AI News Roundup: Pentagon Signs Classified AI Deals, Big Tech’s $700B Bet, Anthropic Surpasses OpenAI in Revenue

Pentagon Signs Classified AI Deals With Seven Tech Giants — Anthropic Left Out

The Department of Defense announced agreements with seven major technology companies to deploy their AI tools on classified military networks. The firms include OpenAI, Google, Microsoft, NVIDIA, Amazon Web Services, SpaceX, and Reflection. Notably absent: Anthropic, which the Trump administration blacklisted after the company insisted the Pentagon include safety guardrails restricting the use of its AI in autonomous weapons and domestic mass surveillance.

The Pentagon labeled Anthropic a “supply-chain risk” — a designation historically reserved for foreign adversaries. Anthropic sued in response, and a federal judge in California blocked the government’s effort last month. In a potential thaw, the White House recently reopened discussions with Anthropic following its Mythos model announcement, and President Trump told reporters the company is “very smart” and could “be of great use.”

Big Tech’s $700 Billion AI Infrastructure Bet

The largest US tech companies now plan to spend as much as $725 billion this year on capital expenditures, primarily on AI data center equipment — up sharply from roughly $410 billion in 2025. The combined figure represents the largest single-year capital expenditure surge in the history of the technology industry, exceeding the GDP of all but the top 20 national economies.

Amazon leads with $200 billion in planned capex, followed by Alphabet at $175–185 billion (doubling its guidance), Meta at $115–135 billion, and Microsoft on track for $120 billion or more. Google Cloud revenue grew 63% year-over-year to $20 billion, the strongest signal yet that AI spending is translating into revenue. Still, some analysts warn of a potential overbuild if demand doesn’t keep pace.

Anthropic Surpasses OpenAI in Revenue at $30B Run Rate

Anthropic’s annualized revenue run rate has crossed $30 billion, up from $9 billion at the end of 2025 — a more-than-threefold increase in roughly four months. That figure now exceeds OpenAI’s estimated $25 billion annualized revenue, marking the first time Anthropic has overtaken its larger rival on this metric.

Enterprise demand is driving the surge: more than 1,000 business customers now spend over $1 million annually on Claude, a figure that has more than doubled since February. Meanwhile, Anthropic secured 5 gigawatts of computing capacity through an expanded partnership with Google and Broadcom to support the growth.

OpenAI Ships GPT-5.5 With Strongest Agentic Capabilities Yet

OpenAI released GPT-5.5 on April 23, calling it the company’s most intelligent model to date. The model excels at writing and debugging code, online research, data analysis, and multi-step software operations. GPT-5.5 is rolling out to Plus, Pro, Business, and Enterprise ChatGPT users, with a Pro-tier variant for heavier workloads.

OpenAI reports improvements on key benchmarks over competitors, including an 82.7% score on Terminal-Bench 2.0. The model is described as significantly more token-efficient than its predecessor, though it comes at a higher price point. It ships with what OpenAI calls its strongest safeguards to date.

EU AI Act Reform Talks Collapse — August Deadline Back in Play

A second round of trilogue negotiations between the European Parliament, the Council, and the Commission ended without agreement on April 28 after 12 hours of talks. The sticking point: how the AI Act overlaps with existing sector-specific regulations governing high-risk AI in products like medical devices and toys.

The collapse matters because the European Commission’s Digital Omnibus for AI proposed deferring the high-risk compliance deadline from August 2, 2026 to December 2027. If the Omnibus isn’t adopted before August, the original deadline applies as written. A third trilogue is scheduled for May 13, but the window for any agreed delay to take legal effect is narrowing fast.

Snap Cuts 16% of Workforce, Citing AI Efficiencies

Snap announced plans to lay off roughly 1,000 employees — 16% of its workforce — along with closing over 300 open roles. CEO Evan Spiegel said AI advancements are helping staff work faster and reduce repetitive work, noting that AI now generates over 65% of new code at the company.

The restructuring is expected to cut annual expenses by over $500 million. Snap’s stock jumped 7% on the news. Activist investor Irenic Capital Management, which holds a 2.5% stake, had been pushing for cost reductions. The move adds to a growing pattern of tech companies citing AI-driven productivity gains as justification for workforce cuts.