The most important number in today's AI selloff is not an earnings estimate. It is the price-to-earnings multiple investors are willing to pay for that estimate. Bloomberg Intelligence now sees S&P 500 earnings growth of about 20.1% in 2027, trimmed from 21.2% a few weeks ago and down from the roughly 30% pace expected this year. A one-point revision is not what is moving prices. A change in what investors will pay for each dollar of profit is.
The arithmetic is unforgiving. Take a company earning $5 per share and valued at 30 times earnings, or $150. If profits grow 20% to $6 but the market decides the business deserves 25 times, the share price lands at $150 again. A full year of excellent execution produces a flat stock. Several large technology names have recently reported strong quarters and still traded lower, which is what this kind of multiple reset looks like in practice. The business did not disappoint; the starting valuation simply left no margin for error.
There is an encouraging side to that math for patient owners. Multiple compression is a one-time repricing, not a permanent drag. Once the market stops paying 30 times and settles nearer 20, returns reconnect with the underlying growth in earnings and cash flow. The technology sector of the S&P 500 now trades near 20 times forward earnings, close to its cheapest level since ChatGPT launched. That does not make it a bargain by itself, but it is a very different starting point than the one buyers faced a year ago.
The detail beneath the headline matters too. While the broad 2027 forecast softened, analyst estimates for information technology actually improved, and the largest platform companies are expected to regain earnings leadership as growth slows elsewhere. When growth becomes scarce, dependable growth tends to command a premium, and companies with genuine AI revenue should pull further away from those that were simply carried along by enthusiasm.
Spending forecasts deserve the same discipline. Some analysts have trimmed their longer-term AI capital expenditure outlooks, and the reason matters more than the cut. Delays that push capacity from 2028 into 2029 mean slower, not smaller. The businesses most at risk are those that need cheap debt and years of heavy investment before customers start paying, because financing conditions have tightened for exactly that model.
So what for long-term investors? Separate the two questions the market is currently blending: is the business still compounding, and is the price still reasonable? Favor companies whose balance sheets can fund growth from internal cash flow, whose earnings do not depend on every optimistic assumption landing, and whose valuation leaves room for the multiple to drift lower without erasing years of returns. Over a decade, owning durable earnings at a sensible price matters far more than guessing the next quarter's sentiment.