New Madrona research based on 150 enterprise IT professionals finds 74% plan to expand AI budgets in the next 12 months, yet fewer than half of AI pilots reach full production and 77% of enterprises reevaluate AI vendors every six months or on a rolling basis. A companion a16z survey of 50 AI buyers finds most want pricing tied to outcomes rather than token usage, indicating pricing models are still unsettled. The implication is that post-pilot enterprise contracts no longer function like traditional multi-year SaaS commitments, so reported ARR growth rates may not reflect durable revenue. For financial institutions this weakens the reliability of ARR-based valuations and revenue backlog claims at AI application-layer startups.
256 other tracked developments touching the same claims.