Zuckerberg's 6500-Word AI Manifesto Anchors a Week of Big AI Infrastructure Moves

Meta CEO Mark Zuckerberg published a 6,500-word essay laying out his vision for AI as a broadly accessible public good, arguing that the greatest risk would be any single entity gaining too much control over the technology. He said Meta plans to offer free AI tools to billions of people, with heavier compute access sold through a dynamic auction mechanism, and envisions individuals each having a personal AI agent handling health, career, finance and relationship needs. The manifesto arrives alongside a string of other AI industry developments.

Microsoft confirmed its 365 Copilot product has surpassed 30 million paid seats, with seat additions more than doubling quarter-over-quarter, as EVP Charles Lamanna described a broad shift in which software engineers are already ceding routine tasks to AI agents, with sales and finance workers set to follow. Microsoft frames the trend as a productivity driver rather than a headcount reduction. Cerebras Systems reported Q1 2026 revenue of $193.4 million, up roughly 94% year-over-year, guided Q2 revenue to $194 million above analyst estimates of $174 million, and holds a $2.2 billion cash position.

The company's WSE-3 chip delivers what analysts describe as 7,000 times the memory bandwidth of Nvidia's H100, underpinning contracts with OpenAI, Amazon and AMD. Separately, Silicom Ltd is being highlighted as a smaller-cap beneficiary of AI inference demand, with analysts pointing to a potential $80-100 billion total addressable market and projecting that a single unnamed customer could add $64 million in revenue by 2027 on a 2026 base of $95 million. A broader Barron's analysis cautions that despite the scale of AI capital spending, profit margin expansion has not yet materialized for most companies involved.