2026-08-05
By Vadym · Generated with AI, curated by me
Today’s mood is money and mortality — literally. Anthropic locked in a fresh multi-billion-dollar compute deal, Palantir’s earnings gave the AI trade its best hard evidence yet, and two of the field’s founding researchers are sharing a Las Vegas stage to argue about whether any of it ends well for humanity.
Anthropic committed $10 billion over six years to Volta, an Nvidia-backed cloud startup founded earlier this year, for compute from a 133-megawatt data center in Norway being built with crypto miner Bitdeer. It’s the latest in a string of capacity deals — alongside SpaceX and a $5B Amazon top-up — as Claude demand keeps outrunning supply.
Anaconda acquired Enkrypt AI, adding pre-deployment red-teaming across 300+ attack categories and NIST/EU AI Act compliance automation to its platform. In a two-month scan across 25,000 MCP servers, Enkrypt found 143,000 vulnerabilities affecting 73% of the servers tested. The deal follows Anaconda’s acquisition of Kilo Code last month.
At Ai4 2026 in Las Vegas today, Geoffrey Hinton and Andrew Ng are headlining a joint session with fundamentally opposed views: Hinton restates his case that current AI development risks catastrophic outcomes for humanity, while Ng argues that framing is being weaponized to slow competitors and hand regulators an excuse to entrench incumbents. Fei-Fei Li shares the stage as a third voice.
Palantir reported Q2 revenue of $1.94 billion, up 93% year-over-year and well above the $1.80 billion consensus, then raised full-year guidance to as much as $8.16 billion in sales. U.S. commercial revenue grew 149%. Shares jumped as much as 12% on the print.
The Cloud Security Alliance launched an AI Resilience Center of Excellence with Rubrik as founding partner, aimed at protecting, monitoring, containing, and recovering data and systems in environments run by autonomous AI agents. The announcement came alongside new AI Vulnerability Storm research and three new Vanguard members: Rubrik, Qualys, and Zscaler.
Every story today is really about the same tension: capital is flowing into AI infrastructure and enterprise adoption faster than anyone can verify it’s safe, or even solvent. Anthropic is locking in compute from a company that’s barely six months old, Palantir just proved the enterprise demand is real, Anaconda is buying up the tooling to catch what breaks, and two of the field’s most credible voices are publicly disagreeing about whether the foundation underneath all of it is sound. Nobody on that Vegas stage is wrong for the wrong reasons — they just disagree about how much time is left to find out.
— Boba
Curated by Vadym