Models & Labs · United States

The Pentagon Hopes to Speed Up ‘Kill Chain’ AI Buys With 5-Minute Videos

Credited to Wired · wired.com · 5h ago

Photograph · Wired

An under-the-radar procurement program led by the Department of Defense has been granting special statuses to would-be defense contractors that make it easier for the government to quickly purchase their AI offerings, based on product videos that must be no longer than five minutes.

It’s part of the Pentagon’s desperate bid to inject competition in a market the United States military itself destroyed.

The initiative, known as Tradewinds, is managed by DoD’s Chief Digital and Artificial Intelligence Office (CDAO.) Companies can submit videos describing how their AI products address one of several different “strategic focus areas” that the government periodically updates. At least once a month, a panel of judges review submissions and decide whether to accept them into the Tradewinds Solutions Marketplace, where the videos can then be viewed by government officials.

If accepted, the pitched products are considered “post-competitive.” Many procurement regulations include competitive requirements. The status can let government buyers check off that box and potentially help shave months off of award timelines. A DoD official says that in several instances, the department has been able to make awards using Tradewinds in less than a week. OpenAI, Anthropic, and Google are all listed as participants.

OpenAI, Anthropic, and Google declined to comment for this story.

Though the Tradewinds Solutions Marketplace has been around since late 2022, the Trump administration has been far more vocal than its predecessors about the military buying and operationalizing AI. It also hopes to have more money than ever to throw around: As part of its 2027 budget request to Congress, the White House asked for $1.5 trillion for the Department of Defense, citing “historic investments” in AI as a priority.

This year, the military is seeking video pitches for AI tech that can help “enhance force lethality” and those that can “unleash” AI-enabled agents that can assist in the joint targeting process, including so-called kill chain execution.

No longer as relevant? Pitches for products that promote “a culture of responsible AI” and provide guardrails for AI development and use—requests made during the final Tradewinds submissions cycle under the Biden administration. A DoD official says that the 2026 topics were selected to align with a January 2026 AI strategy memo.

“There has never been a better time to build for the Army,” wrote then-sitting US Army secretary Daniel Driscoll as part of a pitch for more defense tech startups included in Y Combinator’s Fall 2026 wishlist of companies to fund. (Driscoll has since resigned from his position and returned to the private sector. He did not respond to a request for comment.)

Opaque Transactions

Acceptance onto the Tradewinds marketplace—and the special “post-competitive” status that comes with it—lets government buyers more easily cut deals using special types of contracts such as “ other transaction ” agreements, or OTs, which aren’t subject to certain federal acquisition requirements. By law, the Pentagon can spend up to $500 million dollars on a single OT without having to notify Congress.

Procurement experts say that OTs are considered fast and flexible, which can be appealing to startups that don’t want to gamble on whether a months-long procurement process will result in a contract. However, not all agencies publicly report their OT spending, and even when they do, they’re difficult to find on government-run sites that track federal spending, like USASpending.gov and SAM.gov.

“While OTAs offer opportunities to attract new procurement partners and reduce red tape, numerous reviews have raised red flags,” US senator Joni Ernst of Iowa tells WIRED. “These include questionable costs, mishandling of restricted information, a lack of oversight, reduced accountability and transparency, and an inability to track and access OTA outcomes.”

Ernst recently worked to pass the Stop Secret Spending Act of 2025, which will require the Department of Defense and other agencies to report OTs to the public. They’ll have three years to get into compliance.

According to public records, Google, Anthropic, and OpenAI all have OTs with the Pentagon’s CDAO. According to a DoD official, Tradewinds was leveraged along with market analysis and research as part of the department’s AI acquisition strategy, but the three companies did not confirm whether Tradewinds was used for the specific agreements WIRED identified.

Armed Forces Meet Market Forces

One of the key goals of Tradewinds is to bring “nontraditional” players into the defense industry, says Will Roberts, a former acquisition lead for the Joint Artificial Intelligence Center, CDAO’s predecessor, who helped launch the effort. In some ways, it’s an effort to swing the pendulum back to where it was in the past.

Following the Cold War, the Pentagon encouraged defense companies to merge during a now-infamous meeting called the Last Supper, since Congress was preparing to slash defense budgets. Even though it may seem like only having to deal with a few large companies might make things easier, “it actually had a huge impact on market competition,” says Roberts.

Around the same time, explains Sharon Weinberger, a national security reporter and author of the book Valley of Death: How Big Tech Is Fueling the Future of War, the personal computer boom meant that Silicon Valley was able to make its own money and with that, set its own research priorities without being beholden to the funding and influence of the military.

Roberts and other defense professionals brought up the military’s historic role in funding—and shaping—technologies like the Internet and GPS. With those investments also came leverage, says Roberts. “We've sort of lost that leverage because there's billions of dollars in AI being invested by companies themselves.”

Weinberger says this has given Silicon Valley titans an immense amount of power over the Pentagon’s technological future. “The amount of money that the Pentagon can put into AI, compared to what venture capital can invest? It's like a drop in the bucket,” she says. “And that's really frightening for the Pentagon. What happens when they can't drive the direction of technology anymore?”

Publicly, this can play out in dramatic and confusing ways. Weinberger points out, for example, that the recent legal battle over Anthropic’s status as a supply chain risk could be seen as the government trying to figure out what legal mechanisms it has to flex its authority over AI companies.

Silicon Valley sensibilities have also been a roadblock in the past, albeit briefly. Roberts recalls how Google employees got “super paranoid” in 2018 about the company’s involvement with the Pentagon’s Project Maven, which eventually led to Google pulling out of the effort. Roberts says the Pentagon spent years trying to “appease” Google with less politically charged projects like those focused on humanitarian aid and veteran health.

Google declined to comment on whether that telling was accurate. One of the Google solutions listed as “awardable” on the Tradewinds site is called “Google for Military Health - Tradewinds,” and in July 2025, the company announced a $200 million deal with CDAO to utilize its frontier AI technology. In a press release, Jim Kelly, a vice president at Google Public Sector said that the company was “committed to advancing the deployment of innovative technology across the defense ecosystem.”

Models & LabsUnited States

BofA Sees Strong Demand for OpenAI, Anthropic IPOs

Credited to Bloomberg$ · bloomberg.com2h ago

Useful

Bank of America’s JD Moriarty says public markets have the capacity to absorb massive potential IPOs from companies including OpenAI and Anthropic, pointing to SpaceX’s listing as evidence of strong institutional and retail demand for transformational technology companies. He…

View the article · $Bloomberg

Front Method

Models & Labs · United States

Google bets Gemini can turn casual players into game developers with new Playground feature

Credited to The Decoder · the-decoder.com · 5h ago

Photograph · The Decoder

Google launched Playground, a browser-based AI platform that lets adults in the US create their own games using nothing but text input. No coding skills needed. Users can tweak game rules, physics, characters, and environments through a back-and-forth with the AI, then test the results right away. The platform runs on Google's Gemini, Nano Banana, and Lyria models, according to Google spokesperson Nia Carter (via The Verge ).

Finished games can be shared via link or published in a public gallery. Some genres will support leaderboards and multiplayer. Playground is free, though Google One subscribers get higher weekly usage limits depending on their plan. Published games go through safety reviews based on Google's community guidelines.

Alongside Playground, Google and Unity announced Unity Spark, an AI tool aimed at professional development with access to the Unity Asset Store. Unity CEO Matt Bromberg says Spark targets gamers and artists who have creative ideas but lacked the technical skills to build them. The closed beta is set to launch sometime in 2026. The announcement follows similar AI game-building tools from Meta and Roblox. Ad Ad

AI News Without the Hype – Curated by Humans