Trump says Anthropic is shaping up and a deal is 'possible' for Department of Defense use

Trump said it is "possible" the Department of Defense could reach a deal allowing Anthropic's AI models to be used by the Pentagon, following recent White House talks. The comments come after the DoD labeled Anthropic a supply chain risk in March, which requires defense contractors to certify they do not use Claude in military work. The development is constructive for Anthropic and could reduce a key defense-sector barrier if negotiations advance.

Analysis

This is less about one vendor and more about a reset in procurement power: a White House nod implies the defense AI stack is still up for grabs, and the market should treat this as a signal that policy risk can be negotiated away if the capability gap is large enough. The second-order winner is likely not just the named model provider but the broader secure-inference and deployment layer—prime contractors, cloud integrators, and GPU/networking vendors that can package compliant, auditable AI for the Pentagon. That shifts value away from raw model branding toward whoever can make models usable inside classified and air-gapped workflows.

The biggest loser is the notion that regulatory designations are durable moats. If the government is willing to reverse or soften a supply-chain-risk stance on strategic grounds, then competitors with stronger enterprise distribution but weaker model quality may face pricing pressure as procurement turns into a performance race. Expect a short-term repricing in adjacent defense-AI names as investors extrapolate a faster path to federal adoption, but the real monetization likely takes months: compliance reviews, contract language changes, and security validation are the bottlenecks.

The tradeable catalyst is a potential re-rating of defense tech enablers rather than the policy headline itself. The near-term risk is that talks stall and the issue becomes political theater, which would fade the signal within days; the longer-dated risk is a broader restriction regime if the Pentagon concludes that the security model is still unresolved. Contrarian view: the market may be overestimating how quickly a symbolic White House meeting translates into revenue, while underestimating the probability that this becomes a precedent for selective exceptions across other banned or restricted AI vendors.

Key Decisions for Investors

• Go long PLTR vs. short a basket of legacy defense primes for 1-3 months; if federal AI adoption accelerates, software orchestration and data-layer exposure should outperform pure hardware contractors on margin expansion.
• Buy near-dated call spreads on SMH or NVDA over the next 4-8 weeks; a Pentagon AI procurement thaw would reinforce demand for secure inference, networking, and accelerated compute, with upside skew if the market re-prices government demand optionality.
• Pair trade: long CRWD / short a generalized software index for 1-2 quarters; any move toward controlled AI deployment inside defense should benefit cybersecurity and identity/access tooling as gating spending, not just model vendors.
• Avoid chasing any direct AI-vendor bounce until there is evidence of contract language change; if no formal policy shift emerges within 2-4 weeks, fade the headline into the prior range.
• If you want asymmetric exposure, consider call spreads on a defense cloud/infrastructure name with federal exposure (e.g. MSFT) 3-6 months out; the best risk/reward is in the picks-and-shovels layer, not in the controversial model layer.