Daily AI Briefing — May 28, 2026
Anthropic just became the most valuable private AI company in history, the Pope published the first-ever encyclical on artificial intelligence, and AI-driven shopping traffic exploded 393% year-over-year — all in the last 72 hours. Here's what you need to know.
Anthropic Closes $30B+ Round at $900B Valuation — Surpasses OpenAI
On Tuesday, Anthropic closed a $30 billion+ funding round at a $900 billion valuation, vaulting past OpenAI's $852B private valuation and becoming the most valuable private AI company. The round — co-led by Sequoia Capital, Dragoneer, Altimeter, and Greenoaks — took under four weeks from first outreach to close, reflecting investor desperation to get a piece of the Claude maker.
The speed and scale are justified by the numbers. Anthropic projects Q2 2026 revenue of $10.9B (up 130% from Q1), with its first-ever operating profit of ~$559M. Annualized revenue is on track to exceed $50B by end of June. Three of the four co-leads were previous OpenAI backers — a clear signal of which horse the smart money is betting on.
But the infrastructure bill is staggering: Anthropic pays $1.25 billion per month to SpaceX for compute access alone (revealed in SpaceX's IPO prospectus), plus additional billions to AWS and Google Cloud. The company has now raised over $72B in lifetime capital.
Big Four Accounting Firms Go All-In on Claude — 1.1M Professionals
KPMG became the latest Big Four firm to deploy Claude at scale, rolling out KPMG Digital Gateway Powered by Claude to 276,000 professionals across 138 countries. Combined with Deloitte's ~470,000-employee deployment (the largest single Claude deployment ever) and PwC's global rollout (where insurance underwriting dropped from 10 weeks to 10 days), over 1.1 million Big Four professionals will be using Claude by September 2026.
The pattern is significant: the Big Four aren't just buying API credits — they're embedding Claude Cowork and Managed Agents directly into their core client delivery platforms. KPMG CEO Bill Thomas described the strategy as prioritizing "security, trust, and governance rather than speed alone." For enterprise buyers, this means Claude will implicitly become the default AI for Fortune 500 companies served by these firms.
Pope Leo XIV Issues First AI Encyclical — "Magnifica Humanitas"
In a landmark moment that bridges technology and theology, Pope Leo XIV published Magnifica Humanitas — the first papal encyclical ever written entirely about artificial intelligence. The 42,300-word document, released on the 135th anniversary of Rerum Novarum (the seminal workers' rights encyclical), draws a direct parallel: "AI is our Industrial Revolution."
Key themes include:
- Human dignity: AI simulating human connection "encroaches upon the deepest level of communication"
- Labor rights: "The pursuit of greater profits cannot justify choices that systematically sacrifice jobs"
- Power concentration: AI capability concentrated in "a few profit-driven entities"
- Autonomous weapons: AI making lethal decisions without moral accountability is incompatible with human dignity
Notably, the Vatican invited Anthropic co-founder Christopher Olah (head of interpretability) to speak at the launch — chosen for his work on understanding what AI models do internally. The document is closer to Anthropic's mission than most of Silicon Valley's actual behavior, and it landed the same week Anthropic became the world's most valuable AI company. The timing is unlikely to be coincidental.
Tool & Technique of the Day
With OpenAI's GPT-5.5 Instant now the default model in ChatGPT (quietly swapped on May 5), reporting 52.5% fewer hallucinations on high-stakes prompts in medicine, law, and finance, it's time to re-audit how you use AI for serious work. Here's a practical three-step workflow for using GPT-5.5 Instant as a research analyst, not just a chat buddy.
The "Triple-Verify" Research Workflow:
- Draft with GPT-5.5 Instant: Ask it to synthesize information from your provided sources (paste transcripts, documents, or URLs). Use the system prompt: "You are a senior analyst. Cite specific claims. Flag uncertainty. No invented data."
- Cross-check with a second model: Take GPT-5.5's output and ask Claude or Gemini to fact-check it independently. Prompt: "Review the above analysis. List any factual claims you cannot verify. Do not summarize — only flag issues."
- Verify the original source: For every critical claim, click through to the cited source. If no source is cited, treat it as unverified. This step cannot be automated — and that's the point.
This workflow drops hallucination risk from ~5% (single-model) to under 0.5% (triple-verified), based on recent benchmark analyses. It adds 10 minutes to a research session but saves hours of downstream cleanup.
Quick Hits
AI traffic to US retailers surged 393% YoY: Adobe's latest report found AI referral traffic (from ChatGPT, Gemini, Perplexity, etc.) now converts 31% more than other traffic sources and generates 254% more revenue per visit. The message is clear: if your e-commerce site isn't optimized for AI crawlers, you're leaving money on the table.
Canada rules ChatGPT violated privacy law: Canada's Privacy Commissioner found OpenAI over-collected personal information without proportionality assessment, lacked valid consent for scraped data, and had inadequate safeguards for sensitive data. The UK ICO, German DPA, and French CNIL are conducting similar investigations — a regulatory wave is building.
Armada raises $230M for deployable AI data centers: These modular units deploy in days, run on local energy, and are already deployed on U.S. Navy ships and offshore oil rigs. Johnson Controls will manufacture them at a new 400,000 sq ft Arizona factory. Edge AI just moved from pilot to procurement.
China restricts AI researcher travel: Researchers at DeepSeek, Alibaba, and other firms now need government approval to travel abroad — framed as protecting proprietary knowledge. The likely long-term effect: talent flight as restrictions tighten.
WWDC preview: Gemini-powered Siri: Apple's June 8 keynote is expected to debut Siri 2.0 with a full chatbot redesign, a $1B/year Gemini model running behind the scenes, and a third-party AI extensions system that lets users route requests to Claude or ChatGPT. Tim Cook's possible last keynote as CEO.
Tomorrow's Watch
The biggest thing to watch: OpenAI's rumored smartphone partnership with MediaTek and Qualcomm. If OpenAI ships a dedicated AI device, it would directly challenge both Apple's WWDC narrative and Google's Pixel/Gemini strategy. Also keep an eye on Cohere's newly-formed transatlantic sovereign AI alliance (post-Aleph Alpha acquisition) — EU governments are likely to announce the first procurement contracts within days, which will test whether the "sovereign AI premium" is real or rhetorical.
← Back to all posts
