Startup ARR is increasingly fragile, new research warns
Original: Startup ARR is less secure than ever, new research shows
Why This Matters
Recurring revenue instability could force AI startups to rethink growth metrics and pricing models industry-wide.
New research from Madrona and a16z reveals that 77% of enterprises reevaluate AI vendors every six months, undermining the long-term contract stability that once secured startup ARR. IDC forecasts $4.25 trillion in enterprise tech spend in 2026, nearly all AI-driven.
Venture capital firm Madrona surveyed 150 enterprise IT professionals and found that 74% plan to expand AI budgets over the next 12 months, with the remainder holding spending steady. However, fewer than half of AI pilots reach full production — an improvement over MIT's 2025 finding that 95% of enterprise AI projects failed on ROI, but still a low success rate.
The most consequential finding: 77% of enterprises reevaluate their AI vendors every six months or on a rolling basis. Madrona describes this as a 'fast in, fast out' dynamic that contrasts sharply with traditional enterprise SaaS, where multi-year contracts created durable switching costs. 'In enterprise AI, switching costs are lower and the re-evaluation cadence is relentless,' the firm writes.
This directly threatens the ARR numbers AI startups have used to demonstrate explosive growth — some reporting $0 to $10 million in revenue within three months. A separate a16z survey of 50 technical AI buyers found that more than half prefer outcome-based pricing (e.g., reports processed, tickets closed, leads generated) over usage-based models like token consumption. Partners Tugce Erten and Sarah Wang argue that outcome-based pricing makes AI products 'economically valuable to both sides.' Together, the data suggests enterprise AI revenue remains structurally insecure even post-pilot.