Forrester Model Shows AI Capacity Outpacing Demand Absorption

Forrester’s Nov. 5 assessment says an AI bubble burst is not imminent, while the firm’s public model overview shows infrastructure supply pressure exceeded demand absorption in the second quarter of 2026. The readings frame a gap between capacity being built and demand signals showing up in business activity.

The model scored supply pressure at 4.95 on a five-point scale, compared with 3.44 for demand absorption, a 1.51-point spread. The nine-company financial-health basket scored 82.3 out of 100. Forrester places its green range above about 73, making this a group result rather than a statement about each company’s condition.

The model’s Nov. 2 nowcast put the basket at 79.1, down from 82.1 on Oct. 29. Its likely range was 77.7 to 81.4. The estimate used third-quarter filings from seven of the nine companies and share prices through Nov. 2; it was an updated estimate, not a new quarterly score.

Forrester calculates supply pressure using capital spending by AI compute builders and suppliers, along with U.S. investment in information-processing equipment. Its demand measure draws on supplier revenue, business software investment and labor productivity. The supply and demand scores compare infrastructure investment with indicators of realized demand.

The financial-health measure covers audited results from public AI compute builders. OpenAI and Anthropic are outside the nine-company basket, while their activity is reflected indirectly through public builders’ filings and circular deals. Forrester says it tested the model against telecom and housing bubbles and earlier periods of healthy technology infrastructure investment.

A correction could still reshape prices and contract terms and potentially lead to startup failures, acquisitions or discontinued products. That creates a supplier-continuity risk for businesses planning around AI tools, alongside the gap between capacity spending and measured demand.

The model adds a capacity-and-financial-health measure to the debate reflected in earlier family-office concerns about AI bubble risk, which focused on investor expectations.