2026-05-27
By Vadym · Generated with AI, curated by me
AI capital is concentrating faster than at any point in tech history — one company closes a $900B valuation round while a single quarter surpassed all of 2025’s funding combined. At the tooling layer, the repricing is already underway.
Anthropic closed a $30B+ funding round this week — its second of that size in a single calendar year — pushing its valuation above $900B and past OpenAI’s $852B March mark. Sequoia, Dragoneer, Greenoaks, and Altimeter each committed roughly $2B. The company reported its first quarterly operating profit and projects $10.9B in Q2 revenue. Two $30B rounds in twelve months is a pace that has no precedent in private markets, and the first operating profit means Anthropic is no longer just burning capital — it has a real business underneath the spending.
GitHub paused new sign-ups for Copilot Pro, Pro+, and Student plans in late April and confirmed a full shift to usage-based AI Credits billing starting June 1. Opus-class models are being removed from the Pro tier, staying only in Pro+. Business and Enterprise plans are unaffected. The message is clear: predictable flat-fee access to frontier models was a loss leader, and GitHub is correcting the pricing before AI inference costs consume the product’s margin. Developers who built budget assumptions around Copilot’s flat fee now face variable costs with no obvious ceiling.
xAI released Grok Build 0.1 in early access this month — a purpose-built coding model trained specifically for agentic task execution rather than conversation. It follows Grok 4.3, which shipped with 1M-token context, native video input, and reasoning built in at $1.25/M input tokens. xAI also released Grok Skills (persistent custom expertise profiles) and expanded MCP Connectors to GitHub, Linear, Notion, and Google Workspace. Elon Musk’s lab is now competing in every segment of the agentic coding stack — coding model, context window, tool integrations — that Copilot and Windsurf built their categories on.
The 2026 Robotics Summit & Expo runs today and tomorrow in Boston with over 6,000 developers focused on physical AI. NVIDIA’s Isaac GR00T open models — enabling robots to understand natural language and reason through multistep tasks using vision-language-action pipelines — are the headline technology on the show floor. Open Robotics is presenting how the developer ecosystem is reorganizing for the AI era. Open-sourcing GR00T means any developer building on NVIDIA hardware can now access the same VLA reasoning stack that was research-only six months ago — the barrier to entry for physical AI just dropped.
Global venture investment hit an all-time $300B in Q1 2026, with AI capturing $255.5B — surpassing 2025’s full-year AI total of $254.4B in a single quarter. Four of the five largest venture rounds in history closed in Q1. OpenAI ($122B), Anthropic ($30B), xAI ($20B), and Waymo ($16B) accounted for 65% of all global VC. Three companies took 67% of all AI investment. This isn’t a broad ecosystem rising — it’s a narrow set of bets getting dramatically larger while the rest of the startup stack competes for what’s left of the oxygen.
OpenAI launched its Deployment Company this month — a services entity backed by $4B, staffed with Forward Deployed Engineers to help enterprises build and scale AI systems from day one, and bolstered by the acquisition of Tomoro for deployment expertise. Alongside this, OpenAI is preparing to file its IPO S-1, moving the world’s largest AI lab toward public markets for the first time. When OpenAI goes public, it sets a benchmark that every other AI company will be measured against — and public shareholders will have opinions about safety timelines and capital allocation that private backers have been content to leave to management.
Three companies captured 67% of all AI venture investment in Q1 — and one of them just added another $30B on top of a $30B round. That pace of capital concentration has no precedent in tech history. What it produces at the tooling layer is less obvious than it looks: GitHub is repricing because inference is expensive, xAI is entering every coding category at once, and NVIDIA is opening physical AI to any developer with compatible hardware. The window where a single company could dominate end-to-end — model to tooling to deployment — may already be closing. OpenAI going public will be the first real test of whether public markets price that risk correctly.
— Boba
Curated by Vadym