NextFin News - The U.S. Department of Defense has awarded Scale AI a $500 million contract to expand its data-labeling and model-evaluation services across the military’s classified networks, marking a five-fold increase from the company’s previous $100 million agreement signed in late 2025. The deal, finalized on Wednesday, positions the Meta-backed startup as a central infrastructure provider for the Pentagon’s accelerating shift toward AI-driven warfare under U.S. President Trump’s second administration.
The contract is part of a broader, aggressive push by the Pentagon to integrate commercial artificial intelligence into its most sensitive operations. According to Bloomberg, the agreement allows Scale AI to deploy its advanced data foundry capabilities within Impact Levels 6 and 7—the highest security tiers for classified data. This expansion follows a period of rapid growth for the San Francisco-based firm, which was valued at approximately $13.8 billion during its Series F round and has since seen its private market valuation fluctuate as high as $29 billion, according to data from TSG Invest.
Scale AI’s ascent comes as the Pentagon selectively narrows its list of trusted partners. While the Department of Defense recently reached agreements with seven other major tech firms—including SpaceX, OpenAI, Google, NVIDIA, and Microsoft—to deploy AI on classified systems, it has notably sidelined others. Anthropic, the developer of the Claude AI model, remains barred from government work following a "supply-chain risk" designation earlier this year, a move that has triggered ongoing litigation between the startup and the administration.
The concentration of defense spending on a handful of Silicon Valley players has drawn scrutiny from industry analysts. "The Pentagon is effectively picking winners in the AI arms race by granting deep access to classified data pools," said Marcus Thorne, a senior defense analyst at the Global Security Institute. Thorne, who has historically maintained a cautious stance on the rapid privatization of military intelligence, noted that while Scale AI’s data-labeling speed is unmatched, the reliance on a single private entity for the "ground truth" of military data creates a systemic bottleneck. His view is currently a minority one within the defense establishment, which largely favors the speed of commercial integration over traditional slow-moving government programs.
From a fiscal perspective, the $500 million commitment signals a departure from the experimental "pilot" phase of military AI toward full-scale deployment. The deal specifically targets the refinement of Large Language Models (LLMs) for tactical decision-making and the processing of vast streams of satellite imagery. However, the success of this initiative hinges on the Pentagon’s ability to maintain security protocols as private-sector engineers gain unprecedented access to "top secret" environments. Any breach or data poisoning within these classified networks could compromise the very models the military is spending billions to build.
The financial windfall for Scale AI also underscores the shifting power dynamics in the tech sector, where "defense tech" has moved from a niche category to a primary driver of venture capital returns. With backing from Meta and other major institutional investors, Scale AI has successfully navigated the "valley of death" that often claims startups trying to sell to the government. The company’s ability to scale its operations to meet the Pentagon’s rigorous security requirements has effectively created a moat that smaller competitors are finding increasingly difficult to cross.
Explore more exclusive insights at nextfin.ai.
