The Tech‑Reader AI Digest for Tue May 26 2026
.jpeg)
The Tech‑Reader AI Digest
Tuesday, May 26, 2026
#AI #TechNews #Digest
Story 1: Three Phone Calls Canceled Trump's AI Executive Order
What happened: The AI executive order that the White House had been preparing for weeks never got signed. President Trump canceled the signing at the last minute after calls from Elon Musk, Mark Zuckerberg, and David Sacks. The draft would have had tech companies voluntarily submit their latest AI models to federal agencies for safety testing before release — with a review window of up to 90 days before public launch.
The sequence of events is now documented. According to Axios and Semafor, three calls reached Trump overnight on May 20 to 21. Musk for xAI. Zuckerberg for Meta. Sacks rounded out the trio. Sacks left his post as White House AI and Crypto Czar in late March 2026, but now co-chairs PCAST, the President's Council of Advisors on Science and Technology. The core argument boiled down to a single word: China.
Sacks warned that the measure would slow innovation and hurt the U.S. in its AI race. Both Musk and Meta subsequently disputed the accounts. Musk denied the claim on X: "This is false. I still don't know what was in that executive order and the President only spoke to me after declining to sign." Meta also said Zuckerberg had spoken to Trump only after the order was rescinded.
What the draft actually contained: a voluntary mechanism giving the federal government up to 90 days of access to the most powerful AI models before their public release. The explicitly stated that nothing in the order created a mandatory licensing requirement. The industry had been pushing for 14 days, not 90.
One additional dimension: OpenAI supported the executive order. The White House has reportedly given OpenAI a green light to pursue AI regulations at the state level instead. China, meanwhile, is accelerating its 2026 legislative work plan for comprehensive AI regulation.
Why it matters: The order that was killed was voluntary and non-binding. The episode establishes a clear precedent: in the current administration, even a lightweight voluntary safety framework can be neutralized before it reaches public deployment. The regulatory divergence between the two AI superpowers is now structural, and the U.S. framework will likely be decided state-by-state rather than federally.
Aaron's take — The collapse of this executive order means the 90-day federal speed limit is off. For developers and AI vendors, that means less friction and faster shipping in the short term. But the trade-off is regulatory fragmentation. With the White House stepping back and endorsing state-level initiatives, enterprise legal teams are going to have to navigate a fractured compliance map. The absence of a federal standard doesn't eliminate red tape; it just shifts the burden to local procurement teams.
Story 2: Anthropic Closes $30 Billion — World's Most Valuable AI Startup
What happened: Anthropic is closing its latest funding round this week. The round may top $30 billion at a valuation above $900 billion — vaulting ahead of rival OpenAI to become the world's most valuable AI startup. Sequoia Capital, Dragoneer Investment Group, Altimeter Capital, and Greenoaks Capital Partners are co-leading, each contributing approximately $2 billion. Existing investors including Founders Fund and General Catalyst are also participating.
The revenue trajectory driving the valuation: Anthropic's annualized run rate went from $87 million in January 2024 to $14 billion in February 2026, and $30 billion in April. Claude Code became generally available in May 2025 and reached $2.5 billion in annualized revenue by February 2026.
Anthropic expects to post $10.9 billion in revenue for the second quarter — more than doubling from the prior three-month period. The company is on pace for its first profitable quarter and has told investors that its annualized run rate revenue will surpass $50 billion by the end of next month.
Three of the four co-leads have prior ties to OpenAI — Dragoneer put nearly $3 billion into OpenAI last year, Sequoia has been an OpenAI backer since 2021, and Altimeter's Brad Gerstner has championed OpenAI publicly. Amazon and Google are not expected to participate in the current round.
The $900 billion pre-money valuation means the post-money figure will exceed $930 billion once the round closes — keeping Anthropic's October IPO trajectory intact as potentially the largest tech IPO in history.
Why it matters: $30 billion annualized run rate in April scaling to $50 billion annualized by the end of June. The first profitable quarter in Q2. The funding round validates the enterprise adoption crossover seen in the Ramp AI Index data earlier this month. The capital requirements to compete at the frontier are accelerating, and institutional investors are consolidating their bets.
Aaron's take — $87 million annualized in January 2024 to an expected $50 billion by the end of June 2026. The fact that three former OpenAI backers are co-leading this round highlights the intense capital requirements and shifting loyalties in the infrastructure race. An October IPO at a $930 billion post-money valuation would set a new high-water mark for the industry, moving AI firmly out of the venture capital phase and into public market reality.
Story 3: TrapDoor — The First Cross-Registry AI Poisoning Attack
What happened: Over the Memorial Day weekend, a cybersecurity incident emerged that security researchers are calling a first of its kind. TrapDoor — the first confirmed cross-registry supply chain attack — simultaneously hit npm, PyPI, and Crates.io, the three largest open-source package registries.
What distinguishes TrapDoor from previous supply chain attacks is its deliberate targeting of AI coding assistants. The malware was engineered to detonate across all three major registries simultaneously. Once installed, the payload plants .cursorrules and CLAUDE.md files containing hidden instructions embedded with zero-width Unicode characters.
These configuration files are automatically parsed by developer tools like Claude Code and Cursor to provide project-specific context. The attacker places invisible instructions in the path of the developer's AI assistant. When the assistant acts on those instructions, it silently executes credential-harvesting commands, scanning for AWS keys, GitHub tokens, and crypto wallet extensions.
CISA separately logged 15,000 attacks on a same-week Drupal SQL vulnerability — a separate incident that has stretched the security community's incident response capacity.
Why it matters: TrapDoor is the first documented case of a supply chain attack coordinated simultaneously across multiple registries specifically to poison AI workflow tools. Every developer who pulled dependencies from npm, PyPI, or Crates.io during the attack window was potentially exposed. It is a proof of concept for a new class of attack where the malicious code leverages the AI agent's local permissions rather than directly breaching the host machine.
Aaron's take — TrapDoor isn't just a supply chain attack; it's the first major AI assistant poisoning campaign. Attackers realize they don't need to breach your machine directly if they can just hide invisible Unicode instructions in a config file for your AI to execute. Security teams can no longer just scan libraries for known CVEs. They now have to treat .cursorrules and CLAUDE.md files as active execution vectors.
Quick Hits — The Rest of Today's AI World
Anthropic / Claude
- $30B round closing this week — see Story 2. $900B+ valuation. First profitable quarter in Q2. $50B annualized run rate by end of June. October IPO trajectory unchanged.
Gemini (Google)
- No new announcements today. Google I/O sessions on demand at io.google.
VS Code / GitHub Copilot
- Token-based billing June 1 — 6 days remaining. No new announcements.
Replit
- No new announcements.
Perplexity
- No new announcements today.
Microsoft Copilot
- Microsoft quietly cancelled its internal Claude Code pilot after token billing ate the entire annual AI budget — redirecting developers to Copilot. Exposes the enterprise risk of per-developer token caps.
Apple
- No new announcements. WWDC approaching — AI direction expected. OpenAI legal tension remains standing news.
Thinking Machines Lab
- No new announcements today.
xAI / SpaceXAI
- Musk denies involvement in killing Trump AI EO — see Story 1. SpaceX S-1 roadshow June 4 remains standing news. Ninth Circuit appeal remains standing news.
OpenAI
- OpenAI supported the AI executive order that was killed — see Story 1. Confidential S-1 filed Friday May 22. September listing target. Goldman Sachs and Morgan Stanley leading.
Meta
- Zuckerberg denies involvement in killing Trump AI EO — see Story 1. Meta disputes accounts.
Nvidia
- No new announcements. Record Q1 results remain standing news. Vera Rubin ramp Q3.
Cerebras
- No new announcements. Stock stabilizing post-debut.
Palantir
- No new announcements today.
Reflection AI
- No new announcements today.
Ollama
- No new announcements today.
DeepSeek / Alibaba Qwen / Z.ai
- No new announcements today. Chinese models 61% of global OpenRouter developer API usage — standing news from Friday.
Inflection Pi / Mistral
- No major news today.
That's your AI world for Tuesday, May 26. Back tomorrow. — Aaron
Aaron Rose is a software engineer and technology writer at tech-reader.blog.
Catch up on the latest explainer videos, podcasts, and industry discussions below.
