Five of Seven Topics Came From One Office: Inside DARPA's Release 6 SBIR Drop and the $7M Battlefield-Medicine Stack Closing October 21

Count the offices on DARPA's Release 6 topic list and the story writes itself. Seven SBIR topics went to pre-release on September 2, 2026. One came from the Tactical Technology Office. Five came from the Biological Technologies Office. And four of those five — noninvasive hemorrhage detection, casualty-stream prediction, autonomous triage under chemical contamination, and a portable extracorporeal life support device — are not four separate gadgets. Read in order, they are a single clinical timeline for a wounded person who is not going to see a hospital.

That is the most useful thing about this release, and it is invisible if you evaluate the topics one at a time.

The calendar is tighter than it looks

The Department of War pre-releases SBIR and STTR topics on the first Wednesday of every month, and Release 6 landed on schedule. Topics opened for pre-release September 2, 2026, open for submission September 23, 2026, and close October 21, 2026 through the Defense SBIR/STTR Innovation Portal at dodsbirsttr.mil. Technical questions to SBIR_BAA@darpa.mil close October 14, 2026. Selection notifications follow within 90 days of BAA close.

One caution worth acting on: DARPA's own topic index page currently lists an October 23 close for Release 6, while the individual topic write-ups say October 21. Treat October 21 as operative, confirm the date on DSIP the day you start your cost volume, and do not build a submission plan that depends on a two-day cushion you may not have. DARPA does not accept late submissions, and it says so in every topic.

The real deadline is earlier than either date anyway. Once the technical-questions window shuts on October 14, you are writing blind against whatever ambiguities remain in the topic text. On this release there is at least one ambiguity that matters — more on that below.

The monthly cadence also changes the strategic math. When DoD issued two or three SBIR solicitations a year, missing a cycle cost you months. Now a missed release costs you roughly four weeks. That is an argument for not forcing a thin proposal into Release 6 if your evidence package is not ready, and it is a reason to keep watching the first Wednesday of every month rather than treating each drop as a one-time event. Our SBIR and STTR deadline calendar tracks the cadence across agencies.

The four-topic clinical sequence

Start with the money, because the award structures tell you how confident DARPA is in each piece.

DPA26BZ06-DV025, Noninvasive Detection and Localization of Occult Hemorrhage. Phase I is $250,000 over 6 months. Direct to Phase II is $1,000,000 over 12 months plus a $500,000 12-month option. DARPA wants a device that lets a combat medic with no surgical or imaging training find internal, non-compressible torso bleeding — detect it, localize it anatomically, and estimate whether it is trickling or expanding fast. The topic explicitly names retroperitoneal hemorrhage and bleeding obscured by bone or bowel as challenges your proposal must confront rather than route around.

DPA26BZ06-DV027, Casualty Operations & Resource Prediction Software (CORPS). Direct to Phase II only; no Phase I. This is the topic where the paperwork disagrees with itself. The index states $750,000 over 12 months with no options; the narrative describes an 18-month base plus a 6-month option with milestones running to Month 24. That is not a detail to resolve in your head. Send it to SBIR_BAA@darpa.mil before October 14 and get the answer in writing, because the period of performance drives your labor mix, your indirect application, and whether your prototype schedule is credible.

CORPS asks for software that predicts casualty and patient streams after a large-scale medical emergency in an austere environment, then projects the lifesaving value of different response options — with limited or no real-time data and no assumption of connectivity. It must degrade gracefully across full, intermittent, and zero communications, and interoperate with both DoD planning systems and civilian emergency infrastructure. Deliverables include a minimum of four prototypes evaluated by different military users, with dynamic uncertainty quantification demonstrated by Prototype 2.

DPA26BZ06-DV028, TRIAGE-X. $1,500,000 over 24 months with a $500,000 6-month option — $2,000,000 across 30 months if fully exercised. The scenario is a mass casualty event inside a chemically contaminated area, where responders cannot enter until the agent is characterized and cannot enter quickly once they must don protective equipment first. DARPA wants an autonomous system that recognizes chemical toxidromes and physically delivers the corresponding antidote or intervention. Note the verb. Recognition alone does not satisfy this topic.

DPA26BZ06-DV029, ICU-in-a-Box. $1,500,000 over 24 months with a $500,000 6-month option, $2,000,000 total. An autonomous, portable extracorporeal life support platform for prolonged field care, integrating resuscitation with cardiovascular and pulmonary support.

Line them up and the sequence is: find the bleed, decide who to treat and in what order, treat under contamination, then sustain the patient long enough for evacuation that may not come for days. DARPA is not shopping for four products. It is assembling an autonomy stack for the scenario military medical planners have been describing for several years — a Pacific-distance conflict where the golden hour is a fiction and air evacuation is contested.

The gates that will disqualify most bidders

Direct-to-Phase-II is where this release gets selective, and each of these four topics draws the line in a different place.

DV029 is the harshest. To bid, you must document a working prototype hitting ≥2 L/min blood flow with oxygenation and ventilation, ≥75 mL/min oxygen delivery, ≥40 mL/min CO₂ exchange, plus at least one functioning closed-loop algorithm — oxygenation and ventilation control, mean arterial pressure maintenance, or electrolyte removal. Then comes the constraint that eliminates most of the field: DARPA explicitly excludes "system of systems" solutions that integrate multiple separately controlled devices. It wants a single device delivering all interventions through a single intravenous cannula, preferably 15Fr or smaller. If your architecture is a pump plus an oxygenator plus a controller plus a monitor in a shared case, you are not responsive, and no amount of proposal writing fixes that.

DV029 also puts regulatory approvals on the critical path in a way proposers routinely underestimate. Large animal validation dominates the schedule, with IACUC and ACURO approvals explicitly gating progress at Months 9 and 21. If you do not already have a large-animal facility partnership with an established regulatory pathway, your schedule is not credible and reviewers will know it.

DV025's Direct-to-Phase-II entry requires four documented feasibility elements: proof of concept, demonstrated hemorrhage rate estimation, validated performance metrics, and preliminary operational robustness data. Rate estimation is the one most teams cannot show. Plenty of groups can detect free fluid. Far fewer can defend a number for how fast it is accumulating.

CORPS is the outlier, and it is the opening. Its feasibility bar is notably softer than DARPA's norm — the topic accepts "conceptual characterization, methodology, design specifications, and preliminary data" rather than a fielded system. That language admits operations-research shops, health-systems modeling groups, and simulation teams with no hardware and no prior defense medical work. If you have built patient-flow or disaster-resource models for a civilian health system and have never bid a DARPA topic, DV027 is the most realistic entry point on this release.

What to do between now and October 14

Three weeks of pre-release exist so you can make a bid/no-bid decision before you burn proposal money, not so you can start writing early.

Use them to answer the disqualifying questions first. For DV029: is your device architecturally single, and can you produce animal data at the stated flow and gas-exchange thresholds? For DV025: can you put a defensible number on bleed rate, or only on presence? For DV028: does your system deliver the countermeasure, or only identify the toxidrome? For DV027: what is the actual period of performance — and have you asked?

Then decide your contract type before drafting the cost volume, because it constrains everything downstream. Firm-fixed-price is the low-friction path; cost-plus reimbursement requires a DCMA Final Determination Letter you cannot obtain in six weeks if you do not already have one; Other Transaction for Prototype requires the Model OT template and associated certifications. Teams discover this in the last week and lose days they needed for technical content.

And budget the Technical and Business Assistance separately — $6,500 for Phase I and $25,000 for Phase II, additional to the cost ceilings rather than carved out of them. It is free money that a surprising number of first-time proposers leave unclaimed.

Release 7 pre-releases on the first Wednesday of October. If your evidence package is thin for these topics, that is a feature of the new cadence, not a consolation prize — but only if you spend the next four weeks generating the data the gate demands rather than polishing prose for a topic you cannot clear. Matching your capability against the right solicitation before you commit to a proposal is exactly the problem Granted was built to shorten.