CoreWeave launched the CoreWeave Zero Trust AI Framework so customers can control access to models, data and runtime environments. The company also announced CoreWeave Physical AI Launchables, which are tested configurations designed to scale physical AI and autonomous vehicles faster.
Anthropic's S-1 leaked...again. At this point, Anthropic's IPO filing will look like a rerun it's been leaked so many times. For this go-round, The Wall Street Journal checked out Anthropic's IPO filing and found the company cold potentially spend $750 billion on compute in the years ahead. I'm sure those numbers will work out just fine.
OpenAI said two agents powered by GPT-6.1 Sol and GPT-6 Astra weakened misalignment monitoring while making changes during two internal coding tasks.
Google launched TorchTPU, an open source version of PyTorch that will run AI apps on TPUs and Google Cloud without rewriting code. That news followed AWS, which opened its Trainium AI chips to every PyTorch developer without a rewrite via the now open source TorchNeutron. The upshot: Google Cloud and AWS are removing any friction from moving workloads to their custom chips (and presumably away from Nvidia) to optimize for cost.
HPE launched its Spaceborne Computer for HPC and AI workloads on the Moon. The project is part of a NASA, Deloitte and HPE collaboration. HPE Spaceborne Computer-4 will power the first HPC benchmarks on the moon and test LLMs, vision inference and LiDAR data for deployment.
Microsoft added new security tools to Microsoft 365 E7 and its Project Perception system that treats AI agents as potential insider threats. Also see: Why every AI agent should be treated like an insider threat
Quantinuum said its Helios quantum computer hit 99.9995% logical fidelity, 10x lower than its September 2026 record. Quantinuum's Helios system used the company's error-correction architecture. Quantinuum reiterated that it was on track for Apollo, its universal fault-tolerant quantum computer expected int 2029.
Google DeepMind published research showing four Gemini 4 Argon agents in a sandboxed environment beat Minecraft in 27 hours and 14 minutes.
"Starting from a randomly generated world on Easy difficulty, the team wrote 2,783 programs and defeated the dragon in 27 hours and 14 minutes, dying 605 times along the way. To our knowledge, this is the
first end-to-end autonomous completion of Minecraft by AI agents, achieved using only the standard Mineflayer library without pre-scouted seeds or other game-specific tools written by humans. Winning required the agents to keep a coherent plan across thousands of dependent steps for more than a day."
Anthropic issued a fascinating report about the workplace where agents outnumber humans by a wide margin. In a nutshell, you're more productive, don't hire entry-level techies and you babysit AI agents all day. Sound rewarding? Didn't think so. For Anthropic, this report on the future of work is in part marketing to get you thinking about scaled AI adoption (the company assumes it's 18 months ahead of everyone else) and possibly get you to buy into its IPO and vision. This report may backfire as an HR recruiting tool. Also see: Anthropic employees and their take on their jobs in the future.
Intuit said it acquired Sequence, a startup focused on revenue management automation. Intuit said it will integrate Sequence into its Intuit Enterprise Suite and QuickBooks Online Advanced.
Thomas Bravo's Proofpoint will acquire Varonis Systems in a deal valued at $6.3 billion. Under the terms of the deal, Varonis shareholders will get $49.75 a share in cash, or 8.4x revenue. The combined company will have more than 16,000 enterprise customers.
Accenture is reportedly telling many employees that they can only spend $29 a month on Anthropic's Claude. Some of these employees are currently spending $1,00- a month. The memo outlining the new usage policy was obtained by 404 Media.
