The Pentagon’s New AI Strategy: Speed, Disruption, and the End of Cost-Plus Contracting

The era of defense acquisition moving at a glacial pace is over, replaced by a strategic mandate where "speed wins." This was the core message delivered by Emil Michael, the Department of Defense’s (DoD) Chief Technology Officer (CTO) and Under Secretary of Research and Engineering, during a recent interview with Morgan Brennan. Michael detailed the Pentagon's newly unveiled AI acceleration strategy, alongside fundamental changes to defense contracting designed to inject Silicon Valley velocity and competition into the traditionally slow-moving defense industrial base (DIB). The commentary provides a sharp analysis for founders and investors regarding how the DoD is not just adopting technology, but completely overhauling its operating model to meet current geopolitical threats.

Michael spoke with Brennan about the DoD’s new strategy, emphasizing that the Department of War has historically failed to deploy artificial intelligence to its full potential. With the rapid explosion of AI in the commercial sector over the last two years, the DoD recognized that the time for incremental adoption is past. The acceleration strategy centers on bringing the latest commercial large language models (LLMs) directly to the department’s vast workforce of three million employees. The goal is immediate efficiency gains across every domain, from enterprise use cases like administrative efficiency to highly specific intelligence gathering and, critically, warfighting capabilities. Michael noted that this is about maximizing the capabilities built by companies that have invested hundreds of billions of dollars in their AI models, bringing those cutting-edge tools directly to the warfighter.

The speed of deployment is already evident. Michael confirmed that they have already deployed Google’s Gemini model to three million DoD employees, with nearly a million unique users clamoring for the technology. This rapid adoption signifies a massive cultural shift within the Pentagon, signaling that bureaucratic inertia is being actively challenged by necessity. Looking ahead, the strategy includes integrating other leading models, including Grok from XAI, into both classified and unclassified networks. The underlying insight driving this rapid deployment is simple: if three million employees can be empowered to do the work of ten million because they are more efficient, more empowered, and can use their time more effectively, the increase in fighting force capability is profound. The lessons learned from early adoption, Michael explained, mirror the commercial world: while there are imperfections in new models, the sheer leverage provided by AI in synthesizing massive datasets and increasing efficiency is "incredible."

Of course, integrating commercial, rapidly evolving AI models into the defense apparatus raises immediate concerns regarding guardrails, hallucinations, and security. Michael addressed this head-on, stating that while the model companies build in their own guardrails, the DoD layers its own strict policies on top. The overriding principle for deployment is clear: "The Department of War needs to use your model and its capabilities for every lawful use case. As long as it's lawful, we'll use it." He stressed that the DoD is a workplace with a critical mission, meaning the systems are strictly for professional purposes and adhere to Pentagon security standards, differentiating its use from potentially riskier consumer applications.

Beyond AI deployment, Michael painted a picture of a broader, systemic disruption targeting the long-standing flaws in the defense industrial base. The traditional reliance on a small handful of prime contractors—a number that has shrunk drastically from 50 in the 1980s down to just five today—created a system lacking competition and speed. Worse, the historical contracting model of "cost plus" incentivized delay and cost overruns. Michael explained: "The way it didn't work in the Pentagon was that we were reliant on a small number of prime contractors... and the business models were broken too. We were paying cost plus." This payment structure inherently rewarded projects that took longer and cost more.

The new strategy seeks to reverse this through a pivot to fixed-bid contracts, shifting the financial risk and capital expenditure back onto the contractors. This move rewards companies for delivering on time and under budget, aligning incentives with efficiency. This strategic realignment creates a crucial window for new entrants—companies like Anduril, Palantir, and SpaceX—who operate with modern, venture-backed speed and commercial innovation. The old system required new companies to navigate an arduous process involving lobbyists and complex Requests for Proposals (RFPs) that often dictated solutions rather than defining problems. Now, the DoD is changing the system to present "common operational problems" and asking the industry to return with creative solutions. Michael noted that this change in approach is essential because it "gives new entrants a chance to solve problems... as opposed to getting a long list of requirements that they have to try to fulfill in detail."

The shift extends deeply into the supply chain, as evidenced by the recent $1 billion deal with L3Harris to boost solid rocket motor production, a critical component across multiple weapon systems. This investment was facilitated by the Office of Strategic Capital (OSC), a new DoD vehicle designed to leverage government lending authority to incentivize domestic production of scarce resources. Michael viewed this as a critical strategic move: "We’re bolstering the supply chain underneath the primes so that they can move faster and we can have more capability across many systems." This investment in critical components, which had often been outsourced or suffered from brittleness due to inconsistent demand signals, is about "re-domesticating" production. By offering a "firm demand signal"—promising to buy the product if the companies invest their own capital to build the infrastructure—the Pentagon is effectively transforming these companies into "wartime footing" enterprises focused on growth and rapid delivery, rather than waiting for slow, costly government orders. This powerful combination of technological acceleration and acquisition reform signals the Pentagon’s aggressive commitment to speed, competition, and domestic resilience in the face of escalating global threats.