DARPA's Defense Sciences Office Drops Four SBIR XL Topics on June 3: MANTRAS, Engineered Sleep, Air-Independent Power, and Pathogen-Host Interactomes
June 3, 2026 · 7 min read
Arthur Griffin
The Defense Advanced Research Projects Agency's Defense Sciences Office did not bury its June drop. On Wednesday, June 3, DARPA pre-released four FY26 SBIR XL topics from DSO — one in every quadrant of the agency's hard-science portfolio. The proposals window opens June 24, 2026 and closes on July 22, 2026 at noon Eastern. That is seven weeks from drop to deadline, with the actual submission window compressed to under four weeks once the BAA goes live. Small businesses that did not start positioning before today are already behind.
The four topics — MANTRAS for Rydberg-based atomic sensors, Engineering Sleep for Cognitive Performance, Expeditionary Closed and Air-Independent Power and Energy, and Real-Time Pathogen-Host Interactome Prediction — share an unusual signature for a DSO drop. None is purely exploratory. Each is structured around a manufacturable platform, deployable wearable, fieldable battery, or near-real-time prediction system. DSO has historically been the place where exotic physics, mathematics, and biology get an early federal home. This release is different. DARPA is asking small businesses to take ideas that have lived in DSO basic-research programs for several years and turn them into hardware and software that an operator can put in a rucksack, on a wrist, or on a forward operating base.
That shift matters for how applicants should write. SBIR XL is structured as a Direct-to-Phase II vehicle for small businesses with documented Phase I-equivalent feasibility — meaning DARPA expects you to show up with evidence that the underlying technology works, and to spend the proposal narrative explaining how you will scale it. Reviewers in this cohort will be skeptical of teams that lead with a literature survey. They want a path to a deliverable.
For a parallel read on the agency's broader pivot to manufacturable defense biotech, see our earlier analysis of the DARPA Biological Technologies Office FY26 drop and the BTO May 2026 quadruple-topic release.
MANTRAS — Manufacturing Technologies for Rydberg-based Atomic Sensors
Topic number DPA26BZ03-DV011 is the headline. MANTRAS asks small businesses to demonstrate a low-SWaP — size, weight, and power — ruggedized, manufacturable platform for real-time measurement, data acquisition, and analysis of wideband RF signals using Rydberg-based atomic sensors. The word "manufacturable" is doing the work in that objective statement. Rydberg sensors have been the subject of laboratory demonstrations since the early 2010s, and DARPA has funded a sequence of programs in the underlying physics. What it has not yet seen is a vendor who can build them in volume, at field temperatures, in a form factor that an operator can hand-carry without a cryogenic apron.
The strategic context: Rydberg atom-based RF sensing offers a path around the antenna size constraints that have bounded conventional RF receivers since Marconi. A Rydberg sensor uses laser-excited cesium or rubidium atoms in a vapor cell to detect electric fields directly, with a single sensor head that can in principle cover frequencies from hundreds of megahertz to terahertz. For signals intelligence, electronic warfare, and contested spectrum operations, that is a generational change. The catch is that every demonstration to date has involved an optical bench, two or three lasers locked to atomic transitions, and a vapor cell that does not tolerate vibration or thermal swing. MANTRAS is DARPA's bet that the bench can be collapsed onto a circuit board.
Winning proposals should lead with a credible manufacturing roadmap — vapor cell sourcing, laser stabilization at scale, optical integration approaches like photonic integrated circuits, and a path to production yield. Teams that can point to existing commercial vapor cell fabrication, photonic foundry relationships, or laser packaging capacity will outscore teams that propose to build everything from scratch. Phase II budgets on SBIR XL topics typically run in the $1.5 million to $2 million range, with bridge funding pathways into DARPA program-of-record acquisition if the technology matures.
Engineering Sleep for Cognitive Performance
Topic DPA26BZ03-DV012 asks small businesses to develop wearable systems that enhance the restorative functions of sleep. Operationally, the target user is a warfighter on a 96-hour mission who needs to recover four hours of cognitive function from a two-hour rest window. DARPA's bet is that closed-loop wearables — sensing slow-wave activity, autonomic state, and possibly delivering targeted acoustic or electrical stimulation — can compress the time-to-restoration math.
The competitive landscape for this topic is unusual. Commercial sleep-tech vendors have spent a decade selling consumer wearables on the same general premise. DARPA is asking for something the consumer market has not delivered: clinically defensible cognitive recovery metrics tied to closed-loop stimulation, validated in a population that is sleep-restricted by design rather than by disorder. Small businesses with prior clinical-grade EEG hardware, neurofeedback platforms, or DOD-funded human performance work will have an edge. Proposals that read like a Fitbit pitch will not survive review.
Expect a Phase II structured around a wearable prototype, a validation study, and a manufacturing transition plan. The clinical study design is where most teams will lose ground — DSO program managers are sophisticated reviewers of human-subjects research, and a thin protocol kills proposals at this office faster than at any other DARPA tech office.
Expeditionary Closed and Air-Independent Power and Energy
Topic DPA26BZ03-DV013 targets rechargeable high-energy batteries that are independent of external air sources — meaning sealed systems that do not require atmospheric oxygen, can operate underwater or in chemical-biological-radiological environments, and can be recharged in-theater without specialized infrastructure. The energy-density target language has not yet been published, but DARPA's recent history on closed-cycle power suggests it will press for specific energy figures well beyond current lithium-ion ceilings.
This topic sits at the intersection of two technology curves: solid-state battery chemistries that have begun reaching production in automotive applications, and emerging metal-air chemistries that promise step-change density at the cost of air dependence. The "air-independent" language deliberately rules out metal-air designs, pointing the topic toward sealed solid-state, lithium-sulfur, or lithium-metal chemistries with manufacturing pathways. Small businesses with battery cell manufacturing capacity, electrolyte synthesis expertise, or active relationships with DOD energy-storage program offices like ONR Code 331 or the Army's CCDC Ground Vehicle Systems Center will be best positioned.
Winning proposals will need to address recharge architecture as carefully as cell chemistry. A forward operating base does not have the same charging infrastructure as a vehicle bay. Teams that ignore the recharge half of the topic will lose to teams that treat it as a system design problem.
Real-Time Pathogen-Host Interactome Prediction
Topic DPA26BZ03-DV014 asks small businesses to rapidly characterize host-pathogen interactions to enable medical countermeasure development. The interactome framing is the analytical key: DARPA is asking not for sequence analysis or vaccine candidate generation alone, but for predictive models of how a pathogen's proteins, RNAs, and metabolites engage with host cell machinery in real time. The end product is supposed to inform countermeasure design at speeds that match emerging biological threats — meaning the prediction loop should close in hours or days, not weeks.
This topic overlaps with the BTO portfolio more than the typical DSO topic. DSO is asking the question because the answer requires advances in mathematical modeling, foundation models trained on biological data, and computational platforms that integrate structural biology with cell-level signaling — domains where DSO has built up program management depth. Small businesses with AI/ML platforms applied to biology, structural biology pipelines, or experimental host-cell screening capacity will be the strongest competitors. Teams that frame this as a vaccine discovery topic will be misreading it. The deliverable is predictive infrastructure, not a candidate molecule.
What to do this week if you are competing
Seven weeks is enough time to win an SBIR XL topic if your team has been working in the field. It is not enough time to start a new technology effort. The strategic actions for the next seven days:
- Build the topic file. Pull the full pre-release text from the DSIP portal once it lands on June 24, and read it against your existing Phase I-equivalent data. SBIR XL requires a feasibility documentation memo — start drafting it now.
- Lock your government collaborator. All four topics will favor proposals that name a transition partner — an Army, Navy, Air Force, or Combatant Command program office that has agreed to receive the technology if Phase II succeeds. Cold-call the relevant program offices this week. Memorialize commitments in letters before submission.
- Audit your manufacturing story. Three of the four topics (MANTRAS, Engineering Sleep, Expeditionary Power) are explicitly about manufacturable hardware. Reviewers will be looking for production partners, foundry relationships, or in-house fabrication capacity. If you cannot point to one, recruit one.
- Plan for the new uniform guidance. With OMB's October 1, 2026 effective date for the revised 2 CFR Part 200 approaching, structure your indirect cost narrative defensively. Phase II awards issued after October 1 will be reviewed under the new regime — including the political pre-issuance review.
DARPA's June 3 drop is the clearest signal yet that the agency wants its FY26 portfolio anchored in manufacturable, deployable, fieldable technology rather than further laboratory exploration. Small businesses that can write to that signal — with the documentation, partnerships, and production pathways to back it up — will pull the awards. Everyone else is welcome to apply again next quarter.
For a running view of what DARPA is opening across all six tech offices, see the DARPA SBIR/STTR topics index on Granted.
Sources: