OpenAI’s GPT-6 Lands on Google Cloud: Rivals Unite as the Only Hyperscaler Hosting All Three AI Giants
OpenAI and Google Cloud announced a landmark partnership on October 21: GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna, plus the Codex agentic coding suite, arrive in Google Cloud’s Model Garden as a private preview. Google Cloud becomes the only hyperscaler offering Gemini, Claude, and GPT-6 side by side, with unified billing and existing-commitment drawdown — a structural shift that ends Microsoft’s practical exclusivity and rewrites enterprise AI procurement.
AI Found a 14-Year-Old Bug in PHP: PHP2Shell Forces Government Websites Offline Worldwide as 70% of the Web Races to Patch
Security firm Calif, working through OpenAI’s Patch the Planet initiative, used GPT-5.6 Sol to uncover two critical PHP vulnerabilities dormant for over 14 years — CVE-2026-109618/109619, collectively dubbed PHP2Shell (CVSS 9.2) — enabling unauthenticated remote code execution. After details leaked a day early, the NSA issued a rare public warning, government websites in Canada, Australia and Europe went dark as a precaution, and Japan’s PM personally cautioned the public. With roughly 70% of the web built on PHP, defenders are racing attackers to patch.
SpaceX Targets 13GW AI Compute by 2027 and Terafab Chip Production in 2028: Musk’s Vertical Integration Gamble
SpaceX’s Q3 2026 earnings call delivered staggering numbers: roughly 13GW of AI compute targeted by end-2027 — the power of a dozen nuclear reactors — and volume production at the SpaceX-Tesla Terafab chip fab by Q3 2028, with ASML accelerating equipment deliveries. Q3 revenue hit $14.6 billion, up 161%, and 200,000 Nvidia chips now come online monthly. Musk is rebuilding SpaceX into a vertically integrated AI empire.
CERN Teams Up With Mistral: Timestral Beats Google at 209M Parameters, and Europe’s First Collider-Physics AI Model Is Coming
Mistral fired twice in one day: Timestral, its first time-series foundation model family, tops all three major benchmarks with just 209M parameters — beating Google’s TimesFM-3 — and runs inference on CPUs without a GPU. The same day, Mistral signed a strategic partnership with CERN to build the first foundation model for collider physics, trained on billions of simulated collisions. Europe is betting on open weights plus fundamental science, and the barrier to enterprise predictive analytics just dropped.
“All of This Is Going to Pass”: Jensen Huang Races to Calm Markets as Chip Stocks Log Longest Losing Streak Since 2008
AI stocks are enduring their darkest week since the 2008 financial crisis: the Philadelphia Semiconductor Index has fallen for eight straight sessions, Nvidia has slid more than 12% from its record high, and CoreWeave’s bonds are flashing distress signals. In Berlin, Nvidia CEO Jensen Huang told investors the turmoil will pass — “stay focused on first principles.”
Taiwan’s First AI Robot Sports Tournament: Combat Robots Enter an Elementary School — and a 2-Meter Paper Giant Steals the Show
Taiwan’s first AI Robot Sports Tournament kicked off October 17 at Jian’an Elementary School in Taipei, turning robot combat into a new frontier of tech education. Deputy Mayor Lin Yi-hua championed “learning by doing,” while a 2-meter giant robot built from recyclable honeycomb paper stole the spotlight. With AI education equity at its core, the organizer will launch seven winter camps — its Singapore camp invites one rural student for every five enrollees, narrowing the urban-rural tech divide.
Apple and Meta Bury the Hatchet: Muse Spark Open Puts an 816B-Parameter Open-Weight Model on the Mac
Meta has released Muse Spark Open, an 816-billion-parameter open-weight model for coding and AI agents, in a surprise partnership with Apple — the rival it was feuding with just two weeks ago. Optimized to run locally on two M5 Ultra Mac Studios with 1TB of unified memory, it is the most capable US open-weight model to date, aimed squarely at Chinese open-source rivals.
Jensen Huang Draws the Line: Nvidia Will Cut Off Chips to Unsafe AI Developers — After Talks First
Nvidia CEO Jensen Huang said in London he would cut off chip supply to customers developing unsafe AI — but only after talks first. Calling current doom concerns overblown, he argued for accelerating AI monitoring and containment tech over slowing down, and announced a $250 million, five-year donation to AI safety research at Oxford and UCL.
Character.AI Files for Bankruptcy: Disney Buys Assets for $20 Million as an AI Star Startup Falls
Character.AI, one of the first breakout startups of the generative AI boom, has filed for Chapter 11 bankruptcy and named Disney as lead bidder for its assets in a $20 million deal. With liabilities estimated between $50 million and $100 million, the platform could be wound down if no suitable buyer emerges by early December. The collapse lays bare the harsh economics of the AI companion chatbot market.
Intel’s D-Series AI GPU Is Here: 1.3 PFLOPS at Just 350W, Aiming Straight at Nvidia’s Weakest Flank — Inference Cost
At the OCP Global Summit, Intel officially branded its Crescent Island AI inference chip as the GPU D-Series. The flagship D270 delivers 1.3 PFLOPS of FP8 at just 350W with up to 480GB of LPDDR5X, betting on cost-per-dollar and cost-per-watt rather than peak performance to attack the data center inference market where Nvidia is most vulnerable. Customer sampling has begun.