State of AI: April 2026 newsletter
In early 2026 the US government blacklisted Anthropic over its safety guardrails in Pentagon contracts while OpenAI secured its place on the classified network and Iran attacked AWS data centers used for military AI. Meanwhile, Anthropic’s revenue soared past $30 billion, hyperscaler partnerships expanded, and rival labs raced to release new models amid an industrial-scale distillation clash.
- The White House ordered federal agencies to drop Anthropic over its refusal to remove safety guardrails from a Pentagon contract, while OpenAI stepped in to fill the gap on the classified network.
- Iran struck AWS data centers in the UAE and Bahrain, marking the first direct military attack on commercial cloud infrastructure supporting US AI operations.
- Anthropic's revenue tripled from $14B to over $30B annualized in under a month, driven by enterprise clients paying $1M+/year, though it counts gross cloud-partner revenue unlike OpenAI's net figures.
- OpenAI locked in massive hyperscaler deals (a $50B AWS deal, $100B over eight years) while running at a $25B annualized pace with a $280B revenue target by 2030.
In late February, the White House ordered every federal agency to drop Anthropic’s AI models within six months after the company insisted on safety guardrails—no autonomous weapons or mass surveillance—in its $200 million Pentagon contract. Sam Altman seized the moment, cutting a deal to put OpenAI’s models on the Pentagon’s classified network with similar “red lines” embedded. Within days, State, Treasury and HHS migrated from Anthropic’s Claude to OpenAI’s GPT-4.1, and the Pentagon labeled Anthropic a “supply chain risk.” Anthropic sued the administration for retaliation, and a federal court halted further punishments. Meanwhile, Iran launched drone strikes on AWS data centers in the UAE and Bahrain, claiming U.S. forces were using AI for intelligence and targeting. Two out of three availability zones failed, marking the first direct attack on commercial cloud infrastructure and turning data centers into battlefields.
While geopolitics rattled data centers, Anthropic’s revenue exploded—from a $14 billion annualized run rate in mid-February to over $30 billion by early March. Its 1,000+ enterprise clients each spend at least $1 million a year, and Claude Code and Claude Cowork have become must-have tools. Anthropic reports gross revenue from cloud partnerships, inflating top-line figures, whereas OpenAI nets out the 20 percent share it pays Microsoft. OpenAI counters with deep alliances: a new $50 billion deal with AWS (including $15 billion up front) and a $100 billion commitment over eight years. Microsoft remains OpenAI’s exclusive hyperscaler for Frontier, and the company claims Amazon’s Trainium chips deliver 30–40 percent better performance per dollar than GPUs. By late February, OpenAI ran at a $25 billion annualized revenue pace with forecasts aiming at $280 billion by 2030.