OpenAI CEO Sam Altman published a 13-page blueprint titled Industrial Policy for the Intelligence Age proposing robot taxes, a nationally managed public wealth fund, four-day workweeks at full pay, and auto-triggering unemployment benefits. Policy experts from Carnegie, the Inter-American Development Bank, and Encode AI criticized the timing given OpenAI's upcoming IPO and lobbying history.
Anthropic officially launched Claude Mythos Preview through Project Glasswing, giving 12 founding partners and 40 organizations access to its most capable model for defensive cybersecurity. The model found 181 Firefox exploits versus 2 for Opus 4.6, discovered zero-days up to 27 years old, and costs under $1,000 per exploit. Anthropic committed $100 million in credits but refused public release.
Anthropic's run-rate revenue surged from 9 billion to 30 billion dollars in roughly three months. The company signed a deal for 3.5 gigawatts of next-generation Google TPU capacity through Broadcom, its largest compute commitment ever, with more than 1,000 enterprise customers now spending over a million dollars annually.
Through the Frontier Model Forum, OpenAI, Anthropic, and Google are sharing indicators related to suspected distillation attacks. The companies attribute roughly 16 million unauthorized exchanges to activity involving DeepSeek, Moonshot AI, and MiniMax.
JPMorgan Chase has made AI adoption a formal performance requirement for 65,000 engineers. Internal dashboards track GitHub Copilot usage by individual, and the bank reports 10-20% productivity gains from AI coding tools.
Utah approved Legion Health to let its AI chatbot independently renew 15 psychiatric medications including antidepressants and anxiety drugs. The pilot features a three-phase oversight structure requiring physician review of the first 1,250 prescriptions before autonomous operation.
SpaceX filed a confidential IPO registration with the SEC on April 1, 2026, targeting a $75 billion raise at a $1.75 trillion valuation. After acquiring xAI in February, the company plans to build orbital AI data centers using up to 1 million Starlink satellites, while skeptics question the economics, launch cadence requirements, and regulatory framework.
Google released Gemma 4, a family of four open-source models under Apache 2.0 license, with the 31B variant ranking #3 on Arena AI. Community testing within 24 hours revealed significant speed limitations compared to Alibaba's Qwen 3.5, VRAM inefficiency, and fine-tuning incompatibility with existing tools.
Oracle eliminated up to 30,000 jobs as it committed $156 billion to AI infrastructure, according to estimates and filings cited in the article. The company also disclosed a $2.1 billion restructuring plan despite posting record quarterly profit.
A North Korean state actor compromised the Axios npm package, the most popular JavaScript HTTP client with 100M+ weekly downloads, deploying a cross-platform RAT through a malicious dependency. The attack lasted 3 hours before detection.
A Utah man filed a proposed class-action lawsuit against Perplexity AI in San Francisco federal court on April 1, 2026, alleging the company embedded hidden tracking software that transmitted user conversations to Meta and Google. The complaint claims trackers activate upon login and operate even in Incognito mode. The plaintiff shared financial and tax information with the chatbot. This is the latest in a series of legal challenges for Perplexity, including Amazon scraping disputes and publisher copyright suits.
OpenAI completed the largest private funding round in history, raising $122 billion at an $852 billion valuation. Amazon committed $50 billion, with $35 billion contingent on an IPO or AGI milestone by 2028. Nvidia and SoftBank each invested $30 billion in phased installments. The company generates $2 billion in monthly revenue but projects a $14 billion loss for 2026, making an IPO increasingly likely by Q4 2026.
Anthropic's annualized revenue rose from $9 billion at year-end 2025 to $19 billion by early March 2026, according to the cited reporting, with Claude Code described as the main growth driver. The company is in early discussions with Goldman Sachs, JPMorgan, and Morgan Stanley about a possible listing as soon as October 2026 that bankers estimate could raise more than $60 billion. Its February 2026 Series G valued it at $380 billion.
Court filings unsealed March 31 in the $134 billion Musk v. OpenAI case reveal Elon Musk privately texted Mark Zuckerberg asking him to join a bid for OpenAI's intellectual property. The exchange occurred February 3, 2025, just seven days before Musk's xAI consortium submitted a $97.4 billion unsolicited offer. Zuckerberg declined to participate but did not publicly disclose the conversation. The texts represent a new dimension to a lawsuit Musk has framed as a crusade to protect OpenAI's nonprofit mission. Jury selection begins April 27 in Oakland.
Mistral AI secured $830 million in debt financing from seven banks to build a dedicated AI data center near Paris housing 13,800 Nvidia GB300 GPUs. The facility is part of Mistral's plan to reach 200 MW of compute capacity across Europe by 2027.
A configuration error exposed documents describing Claude Mythos, which Anthropic characterized as substantially more capable at cybersecurity tasks than its earlier models. Cybersecurity stocks fell the next morning, and Anthropic subsequently briefed government officials about AI-enabled cyber risks.
Attackers compromised LiteLLM 1.82.7 and 1.82.8 after stealing PyPI credentials through an upstream security-tool breach. The malicious packages used Python startup mechanisms to harvest credentials, turning a CI/CD compromise into a downstream software-supply-chain incident.
The White House's March 20 national AI framework called for broad federal preemption of state AI laws and rejected creation of a new federal AI regulator. House Democrats responded with the GUARDRAILS Act, setting up a dispute over whether Congress should replace state rules with a single federal standard.
