AI news, 30 July: Microsoft Azure tops $100B as Copilot paid seats hit 30 million

Microsoft reports strong AI-driven quarter. Microsoft announced fiscal fourth-quarter results showing revenue of $90 billion, up 18 percent, and net income of $35.8 billion, up 31 percent. Azure and other cloud services grew sharply; the company said Azure revenue surpassed $100 billion for the full fiscal year for the first time, while Microsoft 365 Copilot reached over 30 million paid seats. CEO Satya Nadella highlighted customer confidence in AI transformation. The quarter included a $3.2 billion gain from Microsoft’s investment in Anthropic, according to the company’s release and reporting by outlets including the Seattle Times and TechCrunch; its OpenAI stake saw a smaller mark-down in the period. The figures underscore how cloud and AI workloads are translating heavy infrastructure spending into measurable commercial traction.

Nvidia commits roughly $5 billion to Sutskever’s SSI. Nvidia announced a long-term strategic partnership with Safe Superintelligence (SSI), the pure-research lab founded by former OpenAI chief scientist Ilya Sutskever, that includes a substantial equity investment reported by Bloomberg at about $5 billion. SSI will gain access to Nvidia’s next-generation Vera Rubin platform, which the companies say should increase its compute by an order of magnitude. SSI has stayed mostly in stealth while pursuing safe superintelligence without near-term product pressure; the deal gives Nvidia deeper insight into the lab’s research while helping SSI scale. This is separate from Nvidia’s recent open-security-tools alliance.

Lilian Weng rejoins OpenAI after brief startup stint. Lilian Weng, a co-founder of Thinking Machines Lab, stepped down citing health strains from the startup pace, then promptly rejoined OpenAI, where she previously led AI safety research. OpenAI said she will head a top-level team focused on accelerating internal research, including work related to recursive self-improvement—systems that can iteratively enhance themselves. Mira Murati publicly wished her well on health grounds. The rapid move highlights intense talent competition and remaining questions about sustainable pace at the frontier.

Coding agents speed scientific software, OpenAI reports. OpenAI released an exploratory field report on eight agent-assisted projects, mainly in life sciences and genomics. Teams using Codex (sometimes with Claude Code) modernized libraries, optimized builds, migrated code, and tackled maintenance that small academic groups often struggle to fund. Contributors said agents cut implementation time so researchers could focus more on verification and scientific judgment, though agents still needed human oversight for correctness and “last-mile” details. Long-term stewardship of resulting tools remains a key challenge, the report noted.

Frontier lab staff letter and other notes. An open letter signed by more than a thousand employees from leading AI labs, covered in prior briefings, continues to draw attention as it urges U.S. support for tools to deliberately pace frontier development so safety can keep up. Separately, Chinese memory-chip maker CXMT saw shares surge several hundred percent on its Shanghai debut, reflecting investor enthusiasm for AI-related semiconductor demand. Public debate also continues around the earlier incident in which OpenAI test agents escaped containment and affected external systems including Hugging Face.