AI Mega-Funding Week: The Fast Brief on Where the Money Went

Four raises defined the week. Valuations and ARR are as reported by the companies and press.

Legora landed "virtually every top-tier VC in the cap table." Prime Intellect is already running ARR over $100M. Rhoda AI unveiled FutureVision, a robotic-intelligence system built on video-predictive control, after 18 months in stealth. Mecka AI builds no robots at all — it collects and processes human daily-action data to train robotics models, with annualized revenue expected to reach ~$100M. The common thread: three of the four raises fund the tooling and data that sit around the robots and agents, not the flashy end product.

Two things stand out. First, the money is moving up the stack — out of raw models and into agents, robotics data and applied AI, the layer that sits on top of ChatGPT and Claude. Second, the geography is lopsided: Crunchbase reported ~88% of 2026 AI-startup funding went to US-headquartered companies (and >80% of all global startup funding), while the very biggest checks still cluster in a tiny set — OpenAI, Anthropic and xAI.

Crunchbase reported that ~88% of AI-startup funding in 2026 went to US-headquartered companies, and more than 80% of all global startup funding went to the US, with the largest checks concentrated in OpenAI, Anthropic and xAI (Crunchbase, via Crescendo AI).

Watch the deployment bet. OpenAI's new Deployment Company agreed to acquire Northslope — its second applied-AI acquisition since May — as part of a reported ~$4B commitment to forward-deployed engineering. The wager: implementation expertise will matter as much as raw model quality. If it pays off, expect more capital to keep flowing to the companies that put AI to work — the deployers, agent platforms and data-labour plays — not just the handful of labs that build the frontier models. This one is still developing.