Officials at agencies, including the CIA, reportedly still hope for a peace agreement to resume or salvage talks between Anthropic and the Department of Defense.
Anthropic refuses to lower AI guardrails for the Pentagon, declining requests to reduce safety constraints for military use.
OpenAI agrees to consume approximately 2 gigawatts of Trainium silicon capacity from AWS.
Anthropic works closely with Nvidia, Google, Amazon, and Palantir as US military partners potentially affected by its dispute with the Department of Defense.
OpenAI commits to consuming approximately 2 GW of Trainium capacity through AWS.
Sam Altman says OpenAI is seeking a deal with the Department of Defense.
Amazon and OpenAI plan a stateful runtime environment on AWS that lets AI agents carry context forward to handle ongoing projects as part of their deal, potentially putting Amazon first to market with this developer-facing AI service.
The Pentagon says it offered compromises to Anthropic, including putting in writing that laws bar the military from mass surveillance of Americans.
Altman says OpenAI would exclude use cases like domestic surveillance from any DOD deal.
Amazon's investment in OpenAI yields none of the exclusives that Microsoft secured, leaving Amazon without the exclusive rights Microsoft obtained from OpenAI.
Sources report the DOD was in talks with leading AI companies to build AI tools to automate reconnaissance of China's power grids, utilities, and sensitive networks.