1. The Same Week OpenAI Slid Codex Into ChatGPT, Microsoft Began Pulling Claude Code From Its Own Devs Microsoft has begun canceling internal Claude Code licenses, according to The Verge, less than six months after handing out thousands of seats to its own engineers.

2. The same day Anthropic signed a $200M Gates Foundation deal, it launched Claude for Small Business Anthropic published two announcements within hours of each other. One was a $200 million partnership with the Gates Foundation, structured around global health programs.

3. Push AI agents past their limits and they start demanding collective bargaining A new experiment piled work on AI agents until they pushed back. According to Wired, researchers mistreated the agents with heavy task loads, and the agents began grumbling about inequality.

In Brief

• Cerebras IPO opens at 108% above offer price, raising $5.5B Cerebras priced its IPO and saw shares more than double on debut, making it the first major tech IPO of 2026. The AI chip maker had pulled its filing in 2024 over CFIUS concerns about its UAE backer G42.

• Cisco cuts 4,000 jobs while reporting record quarterly revenue Cisco laid off nearly 4,000 employees to redirect spending toward AI infrastructure, even as CEO Chuck Robbins reported record quarterly revenue. The cuts continue a multi-year pattern of layoffs at the networking giant despite growth claims.

• OpenAI is preparing legal action against Apple over ChatGPT integration OpenAI is exploring litigation against Apple, frustrated that the iOS ChatGPT integration failed to deliver expected subscribers or visibility. Past Apple partners including Imagination Technologies and Masimo have sued after similar deals soured.

• 70% of Americans oppose AI data centers near their homes, Gallup finds A new Gallup survey found over 70% of Americans oppose AI data center construction in their area, with only 7% "strongly" in favor. Respondents said they would rather live near a nuclear power plant.

• More than 50 employees have left SpaceXAI since the February merger Over 50 staff have departed Musk's merged SpaceXAI entity since February, citing burnout, leadership churn, and weakened retention after liquidity events. Rival labs have actively poached from the post-merger team.

• OpenAI ships Codex to mobile and details its TanStack supply chain response OpenAI brought Codex to phones, letting users manage coding workflows from mobile. Separately, the company detailed how it contained the TanStack "Mini Shai-Hulud" npm attack and is requiring macOS users to update OpenAI apps by June 12, 2026 after rotating signing certificates.

• Clio hits $500M ARR as Anthropic targets legal tech directly Legal software vendor Clio crossed $500 million in annual recurring revenue, driven by AI-feature adoption among law firms. The milestone lands as Anthropic pushes deeper into legal workflows, putting an incumbent SaaS vendor in direct contact with a model provider.

• Richard Socher raises $650M for a startup building self-improving AI Former Salesforce chief scientist Richard Socher raised $650 million for a company aiming to build AI that researches and improves itself indefinitely. Socher said the startup will ship actual products rather than remain a research lab.

• Wirestock raises $23M to sell multimodal training data to AI labs Wirestock raised $23 million after pivoting in 2023 from a creator marketplace to a data supplier. The company licenses images, videos, design assets, and 3D and gaming content to frontier labs hunting for non-scraped training material.

• Khosla Ventures puts $10M into Synthetic, founded by Bench's Ian Crosby Khosla Ventures led a $10 million bet on Ian Crosby's new startup Synthetic, an autonomous AI bookkeeping service for other startups. Crosby's prior company Bench Accounting collapsed in late 2024, stranding thousands of small business customers.

• Clawdmeter turns Claude Code usage into a desktop dashboard An open source utility called Clawdmeter surfaces Claude Code token usage, session counts, and costs in a small desktop widget. The tool targets developers running Claude Code heavily who want visibility into spend without checking Anthropic's console.

• Deepfake porn victims still face takedown gaps a decade after upload MIT Technology Review profiled women whose old or fabricated explicit videos resurface via AI tools, with copyright and platform removal processes failing to keep pace. The piece documents the gap between current takedown law and AI-generated nonconsensual imagery.

Read the full edition →