US Military Used Claude AI in Iran Strikes Despite Trump Ban
3 min. read 
Published on
The US military deployed Anthropic’s Claude AI during joint strikes on Iran. This happened hours after President Trump banned its use across federal agencies. The operation showed deep reliance on the AI despite the new supply chain risk label.
Operation Epic Fury targeted Iranian nuclear sites and military bases on February 28, 2026. US Central Command used Claude for intelligence analysis and target selection. Reports from The Wall Street Journal and Axios confirmed this use in real time.
Defense officials said switching systems was not possible on short notice. Claude runs in classified networks with no quick replacement. The strikes proved AI integration goes beyond easy removal.
Tensions between Anthropic and the Pentagon grew after a January 2026 mission. That operation captured Venezuelan leader Nicolás Maduro using Claude. The company refuses military use for weapons or surveillance.
Key Timeline
- Feb 26: Anthropic CEO Dario Amodei restates usage limits publicly.
- Feb 27: Defense Secretary Peter Hegseth names Anthropic a supply chain risk, allows 6-month transition.
- Feb 28: Operation Epic Fury launches 4.5 hours later; Claude aids targeting.
- Feb 28: OpenAI’s Sam Altman announces Pentagon deal with safety terms.
Anthropic plans court action against the ban. They call their rules safety musts, not options. Pentagon pushes for full access in operations.
OpenAI stepped in fast. Altman shared on X that models will enter classified networks. Safety engineers will monitor from inside the Pentagon. Terms match Anthropic’s rejected ones.
Dispute Details
| Issue | Anthropic Position | Pentagon View |
|---|---|---|
| Autonomous Weapons | Prohibited | Needs flexibility for ops |
| Mass Surveillance | Banned for US citizens | Required in some cases |
| Transition Time | Not specified | 3-6 months needed |
| Replacement Cost | N/A | High due to integration |
This clash highlights AI ethics in warfare. Developers want controls. Military demands speed. Operation Epic Fury sets a precedent for future use.
Replacement could take until mid-2026. Classified systems run Claude deeply. No other AI matches its role yet.
Expert Context
Defense One notes the ban creates irony. AI aided strikes against a fresh prohibition. This tests how ethics shape procurement.
Anthropic’s stand reflects industry trends. Safety first versus operational needs. Courts may decide the balance.
FAQ
Short notice made full cutoff impossible; Claude sits in classified systems.
Joint US-Israel strikes on Iranian nuclear and military targets on Feb 28, 2026.
No autonomous weapons or US citizen surveillance.
They signed a deal hours after the ban, with similar safety rules.
3-6 months, per Defense One reports.
Wall Street Journal, Axios, Reuters.
Read our disclosure page to find out how can you help VPNCentral sustain the editorial team Read more
Improve this guide
User forum
0 messages