JadidJust-in AI Developments, Incidents & Deadlines
May 2026
SMTWTFS12345678910111213141516171819202122232425262728293031

AI news, Friday, May 29, 2026

8 stories

TLDR

  • Anthropic closes a $65B Series H at a $965B post-money valuation — now the most valuable private AI company, ahead of OpenAI and xAI (May 28) Trending — Official confirmation and final numbers for the round flagged as preliminary in yesterday's digest ($30B / $900B). Run-rate revenue crossed $47B this month. (Anthropic | TechCrunch)
  • IBM and Red Hat commit $5B to "Project Lightwell," an AI-powered clearinghouse to find and fix open-source vulnerabilities at scale (May 28) Security Open Source — 20,000+ engineers plus frontier AI; early adopters include Goldman Sachs, JPMorganChase, Citi, Visa, Mastercard. (Red Hat | SiliconANGLE)
  • OpenAI publishes its Frontier Governance Framework, mapping safety/security practices to California's Transparency in Frontier AI Act and the EU AI Act GPAI Code of Practice (May 28) Policy — A public template for the compliance documentation regulators will start demanding. (OpenAI)
  • California's ~30 AI bills clear their chamber-of-origin crossover deadline today (May 29) (May 29) Policy — Survivors include chatbot-disclosure (AB 1609), automated employment decisions (SB 947), and workplace surveillance (AB 1883). Final votes due before July 2 adjournment. (Transparency Coalition)
  • AI-linked layoffs accelerate: Meta (~7,000), Cisco (4,000), and Cloudflare (1,100+) cut staff in May as firms restructure around AI (May 2026) Hiring — ~50,000 of 2026's ~300,000 announced tech cuts are explicitly AI-driven. (Yahoo Finance | CBS News)

Action Items

Things your engineering team should look at or do this week.

  • Map your OSS supply-chain exposure against Project Lightwell's scope before deciding whether to buy in. Lightwell delivers AI-validated patches through commercial subscriptions and explicitly builds on the Anthropic Mythos OSS-vulnerability findings already in your backlog. Inventory which critical dependencies you currently patch manually and whether a clearinghouse subscription would close the gap faster than your own pipeline. (Red Hat)
  • If you ship a GPAI model or AI product into California or the EU, use OpenAI's Frontier Governance Framework as a gap-check template. It explicitly maps practices to the CA Transparency in Frontier AI Act and the EU AI Act GPAI Code of Practice — both now near enforcement. Compare your own risk-assessment, model-reporting, and incident-response documentation against it. (OpenAI)
  • Customer-facing teams: track California AB 1609 (chatbot disclosure). It passed the full Assembly May 27 and is now in the Senate; if it becomes law it will impose disclosure obligations on customer-service chatbots. Flag any production chat surfaces that would need a disclosure layer. (Transparency Coalition)

Security

  • IBM + Red Hat "Project Lightwell" — $5B AI-powered open-source security clearinghouse (May 28) Open Source Business — A "trusted enterprise clearinghouse" that uses frontier AI plus 20,000+ engineers to identify, validate, and ship fixes across open-source code, delivered via commercial subscription so enterprises can pull vetted patches into existing supply chains. The announcement explicitly cites Anthropic's Mythos Preview (~3,900 high/critical OSS bugs) as evidence of the problem's scale. Early adopters are nearly all major banks — Bank of America, BNY, Citi, Goldman Sachs, JPMorganChase, Mastercard, Morgan Stanley, RBC, State Street, Visa, Wells Fargo. (Red Hat | IBM Newsroom | Help Net Security)

Industry & Business

  • Anthropic raises $65B Series H at $965B post-money — passes OpenAI and xAI as the most valuable private AI company (May 28) Trending — This is the official close of the round reported preliminarily yesterday (then framed as ~$30B at ~$900B). Co-led by Altimeter, Dragoneer, Greenoaks, Sequoia, Capital Group, Coatue, D1, GIC, ICONIQ, and XN, with $15B of previously committed hyperscaler/strategic money (incl. $5B Amazon, plus Micron, Samsung, SK hynix). Run-rate revenue crossed $47B this month; widely read as the last private round before an expected IPO window. (Anthropic | TechCrunch | Crunchbase News)
  • Saris raises $28.8M Series A (led by 8VC) for agenticDescribes AI systems where a model uses tools and decides its own steps instead of following a fixed script.Anthropic ↗ back-office automation in banks and credit unions (May 28) Agents — Agents trained on each institution's own workflows, running under human supervision; claims automation of up to 70% of consumer/mortgage/commercial lending tasks with ~35% cost reduction. Integrations with Fiserv, Encompass, and MeridianLink. A clean example of the "agents inside regulated, expensive workflows" thesis dominating current AI funding. (BusinessWire | Axios)
  • AI-linked layoffs accelerate across big tech (May 2026) Hiring — Meta (~7,000), Cisco (4,000), and Cloudflare (1,100+, while disclosing internal AI usage up 600% in three months) all cut staff this month. ~50,000 of 2026's announced cuts are explicitly tied to AI restructuring; the quieter signal is weakening demand for junior and entry-level roles. (Yahoo Finance | CBS News)

AI in Production

  • Anthropic platform wobbles: Console/claude.ai billing errors (May 28) and elevated Claude Opus 4.8 errors (May 29), both resolved (May 28–29) — Chat and API traffic were unaffected during the billing-management incident; the Opus 4.8 error spike was short-lived. Worth a glance if you saw transient 5xx on Opus 4.8 calls Friday. (Claude Status)

Policy & Safety

  • OpenAI publishes Frontier Governance Framework (May 28) Trending — A public governance document that takes the existing Preparedness Framework and maps it onto specific regulatory obligations: risk assessment/mitigation across cyber, CBRN, harmful manipulation, and loss-of-control, plus model reporting, security risk management, incident response, and external expert input. Explicitly aligned to California's Transparency in Frontier AI Act and the EU AI Act GPAI Code of Practice. (OpenAI)
  • California AI bills clear the May 29 crossover deadline (May 29) — Nearly all of ~30 AI-related bills passed their chamber of origin. Notables still alive: AB 1609 (customer-service chatbot disclosure, passed Assembly May 27), SB 947 (automated employment decision systems), AB 1883 (workplace surveillance tools), and AB 1651 (AI in bar-exam administration). Final passage must happen before the July 2 summer adjournment. (Transparency Coalition)

Tutorial / How-To

  • OpenAI: "Building self-improving tax agents with Codex" (May 28) Agents — A walkthrough of using Codex to build agents that iteratively refine their own task performance — a concrete pattern worth reading if you're prototyping domain-specific agent loops, regardless of the tax framing. (OpenAI News)

Type to search · ↑↓ to move · enter to open