A Sequoia partner warns the AI gold rush is built on shaky, 'experimental' revenue (www.businessinsider.com)

Sequoia partner Alfred Lin warned that much of the current AI boom is funded by "experimental revenue" — short-term pilots and one-off deals that founders are sometimes annualizing and presenting as recurring ARR. He said these pilots help finance R&D but may evaporate once experimentation ends, so headline revenue can be misleading if customer retention and true repeatability aren’t proven. Lin urged investors to prioritize revenue quality and the velocity of core metrics (engagement, retention, expansion) over raw growth figures, arguing slower but high-quality revenue is preferable to fast, hype-driven numbers. The caution matters because AI startups are raising record capital and valuations, prompting comparisons to past bubbles. Technically, the warning targets how ARR is computed (multiplying a pilot month's revenue by 12), the importance of churn/retention as a signal of product-market fit, and the need to dig into unit economics and customer lifecycle metrics rather than rely on topline growth. For VCs and founders that means focusing diligence on repeatable purchasing behavior, sustained usage signals, and expansion potential — otherwise "experimentation" can mask fragile business models even amid exuberant market valuations.