The S&P 500 could theoretically reach 10,000 by 2027, but the arithmetic makes that a demanding bull case rather than a sensible base forecast. From 7,605.87 at 11:24 a.m. ET Thursday, the index would need to gain 31.5%. That could happen through earnings growth alone, or through a combination of profit gains and higher valuations. But with bond yields near 5%, the latter is getting harder.
The round-number target resurfaced after James Thorne, chief market strategist at Wellington-Altus, described a path to 10,000—not a prediction. His published scenario requires S&P 500 earnings approaching $440 and a price-to-earnings multiple near 23. Those assumptions produce about 10,120, so the framework is internally consistent. But both assumptions need to work at once.
The 0 earnings hurdle
The earnings leg has momentum. FactSet said on September 4 that analysts lifted the third-quarter bottom-up estimate by 1.2% during July and August, to $89.69. More unusually, its calendar-2026 estimate rose 6.1% over those two months to $361.38, when estimates normally fall at this stage of a quarter.
But $440 is still 21.8% above that 2026 estimate. Other bullish institutional forecasts stop short of it: Goldman Sachs projects $385 for 2027, while Citigroup introduced a preliminary $400 estimate. At 23 times earnings, those figures imply roughly 8,855 and 9,200, respectively—not 10,000.
The gap is the useful part of the exercise. If earnings reach $440, 10,000 needs a multiple of 22.7. If the market instead stays near the roughly 21 times forward earnings cited by Goldman, profits would need to approach $476. In other words, investors need either a material earnings beat, a valuation expansion, or some combination of the two.
AI has to become profit beyond the chipmakers
The bull case is not simply that companies will spend more on artificial intelligence. Goldman estimates the largest hyperscalers will invest $754 billion this year, up 83% from 2025, and $905 billion in 2027. It expects AI-infrastructure beneficiaries to generate roughly half of S&P 500 earnings growth in both years.
That spending supports semiconductor, hardware, power and construction profits now. The harder second step is whether customers earn enough from AI applications to keep buying capacity. If productivity gains spread across the other 493 companies, $440 becomes less implausible. If the returns remain concentrated among a handful of platform and chip companies, the index will be more exposed to any capex pause or margin disappointment.
FactSet’s latest revisions offer a modest breadth check. Seven of 11 sectors saw their 2026 estimates increase during July and August, although Energy—not Technology—led the third-quarter revision after oil surged. That is encouraging breadth, but it does not yet prove the broad productivity acceleration embedded in the 10,000 scenario.
The rate market is the immediate obstacle
The valuation half of the case looks harder on Thursday’s tape. The U.S. 10-year Treasury yield was 4.922% late Thursday morning, up from 4.837% at Wednesday’s close, while Brent crude futures traded at $105.54, about 4.3% above the prior close. The S&P 500 was down 0.40% at the same point.
A near-5% risk-free yield competes directly with a roughly 4.4% earnings yield at 23 times profits. That does not make the multiple impossible, but it leaves little valuation cushion. Persistently expensive energy can also raise inflation, squeeze non-energy margins and keep the Federal Reserve tighter—three ways to damage both legs of the 10,000 calculation.
What index investors should watch
- Earnings: estimates must move beyond the high-$300s and toward $440 without relying on one-off gains.
- Breadth: profit revisions need to spread from AI infrastructure and Energy into industries that are supposed to benefit from productivity.
- Rates: a sustained decline in the 10-year yield would make a 23-times multiple easier to defend; another move higher would raise the earnings bar.
The clean conclusion is that 10,000 is arithmetically possible but not yet supported by mainstream earnings forecasts. The next 2,394 index points depend less on enthusiasm for AI than on proof that the investment boom can produce roughly $440 of durable earnings while the bond market permits investors to value those profits at almost 23 times.
