AI has gifted the enterprise world a parade of never-before-seen moments, and the latest one is a real doozy: companies that used to cling to their software vendors like barnacles to a ship are now treating AI startups like a fling - fun for a bit, then reevaluated every six months. Market researcher IDC predicts these fickle enterprises will splash out $4.25 trillion on technology in 2026, almost all of it AI-driven. But as new research from venture capital firm Madrona shows, they're not exactly committed.

Madrona surveyed 150 enterprise IT professionals and found that 74% plan to expand their AI budgets in the next 12 months, with the rest holding steady. Good news, right? Well, these same enterprises admit that fewer than half of their AI pilots ever make it to full production. That's actually an improvement: last year, MIT famously reported that 95% of enterprise AI projects failed on ROI. So 'fewer than half succeeding' is a low bar, but hey, it's progress.

The real kicker from Madrona's report is that even when enterprises do adopt AI tech, they don't commit long-term. A whopping 77% re-evaluate their AI vendors every six months or on a rolling basis. As Madrona puts it, this creates a 'fast in, fast out' dynamic that's fundamentally different from traditional enterprise SaaS, where multi-year contracts provided a 'moat of inertia.' In enterprise AI, switching costs are lower, and the re-evaluation cadence is relentless.

This has massive implications for those astronomical ARR numbers startups love to flaunt. The initial AI boom of 2025 was fueled by enterprise trial budgets. This year was supposed to be when big customers settled in and committed. Those contracts allow startups to claim revenue growth like $0 to $10 million in three months. But for the first time ever, enterprise revenue remains insecure even after an AI product graduates from pilot purgatory.

Part of the problem? Many AI startups haven't figured out how to price their wares. New research from VC firm Andreessen Horowitz, which surveyed 50 technical AI buyers, found that more than half want fees tied to outcomes - like work produced - rather than usage, like tokens consumed. Charging per token is so SaaS-era. Once an enterprise knows it needs email or HR software, it's just a matter of headcount. But for AI, pricing around 'recognizable work' helps startups prove their worth. When fees revolve around reports processed, tickets closed, or leads generated, the product becomes 'economically valuable to both sides,' write a16z partners Tugce Erten and Sarah Wang.

All this suggests we've entered a new era of enterprise experimentation. That's great for startups - enterprises are more willing to try their tech. But it also means an enterprise contract no longer guarantees long-term revenue. Whether enterprises will ever revert to their old, committed buying habits is anyone's guess. For now, startups might want to enjoy the ride while it lasts.