are easy to measure. That is a genuinely open scientific question, and it is the conceptual spine of the whole program: if you cannot define restorative sleep in measurable terms, you cannot diagnose its absence at home or design a device to fix it.

Everything else in the solicitation flows from that reframe.

Two technical areas, one closed loop

REST is organized into two technical areas that are meant to connect end to end.

TA1 — Measure and Diagnose. The goal is consumer-grade systems that measure sleep physiology with clinical accuracy in the home. ARPA-H explicitly floats wearable EEG, smartwatches, rings, and under-mattress sensors capable of resolving sleep microstructure and classifying insomnia subtypes — not just a nightly "sleep score," but clinical-grade phenotyping in a bedroom rather than a lab. This is the diagnostic half: get accurate, health-predictive measurement out of the sleep clinic and into everyday devices.

TA2 — Control and Treat. The goal is noninvasive, closed-loop intervention systems that adapt treatment in real time overnight — through neuromodulation, acoustic stimulation, haptics, or targeted nerve stimulation. "Closed-loop" is the operative phrase: the system senses the sleeper's physiology and adjusts its intervention on the fly, personalized to the individual, rather than delivering a fixed program.

Put together, TA1 and TA2 describe a single ambition — a device (or family of devices) that can measure whether your sleep is restorative and intervene to make it more so, all noninvasively and all at home. That is what ARPA-H means by moving sleep out of the lab.

Why this is an interdisciplinary teaming problem

REST is not a program a single lab wins. ARPA-H is explicit that it wants bold, interdisciplinary approaches spanning sleep medicine, neuroscience, psychiatry, biomedical engineering, wearable sensors, signal processing, machine learning, clinical-trial expertise, and — critically for a device that will eventually seek market entry — regulatory science. The agency notes that teaming is anticipated as necessary.

That range is not window dressing; it maps directly onto the two technical areas. TA1 needs sensor hardware, signal processing, and the clinical sleep expertise to know which physiological features actually matter. TA2 needs neuromodulation and stimulation engineering plus the clinical-trial machinery to prove an intervention works. And the whole thing needs a machine-learning layer to turn raw physiology into the health-predictive definition of sleep quality that the program is built around.

The practical implication: a competitive REST team probably braids together a device company, an academic sleep-medicine group, an ML capability, and regulatory expertise. Assembling that consortium takes time — which is exactly why the August 12 solution summary deadline is tighter than it looks. Teams that start building the consortium after Proposers' Day are already behind teams that walked in with one.

The Other Transaction mechanism changes the game

REST's use of Other Transaction authority is not a footnote. OT agreements sit outside the Federal Acquisition Regulation and the standard grant rules, which gives ARPA-H flexibility on intellectual property, milestone structure, and non-traditional performers. For companies that have historically avoided federal funding because they feared losing control of their IP or drowning in FAR compliance, OT is the mechanism designed to lower that barrier. We have written before about how ARPA-H's OT-based, multi-phase model works in practice — for example in the BOSS biostabilization program — and the same logic applies here.

The flip side is that OT programs are aggressively milestone-driven. A 66-month, three-phase structure means funding is gated on hitting technical targets, and teams that miss milestones can be cut. Proposers should read the phase structure as a set of go/no-go decision points and build a technical plan that front-loads the riskiest scientific questions — chiefly, can you actually define restorative sleep in measurable terms? — into the early phases where they can be de-risked before the big money flows.

Treat the solution summary as the real competition

The single most important strategic fact about REST is procedural: full proposals are invitation-only, and the invitation depends entirely on the solution summary. A brilliant full-proposal team that writes a mediocre solution summary never gets to submit the full proposal. So the summary is not a screening courtesy — it is the competition.

For a mandatory, short-form solution summary, three things tend to separate advancers from the rejected pile:

• Lead with the health-outcome thesis. REST is built on the claim that sleep quality predicts health outcomes better than sleep quantity. A summary that states, up front, a specific and testable definition of restorative sleep — and how the team will validate it — is speaking the program's language. A summary that leads with hardware specs is not.
• Show the closed loop end to end. If you are proposing in TA1, make clear how your measurement feeds a downstream intervention. If TA2, make clear what signal you are closing the loop on. ARPA-H is buying an integrated system, and summaries that demonstrate they understand the whole loop read as more credible than point solutions.
• Prove the team already exists. Because teaming is "anticipated as necessary," a summary that names the sleep-medicine, engineering, ML, and regulatory partners — and shows they are already committed — signals implementation readiness that a solo applicant cannot match.

Where REST sits in the 2026 ARPA-H portfolio

REST is one entry in an unusually active ARPA-H year. It joins programs like PROSPR on aging and healthspan and the agency's broader push into proactive, at-home health monitoring. For small companies, it is worth reading REST alongside the ARPA-H SBIR/STTR solicitation, which offers a lower-barrier, non-dilutive path (up to $600K Phase I, $3.5M Phase II) for teams that are not yet ready to lead a 66-month OT consortium. A sensor startup with a novel at-home sleep measurement, for instance, might pursue SBIR to mature the core technology while positioning to join a larger REST team as a component provider.

The deadline discipline is unforgiving. Proposers' Day is July 13; the solution summary is due August 12; and if you are not invited, Fall 2026's full proposal window simply does not open for you. For a health-tech team that has been building toward at-home sleep diagnostics or noninvasive neuromodulation, the next five weeks — not the next five years — are where REST is won or lost.

Tracking ARPA-H and federal health-research funding? Search Granted for live opportunities, funder profiles, and past-recipient data to sharpen your next proposal.