Startup ARR is now less secure than at any point in the recent past, according to new research examining how enterprises buy and retain artificial intelligence products. The findings signal a shift in the economics behind the fast revenue growth many AI startups have reported.
The backdrop is a historic surge in corporate technology spending. Enterprises are projected to spend EUR 4 trillion on technology in 2026, a figure driven almost entirely by AI. That spending appetite is not slowing: a survey of 150 enterprise IT professionals found that 74% plan to expand their AI budgets over the next 12 months, while the remainder intend to keep spending flat.
AI pilots are improving, but adoption remains fragile
Despite the enthusiasm, enterprises report that fewer than half of their AI pilots ever reach full production. That is actually an improvement over prior results, when 95% of enterprise AI projects were reported to have failed on return on investment. A success rate under 50% is a low bar, but it marks progress from a 5% rate.
The more revealing detail is what happens after deployment. Even when an enterprise rolls out an AI product, it rarely commits for the long term. Roughly 77% of enterprises reevaluate their AI vendors every six months or on a rolling basis. The result is a “fast in, fast out” dynamic that departs sharply from traditional enterprise SaaS, where multi-year contracts created a moat of inertia. In enterprise AI, switching costs are lower and the reevaluation cadence is constant.
Why enterprise revenue is no longer locked in
This has broad consequences for the annual recurring revenue (ARR) figures startups tout. Enterprise trial budgets fueled the initial AI boom of 2025, and 2026 was expected to be the year large customers settled in and committed long term. Enterprise contracts are what let so many AI startups claim extraordinary revenue growth, including companies moving from $0 to EUR 9 million in three months.
For the first time, however, enterprise revenue stays insecure even after an AI product graduates from a pilot and gets adopted. Part of the problem is pricing. A separate survey of 50 technical AI buyers found that more than half want AI fees tied to the work produced or other outcomes, rather than to usage metrics such as the number of tokens consumed.
The move toward outcome-based pricing
Charging by usage, such as tokens, echoes the SaaS-era model, where a company that needs email, HR software, or cloud storage simply pays based on employees or data volume. AI is different. Pricing “around the recognizable work” is what helps a startup prove its value. When fees revolve around outcomes such as reports processed, tickets closed, or leads generated, the product becomes economically valuable to both sides.
Together, these findings point to a new era of enterprise experimentation. Companies are more willing to try new technology, which opens doors for startups, but an enterprise contract no longer guarantees long-term revenue. Whether enterprises eventually return to their older, more committed buying habits is still unknown.
Source
Image: techcrunch.com
