Fidelity: A 30% Compression in Tech Multiples Already Prices In Fears of an AI Slowdown

The tech sector has gone through an unusually large revaluation, and according to Denise Chisholm of Fidelity, that is more an argument for further gains than against them. Speaking on Bloomberg Tech with Ed Ludlow, she said multiples in tech stocks have compressed by about 30%, even though earnings expectations keep rising. Her conclusion: the market has already priced in fears that spending on AI and technology will slow, which leaves the sector room to move higher, or to "climb a wall of worry."

What a 30% multiple compression means

A multiple shows how much investors are willing to pay per unit of earnings. When it falls by a third while earnings forecasts are rising rather than falling, the market is effectively pricing in the worst case. Valuations are shrinking not because the business has deteriorated, but because uncertainty about its future has grown.

Chisholm ties that uncertainty to one specific fear: that the wave of investment in AI infrastructure, meaning accelerators, data centers and cloud capacity, is starting to run out of steam. This spending has driven revenue and expectations across the whole chain in recent years, from chipmakers to software companies. If investors worry that the flow will slow, it makes sense for them to cut valuations in advance.

Capex is slowing, but the sector doesn't necessarily shrink

Chisholm's second point is arguably the more interesting one. Historical data, she says, shows that a slowdown in capital expenditure in the tech sector does not necessarily mean the sector contracts. It is easy to confuse two things here: the growth rate and the absolute level. If companies spend 15% more on infrastructure than a year ago, compared with 40% before, that is a slowdown. But it is still growth, not a pullback.

The brief summary of the interview does not show which periods and metrics she uses, so the argument cannot be checked against the numbers. The logic is clear, though: the market often reacts to the second derivative, that is, the change in speed, while earnings depend on the level, which stays high.

Key takeaways

• Tech stock multiples have compressed by about 30%.
• Earnings expectations in the sector keep rising.
• The main source of anxiety: a possible slowdown in spending on AI and technology.
• In Chisholm's assessment, historical statistics do not link a slowdown in tech capex to a mandatory contraction of the sector.
• Conclusion: valuations already reflect part of the negativity, which leaves room for upside.

How to read this signal

This position has practical implications for investors and anyone following the AI industry. First, look past the headlines about an "AI bubble" and watch two lines together: revisions to earnings forecasts and comments from large companies about their capex plans. A divergence between them, with valuations falling while earnings rise, is exactly the case Chisholm describes.

Second, it is worth distinguishing between "slowdown" and "cut" in company reporting. "Capex will grow less than expected" and "capex will be cut" have different consequences for suppliers of chips, servers and data center power.

Third, this is the opinion of one strategist, not a guarantee. If AI spending does fall in absolute terms, the historical analogy will stop working, and the cheap multiples will turn out to be justified. That makes the upcoming quarterly reports and capex guidance the main test of her thesis.