Models, infrastructure, agents, and applied intelligence systems.
The Stanford AI Index 2026 documents AI coding reaching near-perfect scores, a $581 billion investment year, and a 2.7% US-China performance gap that a single model release can now flip.
Anthropic's annualized revenue hit $30 billion in April, surpassing OpenAI's $25B — the first time a challenger AI lab has led the company that invented ChatGPT on revenue.
Within 24 hours on April 23 and 24, Meta announced 8,000 layoffs effective May 20 and Microsoft launched its first-ever voluntary buyout, redirecting human-labor budgets toward a $700 billion AI buildout.
Within thirty hours of each other, OpenAI doubled GPT-5.5's API price to $5/$30 per million tokens while DeepSeek V4-Pro landed at one-ninth the cost with near-equal coding benchmarks.
After SpaceX optioned Cursor for $60 billion, Cognition AI seeks a $25 billion valuation for its Devin agent; it already owns Windsurf.
Bloomberg reports Google plans to invest up to $40 billion in Anthropic, turning frontier AI into a contest over capital, cloud capacity, and enterprise distribution.
Bloomberg reports Alphabet will commit up to $40 billion to Anthropic, deepening its bet on a startup that also competes with Gemini and giving Anthropic a second hyperscaler patron alongside Amazon's $25 billion equity and $100 billion compute deal.
Reuters and the IEA point to the same shift: AI expansion is no longer limited by chips alone, but by electricity, transmission and permits.
DeepSeek's latest releases matter less as model marketing than as a serious open-source bid for the software layer that governs AI clusters, training jobs and inference economics.
Anthropic's 10-year, $100 billion pledge to AWS for 5 gigawatts of Trainium capacity turns a cloud contract into an industrial policy statement about who gets to train frontier AI at scale.
SK Hynix's record quarter shows the AI memory story is really about HBM scarcity, packaging bottlenecks, and margin power that rivals still struggle to match.
Florida's criminal probe into OpenAI over the FSU shooting pushes AI safety out of the realm of reputational damage and into potential criminal-liability risk.
OpenAI's decision to kill Sora barely six months after launching a standalone app turns a flashy AI-video bet into a harder lesson about compute costs, copyright risk and weak monetization.
Reuters and the Financial Times report Nvidia is redesigning its China AI chips after tighter U.S. curbs, turning one product line into a test of pricing, policy and competitive power.
Reuters said OpenAI's latest SoftBank-led financing values the company at about $300 billion and could bring in up to $40 billion, giving investors rare data on private AI economics.
Scrutiny over Nvidia chip flows through Singapore and Malaysia is turning AI demand into a test of export controls, corporate compliance and regional geopolitics.
Nvidia said new U.S. export controls on its China-bound H20 chips will trigger a $5.5 billion charge, turning Washington's policy shift into an immediate earnings hit and forcing investors to reset revenue expectations.
Microsoft's reported pullback from some data-center leases is a rare test of whether AI demand and pricing can justify the sector's infrastructure spending.
Reuters says the Bank of England is testing how AI could threaten financial stability, raising the prospect that model risk becomes a capital-markets and regulatory issue.
xAI’s fast expansion in Memphis has turned into a fight over temporary gas turbines, emissions and permits, underscoring how AI growth is running into local infrastructure and political limits.
Cloudflare's Pay Per Crawl is more than an anti-scraping tool. It could turn AI access to the web into a metered market controlled at the network edge.
The Bank of England's warning on Anthropic links frontier AI directly to financial supervision, underscoring how new model releases can quickly become banking-regulation events.
Abu Dhabi's MGX, the UK's pension protection fund, and Japanese banking giant SoftBank have committed to OpenAI's flagship infrastructure program, creating an unprecedented financial coalition aimed at securing AI dominance.
SoftBank and Nvidia back the AI search startup at a 9 billion valuation, up 125% year-over-year, as Perplexity targets enterprise APIs and a 50 million push into Asian markets.
OpenAI’s retreat from a Norway data-center deal, and Microsoft’s takeover, highlights a new AI constraint: securing power, permits, capital, and political approval for hyperscale compute.
A lawsuit over Google’s Gemini chatbot and a man’s suicide turns abstract AI safety promises into a concrete test of product liability, guardrails and acceptable risk.
Google’s Agent2Agent protocol pushes AI beyond single-agent demos, defining how autonomous systems discover peers, delegate work and return results across vendor boundaries.
Meta's acquisition of Chinese-founded AI startup Manus for over $2 billion highlights the escalating battle for AI talent, as the startup relocated to Singapore before sealing its deal with the US tech giant.
OpenAI has discontinued Sora, its AI text-to-video generator, citing unsustainable economics and compute constraints. The company is pivoting the Sora team toward robotics research and world simulation as it prepares for a potential IPO.
OpenAI has flagged its dependence on Microsoft as a significant risk factor in investor documentation ahead of a potential IPO, highlighting the operational and competitive complexities of one of technology's most consequential partnerships.