Court rules Trump can blacklist Anthropic for refusing to enable Claude features

The US Court of Appeals for the D.C. Circuit upheld, in a 2-1 ruling, the Trump administration's blacklisting of Anthropic under the Supply Chain Security Act. The court found the government had authority to act over Anthropic's withholding of certain AI capabilities despite no malicious intent, leaving the company excluded from relevant government supply-chain activity. The decision highlights escalating policy and defense-sector risks around AI model constraints, military reliability, and lethal-targeting safeguards.

Analysis

The investable implication is less Anthropic-specific than a procurement-regime shift: federal buyers can now impose behavioral requirements on frontier-model vendors without proving malicious conduct. That increases the value of vertically integrated, auditable government AI stacks—PLTR, MSFT and GOOGL—while raising compliance costs and contract-risk discounts for model providers whose safety policies can conflict with classified or operational use cases. The near-term revenue impact is likely immaterial for hyperscalers absent disclosed contract scope, but the precedent improves their negotiating position in defense AI procurement over the next 6-18 months.

PLTR is the clearest public read-through because expanded demand for governed deployment, data lineage and human-in-the-loop workflows favors its existing defense integration rather than standalone model capability. RTX, LMT, NOC and LHX could benefit secondarily if autonomy programs shift toward approved software-and-sensor packages, although program budgets—not the court decision—remain the earnings catalyst. AWS faces a modest strategic negative: exclusion of a major model ecosystem from sensitive workloads could reduce differentiation in government cloud, but this is unlikely to move AMZN estimates unless agencies broaden restrictions to cloud hosting or model-training access.

Consensus may overprice a direct winner-take-all outcome for rival models. A ruling validating agency discretion also creates bilateral risk: DoD can impose operational constraints on any provider, potentially slowing deployment cycles, requiring indemnities, and compressing software margins through bespoke compliance. The thesis is falsified if procurement guidance remains narrowly confined to one vendor, or if quarterly federal AI bookings/backlog at PLTR and defense IT contractors fail to accelerate within two reporting cycles.

Key Decisions for Investors

• Watch, do not chase, PLTR: initiate only on a 10-15% pullback or after independently verifiable federal-bookings acceleration; target a 3-6 month relative long versus IGV, with a stop if next-quarter U.S. government revenue growth decelerates or management does not cite incremental AI deployment.
• Pair long PLTR / short BBAI for 3-6 months: the likely spend shifts toward proven classified-workflow integration, while BBAI has greater financing and execution sensitivity. Size modestly because both can trade on broad AI beta; exit if BBAI secures a material, funded DoD award or PLTR government backlog disappoints.
• Accumulate LHX or NOC on weakness rather than buying headline strength: use 6-18 month exposure to defense-autonomy integration, where sensor, communications and mission-system content is more monetizable than foundation-model access. Reassess after FY2027 defense-budget submissions; a flat/down autonomy and C4ISR allocation invalidates the premise.
• Set an AMZN alert rather than a short: investigate whether any restriction affects AWS-hosted Anthropic workloads, government-cloud eligibility, or committed-spend obligations. A disclosed material workload migration to Azure or Google Cloud would support a tactical long MSFT/short AMZN spread; without that data, the cloud read-through is too small for a trade.