Instinct raises $1 billion at $10 billion valuation

The article frames the Instinct deal as an open question: whether $10 billion marks the peak of an AI bubble or the start of a genuinely new kind of company. Its stated answer is that the valuation itself cannot be assessed, because Instinct has disclosed neither revenue nor user numbers — the same problem as WhatsApp's $19 billion price in 2014, which rested on users rather than sales. Where the article does commit is on the question of whether a new company shape has begun: it says yes, and points not to valuations but to how work is divided at small companies that do earn revenue.

WhatsApp kept its headcount near 50 by choosing a business — messaging without advertising — in which sales and marketing departments never arise. Gamma, Bolt, Cursor and Cognition took the opposite path, operating businesses where such work does arise and handing it to AI instead. Gamma left 75% of inquiries to AI; Bolt reached $20m ARR in 60 days with 15 people; Cursor wrote 35% of merged code with AI agents; and at Cognition, Devin committed 89% of engineers' code. The article's cautionary counterpoint is Klarna, which in May 2025 admitted that prioritizing cost over quality hurt service and resumed hiring humans. On the cost side, the article notes that AI companies pay for computing power where software companies paid salaries: a16z estimated generative AI app companies spend roughly 20–40% of revenue on inference and per-customer tuning, and Bessemer's 10 'supernova' companies averaged $1.13m ARR per employee but only 25% gross margins, often negative. Instinct's founder Noah Shinn says he spends 40% of his time thinking about compute. The article's conclusion is that the new company shape is designed around where departmental work goes, not how many people are hired — with humans moving from volume work to deciding, judging, and handling exceptions.