Forrester says the AI bubble won’t burst soon, but firms should prepare anyway

Forrester Research has looked at the AI boom and reached a conclusion that will satisfy almost nobody. The bubble is real enough to worry about, but it is not about to pop.

The analyst firm’s message to businesses is a bit like a weather forecast that says no storm today, but you should still know where the umbrella is. In its view, a correction could still shake up AI suppliers, contracts, and pricing, even if a full-blown crash is not on the near-term calendar.

Inside Forrester’s AI Bubble Model

On November 5, 2026, Forrester published a report titled “An AI Bubble Burst Is Not Imminent.” It introduces a quantitative tool the firm calls its AI Bubble Model.

The model works like a scale with two pans. One side measures supply pressure, meaning how much AI infrastructure is being built. The other measures demand absorption, meaning how much real appetite exists to actually use those capabilities.

The two sides are not balanced. As of Q2 2026, supply pressure scored 4.95, close to the top of the scale, while demand absorption came in lower at 3.44.

So why isn’t the firm calling a crash? The answer sits in a third measurement: the financial health of the companies paying for the build-out.

Forrester analyzed nine firms and gave them a combined financial health score of 82 out of 100, which it classifies as robust. A more recent nowcast was slightly less cheerful, showing the score at 79.1 as of November 2, 2026, down from 82.1 on October 29.

Why cash flow is the firewall

The core of Forrester’s argument is about who is writing the checks. According to the report, approximately 90% of AI computing infrastructure funding comes from large tech companies using their own cash flow.

“Fears that an AI bubble is about to burst are premature,” said Brian Hopkins, VP of Forrester’s Emerging Tech Portfolio.

Hopkins is not saying everything is fine. Forrester flags that the real funding risk is concentrated in a smaller group of builders under genuine financial strain.

Forrester says it tested the model against earlier episodes before applying it to AI. According to the firm, it back-tested successfully against the telecom and housing bubbles, as well as against what it describes as healthy build-outs by big tech between 2012 and 2017.

A shift from last year’s warnings

The new report refines Forrester’s own earlier thinking. As late as October 2025, the firm warned that uncertainty over return on investment could push a $50 billion backlog of planned AI spending into 2027.

The updated view does not claim that worry was wrong. It argues that oversupply on its own does not guarantee a correction.

The missing variable, according to Forrester, is demand realization. In particular, the firm points to demand that goes beyond simple productivity gains as the big unknown.

The macro math illustrates the gap. Forrester’s report indicates that for AI infrastructure investment to lift US GDP by 5% by 2035, a dramatic acceleration would be needed beyond current productivity growth estimates of around 2 percentage points.

What this means for businesses and the market

For enterprises buying AI services, Forrester’s advice is practical rather than panicked. The firm urges companies to prepare for possible supply chain problems and pricing swings, and to build contingency plans and exit strategies.

The nine-firm financial health score is the number to watch. A slide from 82.1 to 79.1 in a matter of days does not signal crisis, but a sustained decline would chip away at the main reason Forrester gives for its calm outlook.

The other number to watch is the gap between supply pressure at 4.95 and demand absorption at 3.44. If demand catches up, the build-out starts to look like foresight. If it stalls while supply keeps climbing, the telecom comparison gets uncomfortably close.