DIU names spectrum challenge winners, but regulatory authority remains unchanged

Airbase and Xtremis.AI won Project Spectrum Strike after a Fort Carson demonstration. The contest tested software to speed coordination; it did not transfer federal spectrum-assignment or equipment-certification authority.

The Defense Innovation Unit said Tuesday that it selected Airbase and Xtremis.AI as winners of Project Spectrum Strike, a prize challenge seeking software that can automate parts of the federal spectrum-coordination workflow. In its announcement, DIU said the companies demonstrated their systems at Fort Carson, Colorado, alongside the second Drone Dominance Gauntlet.

The result is a meaningful signal about the Pentagon’s interest in faster radio-frequency coordination. It is not, on the evidence publicly available, a disclosed production contract or a change in who has legal authority to assign federal frequencies and certify spectrum-dependent equipment. DIU said that Other Transaction agreements will scale the effort; it did not identify an agreement number, award value, recipient or scope in its announcement.

That distinction matters. Project Spectrum Strike’s public solicitation description sought deployable, technology-readiness-level-7-or-higher software for automated deconfliction and authorization. It described a $2 million, multi-round prize challenge and a possible transition to procurement through Other Transaction agreements or conventional contracts. The description set a target of cutting an asserted 75-to-180-day process to fewer than five days. Those are challenge requirements and objectives, not independently published evidence that every underlying certification, assignment or regulatory review can be completed on that timetable.

A Fort Carson test, then a selection

A July 15 Drone Dominance Program announcement said DIU had advanced 12 companies from 43 submissions to the second round. It said the final round would test agentic-AI routing tools and machine-to-machine application programming interfaces during Drone Dominance Gauntlet 2 at Fort Carson. DIU now says the demonstration took place, with spectrum requests generated, validated, modeled and routed in minutes in a real electromagnetic environment.

The agency’s post also said the National Telecommunications and Information Administration, Federal Communications Commission, Federal Aviation Administration, Department of Homeland Security, Defense Information Systems Agency and Army personnel participated in the evaluation. The Offset has treated those statements as DIU’s account of the event; it did not locate a separate joint release reporting the results.

The adjacent Drone Dominance program is a distinct procurement initiative. Its program overview says it is sponsored by the Office of the Secretary of War, administered by the Test Resource Management Center and DIU, and uses Gauntlet results to inform prototype delivery orders issued by Naval Surface Warfare Center Crane. That context explains the emphasis on counter-unmanned aircraft systems in DIU’s post, but it does not establish that the Spectrum Strike winners received such orders.

Faster routing is not the same as an authorization

The public case for automation is straightforward: a radio, radar, drone or electronic-warfare system can be delayed by manual coordination across organizations and databases. But federal spectrum management is more than a form-routing problem.

NTIA, which manages the federal government’s spectrum use, sets policy and regulations for federal use, assigns frequencies, maintains spectrum-use databases and reviews new federal telecommunications systems to certify that spectrum will be available. The Government Accountability Office similarly describes a federal frequency assignment from NTIA as the authorization analogous to an FCC license for a nonfederal user. It notes that NTIA and the FCC jointly determine allocations for federal, nonfederal and shared use.

Those authorities mean software can improve intake, data validation, engineering models and interagency coordination without itself replacing the technical and regulatory judgments that protect against interference. DIU’s claim that the systems routed demonstration requests in minutes therefore supports a narrower finding: the finalists showed an automated workflow. It does not by itself demonstrate that a live federal assignment, a certification decision or a safety-of-flight determination was completed in minutes.

The two winners approach the problem from different starting points. Airbase describes its product as a unified platform for spectrum licensing, coordination and intelligence. Xtremis says its spectrumOS platform combines distributed sensing, RF modeling, APIs and automated spectrum-allocation tools; it says subscriptions are currently restricted to U.S. government agencies and qualifying government partners. Those are company descriptions, not independently validated performance findings from the challenge.

What an OT would mean — and what it would not

DIU often uses Other Transaction authority to move prototype projects faster than traditional procurement. Its public contracting explanation says it can award prototype agreements in as few as 60 to 90 days. Under 10 U.S.C. § 4022, DIU is among the Defense Department officials authorized to carry out relevant prototype projects, subject to statutory conditions; competitive procedures must be used to the maximum extent practicable.

A future OT agreement would thus be evidence of a prototype contracting decision, not proof that a product has displaced NTIA’s role or resolved every operational, safety, host-nation and regulatory constraint. And a prize selection is not the same thing as that future agreement.

The public record establishes the challenge, the Fort Carson test plan and DIU’s announced winners. It also establishes why the problem is difficult: faster software has to fit within the agencies and engineering processes that prevent harmful interference. The next material evidence will be a disclosed agreement, value, performance scope or operational results — none of which DIU supplied in its winner announcement.