The AI bubble can wait, but your exit plan can't

Your AI projects could be disrupted if key vendors face financial trouble or cease support in the next year.

Spending on AI infrastructure is outpacing measurable returns, according to Forrester Research. While fears of an immediate AI bubble burst are premature, analyst Brian Hopkins notes that approximately 90% of current AI computing infrastructure investment comes from large, financially stable tech companies using their own cash flow. The primary risks are concentrated among a few financially strained builders and circular financing arrangements with AI labs. Forrester's AI Bubble Model suggests that AI needs to boost US GDP by 5% by 2035 to justify current infrastructure investments, with only about two percentage points projected from productivity gains. This necessitates finding new value streams beyond enhanced employee productivity, as echoed by Bain & Company's estimate that the AI industry needs to generate $6 trillion in annual revenue by 2031. Companies are advised to monitor the financial health of AI infrastructure suppliers, paying close attention to their funding models and revenue recognition, and to stress-test their AI plans against potential supplier failures.

The rapid build-out of AI infrastructure, funded by significant capital expenditure from hyperscale operators, is not yet justified by demonstrable productivity gains. This disconnect suggests a potential market correction where suppliers might reduce investment, merge, alter pricing, or discontinue services. Enterprises must proactively assess the financial stability of their AI vendors and plan for contingencies, such as service discontinuations or SLA reductions, to avoid project disruptions. The core challenge lies in ensuring AI's long-term value creation extends beyond simple productivity enhancements to meet the substantial revenue demands required to sustain current infrastructure investments.

Forrester's analysis highlights that vendor reliance on external investment and customer prepayments, rather than consistent revenue recognition, serves as a critical red flag. This suggests that customers need to scrutinize their AI suppliers' financial structures and business models closely. The risk of oversupply also looms, which could exacerbate price volatility and further destabilize the market. Consequently, organizations should develop robust exit strategies and be prepared to migrate workloads and data should a critical vendor experience financial distress, thereby protecting their investments and ongoing AI initiatives from unforeseen market corrections.