The Archive · May 2026
May 2026
25 entries from May 1 to May 25, 2026.
25Entries
3Unique tags
90Companies
04Weeks of notes
May 1 — May 25, 2026
- 25 MAYMSCI's global momentum index posts its strongest two-month outperformance on record (17pp over ACWI since end of March) on AI-infrastructure names, as Epoch AI data shows memory has grown to ~63% of AI chip component costs — reframing the buildout bottleneck from fab capacity to HBM-and-CoWoS — while John Jumper's pivot from AlphaFold-style science AI to general coding work at Google reads as bifurcation, not absorption, of the AI-for-science thesis.ai-digestdailyai-news
- 24 MAYAnthropic enters early talks to rent Microsoft Maia 200 inference chips on top of its existing AWS/Google/NVIDIA footprint, [[DeepSeek]] formalises its 75% V4-Pro discount as permanent pricing, and UC Berkeley Law institutes a near-total AI ban for graded coursework — running counter to the T-14 majority moving toward mandatory AI training.ai-digestdailyai-news
- 23 MAYAnthropic ships the first public progress report on Project Glasswing, its interpretability and alignment research initiative — the post hits the HN front page with 371 points and sustained technical discussion.ai-digestdailyai-news
- 22 MAYClaude Code ships v2.1.147 with background sessions and a tunable /code-review, then patches it five hours later with v2.1.148 to fix a Bash exit-code-127 regression — while a heavy HuggingFace paper day lands π-Bench, ACC trajectory compilation, and Gated DeltaNet-2 in a single drop.ai-digestdailyai-news
- 21 MAYOpenAI files confidentially for a September IPO at a ~$850B private valuation the same day Nvidia beats and raises but the hyperscaler-ASIC narrative finally bites the stock.ai-digestdailyai-news
- 20 MAYAndrej Karpathy joins Anthropic's pre-training team the same day Google counters at I/O 2026 with Gemini 3.5 Flash, a 24/7 'Gemini Spark' standing agent, and a $7.99 consumer AI tier — while Anthropic ships MCP tunnels and self-hosted sandboxes for Managed Agents at Code with Claude London.ai-digestdailyai-news
- 19 MAYAnthropic centerstage — a $300M-plus Stainless acquisition pulls SDK infrastructure in-house the same day Claude Mythos's cyber-flaw cache lands in front of the Bank of England and the IMF, with the Musk v. OpenAI jury verdict closing the day's most-watched courtroom arc.ai-digestdailyai-news
- 18 MAYApple's standalone Siri app — partly powered by Gemini through Private Cloud Compute — formalises the post-OpenAI integration era as the Musk v. Altman jury begins deliberations.ai-digestdailyai-news
- 17 MAYOpenAI announces Malta as the first ChatGPT Plus national-distribution partnership under its 'for Countries' program — frontier-lab distribution becoming statecraft.ai-digestdailyai-news
- 16 MAYFirst federal labor data shows two straight years of AI-exposed occupations underperforming the broader US job market — but the divergence pre-dates ChatGPT, so the causal attribution remains contested.ai-digestdailyai-news
- 15 MAYCerebras opens trading 89% above its $185 IPO price and closes +68% on a $5.55B raise, with OpenAI holding warrants for ~11% of the company tied to a $20B+ multi-year compute purchase commitment — the year's largest AI chip IPO is structurally underwritten by a single buyer.ai-digestdailyai-news
- 14 MAYAnthropic is in early talks for a fresh raise at a $900B+ pre-money valuation, three months after closing its $30B Series G at $380B post-money — a second magnitude jump from the same lab inside a single quarter.ai-digestdailyai-news
- 13 MAYThinking Machines Lab debuts its first model — TML-Interaction-Small, a 276B-parameter MoE designed for sub-half-second voice-and-video interaction — framed as a structural critique of OpenAI Realtime's scaffolded approach to interruptions.ai-digestdailyai-news
- 12 MAYGoogle Threat Intelligence Group publishes the first publicly attributed criminal use of an AI-built zero-day exploit, identifying telltale LLM authorship artifacts in a 2FA bypass aimed at an open-source web admin tool.ai-digestdailyai-news
- 11 MAYAlphabet raises 2026 capex guidance range to $180–190B (from $175–185B) and prepares its debut yen-denominated bond as the AI infrastructure financing story moves from 'is this circular?' to 'what does the funding stack actually look like?'ai-digestdailyai-news
- 10 MAY[[NVIDIA]]'s 2026 AI equity commitments cross $40B as the circular-financing critique becomes mainstream-analyst consensus; Box Elder County approves the 9 GW Stratos AI data center over a withdrawn water-rights filing and a planned referendum; and Fields Medalist Tim Gowers reports ChatGPT 5.5 Pro solving previously-open math research problems unaided in under an hour.ai-digestdailyai-news
- 09 MAYAnthropic locks $1.8B / 7-year Akamai compute capacity (Akamai's largest contract ever) the same week as the xAI Colossus 1 lease, while Cloudflare ships its first mass layoff in 16 years (1,100, ~20%) on a record-revenue earnings call and explicitly blames agentic AI.ai-digestdailyai-news
- 08 MAYAnthropic leases the entirety of xAI's Colossus 1 — 222k GPUs and 300+ MW — to relieve Claude serving-capacity strain, while Claude Code re-accelerates with five releases in four days and refutes the 'three quiet weeks' hypothesis.ai-digestdailyai-news
- 07 MAYApple confirms iOS 27 will let users swap Claude, Gemini, and other third-party AI models into Siri, Writing Tools, and Image Playground via a new Extensions framework — a structural opening for a historically closed-garden platform, set for fall 2026.ai-digestdailyai-news
- 06 MAYSamsung joins TSMC at $1T market cap on AI memory demand the same week OpenAI confirms $50B 2026 compute opex and Google / Microsoft / xAI sign formal CAISI evaluation agreements — three reads on the same week's AI infrastructure and governance arc.ai-digestdailyai-news
- 05 MAYOpenAI and Anthropic both announce PE-backed enterprise deployment ventures on the same day, with Sierra's $950M raise rounding out a $12B+ trifecta of enterprise-AI capital announcements.ai-digestdailyai-news
- 04 MAYAnthropic ships Claude Security GA and creative connectors alongside new sycophancy research, as hyperscaler AI capex clears $700B for 2026.ai-digestdailyai-news
- 03 MAYKKR closes $10B+ for Helix Digital Infrastructure under ex-AWS chief Adam Selipsky as Anthropic ships Claude Code Security in beta to Enterprise and Mistral launches cloud-resident coding agents — three different bets on what 'AI infrastructure' means in practice.ai-digestdailyai-news
- 02 MAYPentagon signs eight-company classified-network AI deals while pointedly excluding Anthropic, the same week Fed Vice Chair Bowman flags Mythos and Project Glasswing as new supervisory territory for banking regulators.ai-digestdailyai-news
- 01 MAYAnthropic is reportedly fielding a $50B round at valuations up to $900B as Meta lifts 2026 capex guidance to as much as $145B — and OpenAI's GPT-5.5 Cyber follows Anthropic's Mythos into gated rollout, hardening a two-lab convergence on pre-deployment security gating.ai-digestdailyai-news