A general-platform story gets marked down directly; the money goes to vertical agents that clear four bars: deep vertical focus, a customer money pain, production-grade reliability, and proprietary data plus deep integration. Agentic AI raised $2.66B this year through April versus $1.09B a year earlier (publicly disclosed rounds only).
The mainstream story is that one general agent platform wins everything. The diary parts from it on where the money went: funding ran only to coding agents and customer service, and the general harness sits between model companies (Mythos has already moved into cybersecurity) and vertical agents, the most dangerous spot for valuation. The author admitted the same day to having been too optimistic on it.
Agentic AI raised $2.66B this year through April versus $1.09B in the same period last year; average round size from Q4 2025 to Q1 2026 was $155M (diary of June 3, 2026; publicly disclosed rounds only, undisclosed rounds excluded, an incomplete year-to-date window). Jiazi Guangnian counted 130 large-model exhibitors at WAIC, of which 83 switched their label to Agent (diary of July 18, 2026; one outlet's count of exhibitor labels, not an industry census, read for trend and not precision).
For founders: pick one vertical and go deep, give evidence in the deck for each of the four bars, and be ready to answer "what will your exhibitor badge say in two years". For investors: put general harnesses on a six-month watch, mark down any that have not moved into a specific business within those six months, and price demos that fail the four bars as hype.