Pentagon CTO declares the defense-tech science fair over
Emil Michael's office is concentrating innovation programs around deployment, commercial adoption and measurable military use.
By Ryan Merket · Published
Primary source: X
Why it matters
Defense startups are being told to bring deployable products, production plans and real military customers. A successful Pentagon demo alone will no longer prove a program worked.
The Pentagon's technology office declared that "the peacetime science fair is over" in an X article published Monday, sharpening its demand that military research programs produce technology that can move into operational use.
The August 31st message came from the Office of the Under Secretary of War for Research and Engineering, which serves as the department's chief technology office. Emil Michael, the former Uber executive who leads the office, is the senior official responsible for setting the Pentagon's research, engineering and technology priorities.
The six-word declaration compresses the acquisition strategy Michael's office has been putting in place: centralize control of the Pentagon's scattered innovation organizations, give commercial technology a more direct path into military programs and judge projects by whether they reach operators.
Innovation moves under the CTO
The Pentagon reorganized its innovation structure earlier this year, placing the Defense Innovation Unit, the Strategic Capabilities Office, the Defense Advanced Research Projects Agency, the Chief Digital and Artificial Intelligence Office, the Office of Strategic Capital and the Test Resource Management Center inside an innovation group led by the CTO.
Under that structure, Michael's office sets priorities while the component organizations retain responsibility for execution. Military departments and program offices still control the final step of inserting technology into weapons, software and operational systems.
That division matters for defense startups. The Pentagon has accumulated offices, pilot programs and contracting mechanisms intended to bring commercial technology into the military. Founders can still spend years navigating demonstrations, budget cycles, security reviews and procurement offices before a successful test becomes recurring revenue.
The CTO's "science fair" language sets a harder standard for those programs. A prototype that performs well during a controlled demonstration is only the beginning. The Pentagon wants manufacturing capacity, integration plans, operational testing and a program office capable of buying the product at scale.
The reorganization also gives Michael a wider view of the funding chain. DARPA supports high-risk research, DIU works with commercial suppliers, the Strategic Capabilities Office adapts technology for military missions and the Office of Strategic Capital finances companies working in areas tied to national security. Putting those organizations under a common technology agenda is designed to reduce the gaps between research, contracting and production.
Deployment becomes the metric
Secretary Pete Hegseth previewed the shift during remarks at SpaceX, where he argued that the Pentagon needed to shorten the distance between American technical research and fielded military capability. His formulation contrasted a domestic "peacetime science fair" with competitors pursuing a wartime arms race.
Michael's office has since organized its work around six critical technology areas: applied artificial intelligence, biomanufacturing, contested logistics, quantum and battlefield information dominance, scaled directed energy and scaled hypersonics. The list favors technologies with identifiable military missions and production requirements.
Artificial intelligence is the clearest example. The Pentagon's AI acceleration strategy established seven pace-setting projects spanning warfighting, intelligence and enterprise operations. The department has also deployed GenAI.mil as a common generative AI platform for its workforce, part of an attempt to move AI use beyond isolated experiments and into daily operations.
The department says GenAI.mil reached 1 million unique users shortly after deployment. That figure is self-reported and does not show frequency of use, workload value or whether adoption has changed operational outcomes. It does show how the CTO's office intends to measure progress: named systems, counted users and defined deployment targets rather than a portfolio of disconnected pilots.
The opening for startups
The policy creates an opening for defense-tech founders who can arrive with products that already work, suppliers that can scale and financing that can survive the Pentagon's purchasing cycle. It also raises the burden of proof. Technical novelty alone will carry less weight when contracting officials are under pressure to show deployment, production and military utility.
The Pentagon's own structure remains a constraint. The under secretary can set technical direction and consolidate innovation offices, while appropriations, military-service requirements and program managers determine what gets bought. Startups will still need an internal customer with a budget and authority to integrate their product.
Michael's bet is that central oversight can force those pieces to line up earlier. The August 31st declaration puts the Pentagon's research establishment on notice: a demonstration without a path to procurement no longer counts as the finish line.