The 12-month window

The piece frames the threat as capability expansion by foundation model labs, but the more immediate pressure may be distribution. OpenAI's acquisition spree, covered here just two days prior, suggests the window closes not only when models get smarter but when incumbents buy the surface area startups were counting on.

Factory's $1.5B raise and Upscale AI's reported $2B valuation round (both covered April 16) look different in this light: investors are betting those companies can build enough operational depth before the window narrows, not that the window stays open indefinitely. The MIT Technology Review piece from the same week makes a complementary argument, that enterprise AI advantage lives in operational infrastructure rather than model capability, which is essentially the bull case for every startup playing in a foundation model gap. InsightFinder's $15M raise for AI observability fits the same thesis: if you own the diagnostic layer, you may survive even after the underlying model gap closes.

Watch whether any of the well-funded gap-players, Factory being the clearest example, announce enterprise contracts that explicitly displace a foundation model provider's native tooling within the next two quarters. That would be the first concrete signal that operational depth is actually converting into defensibility before the window moves.