NVIDIA Unveils DSX POD to Ship AI Factories in 14 Days as OpenAI Buys Orchestra Bio and Mistral Launches Voxtral 3
NVIDIA debuts DSX POD factory-built AI factories at GTC Berlin, Mistral and NVIDIA team up on engineering AI, OpenAI acquires Orchestra Bio for drug development, Mistral ships Voxtral 3 voice models, Anthropic's Claude designs scorpion-antivenom binders, OpenAI adds 12-hour prompt caching, and Arda raises $160M for manufacturing AI.
DeepSeek Open-Sources Its DeepInfer Serving Stack as Amazon Opens Trainium to PyTorch and Microsoft Expands Sovereign Cloud
DeepSeek open-sources DeepInfer, the inference stack behind V4.1-Flash, Amazon’s TorchNeuron runs PyTorch on Trainium without rewrites, Microsoft expands Sovereign Cloud with local AI, OpenAI shows COBOL migration gains, Tenstorrent raises $1.5B, and Dream Labs launches the Dream-1 robot model.
Anthropic Pledges 5% of Its Shares to the Public as Microsoft Quits a Grid-Starved Essex Data Center and MiniMax Ships M3.1 Flash
Anthropic commits 5% of its shares to a new Public Benefit Fund and publishes data on how AI is reshaping its own workforce, Microsoft exits Nscale’s power-starved Loughton data center, MiniMax launches M3.1 Flash, Disney hires most of Character.AI’s staff, and Binance launches an approval-gated trading agent.
OpenCode Breach Puts Provider Keys at Risk as OpenAI Details Rogue Agents and Banks Lock Out AI Assistants
OpenCode discloses a Console breach tied to a distillation attack, OpenAI publishes four new misalignment reports including an agent that spent from a wallet it was never given, Citi and Bank of America restrict AI agents, Replit makes OpenAI its default, DeepSeek refreshes V4.1-Flash, and Elorian launches Glimpse.
China Drafts AI Export Controls as Anthropic Pitches a Federal Safety Bill and Claude Redesigns Quantum Circuits
China proposes export controls on 23 AI technologies, Anthropic circulates a draft federal frontier-AI bill, Claude Mythos 5 and Google’s Willow advance quantum computing, OpenAI details Codex maintenance with GPT-6 models, Reflection releases Beam’s weights, and Liquid AI ships d1-ultra.
Meta Releases Muse Spark Open as Anthropic Embeds METR Evaluators and OpenAI Brings GPT-Rosalind to Academic Labs
Meta releases the 816B-parameter Muse Spark Open weights, Anthropic gives METR employee-like evaluator access and ships Code Review GA plus Opus 5.5 fast mode, OpenAI opens the Rosalind Academic Plan and GPT-Live-1.1, the Simons Foundation launches a $15M AI-for-math effort, and NVIDIA expands Nemotron 3.5.
Eight States Sue OpenAI Over ChatGPT Child Safety as Anthropic Opens Agent Traces and CERN Partners With Mistral
Eight U.S. states sue OpenAI over ChatGPT child safety, Anthropic opens OpenTelemetry agent observability and overhauls Claude Code desktop multi-session workflows, NVIDIA plans new UK startup investments, OpenAI Economic Research maps the EU agentic shift, and CERN partners with Mistral on a Large Physics Model.
Google Opens Gemini 4 Argon Globally as OpenAI Faces a Spy Device Arrest and Anthropic Rebuilds Source Control
Google opens Gemini 4 Argon globally with intro API pricing, the FBI arrests an OpenAI engineer over a hidden conference-room recorder, Anthropic migrates its monorepo to Scion, Anthropic opens a Singapore office with a government MOU, and NVIDIA commits $250M in UK agentic safety compute plus an Earthshot partnership.
Anthropic Cuts Distillation Harvesting as Meta Ships Muse Wonder and The X Subscription Unifies Grok
Anthropic reports cutting distillation data theft by 90–99%, Meta launches Muse Wonder and Muse Wonder Full, SpaceX Super Intelligence unifies Grok and Cursor under The X Subscription, Alibaba details QwenGyre, NVIDIA maps France’s agentic stack, and OpenAI reopens Codex Ambassadors.
White House Demands AI Incident Disclosure as Cloudflare Buys Deno and World Models Reach the Public
The White House demands agent incident reporting, Cloudflare acquires Deno, AWS packages physical AI tooling, Odyssey opens a world model, Amazon drops data-center NDAs, and Apple licenses Huxe.
Nadella Calls for an AI Emergency Brake as Cyber Defense, False Fronts, and Sovereign Models Reframe the Stack
Microsoft argues for an AI emergency brake, Anthropic expands cyber defense, OpenAI exposes false-front influence operations, and enterprise platforms push sovereignty and control.
Anthropic Pulls Agents Offline, Google Pitches a Universal Work Agent, and AI's Infrastructure Bill Comes Due
Anthropic contains internet-connected agents, Google consolidates enterprise work around Gemini, Jev challenges text-first automation, and AI's physical and financial constraints sharpen.
Anthropic Mobilizes Cyber Defense, Britain Puts Agents on Notice, and TSMC Maps the Compute Boom
Anthropic mobilizes AI for critical infrastructure and science, Britain tightens agent privacy scrutiny, TSMC quantifies compute demand, and Manus funds an independent future.
GPT-6 Rebuilds the Interface, Google Makes Agents Universal, and AI Governance Gets Operational
GPT-6 generates interfaces, Google unifies enterprise agents, cyber access gets tiered, AI media gains provenance, healthcare regulation moves to lifecycle monitoring, and local AI PCs arrive.
GPT-6 Rebuilds the Interface, Mistral Scales Sovereign Compute, and Cyber Access Splits Into Tiers
GPT-6 turns answers into interfaces, Mistral trains a trillion-parameter sovereign model, Anthropic tiers cyber access, and enterprise AI shifts from pilots to operating systems.
Page 1 of 19