
Essentially, Goldman Sachs is telling AI stock investors to look past the flashy headline profits and focus on the bills that are coming due.
For the better part of the past few years, investors treated the AI spending boom as a clear win for Big Tech.
Nvidia’s chips continued to sell like crazy, cloud demand surged, and S&P 500 profitability climbed to record levels.
The trade felt simple: more AI spending meant greater earnings power.
To put things in context, Bianco Research recently said the S&P 500’s 7.3% gain since Feb. 27, 2026, was almost entirely driven by AI-related stocks, with the rest of the index basically flat, as cited by MarketWatch.
However, according to a Seeking Alpha report, Goldman’s message is less comfortable for AI stock investors.
Though the rally still has support from the bottom line, the cost of staying in the AI race is becoming harder to ignore.
Goldman Sachs’ message to AI stock investors is that the profit story has become a lot more complicated.
More AI:
Investors were treating the AI buildout as a simple earnings tailwind.
As I mentioned earlier, chip demand was robust, cloud sales were rising, and the S&P 500’s record profitability helped back up rich valuations.
Goldman strategist Ben Snider now feels the same spending wave could start pressuring returns at the largest tech giants.
The issue is not that AI demand has disappeared. It is that hyperscalers are spending aggressively to stay in the race.
Goldman estimates major cloud operators will spend about $770 billion on capital expenditures in 2026, roughly equal to 100% of operating cash flow.
Naturally, that higher data-center spending can drag on asset turnover, lift depreciation costs and weaken return on equity.
So the singular growth argument is also being judged on whether AI can outrun the heightened costs of building out the infrastructure.
Goldman’s warning lands hardest on the stocks that are doing the market’s heavy lifting.
According to Goldman Sachs, Nvidia, Microsoft, Alphabet, Amazon, Meta, Apple, and Broadcom collectively generate a 44% return on equity, up 9 percentage points over the past three years.
Additionally, according to MacroMicro’s April 2026 data, the Magnificent 7 had a combined market cap of $22.35 trillion, representing 34.78% of the S&P 500, meaning roughly one-third of the benchmark was linked to just seven companies.
The big risk is that AI capex switched up that math.
Nvidia and Broadcom benefit directly from infrastructure demand, while Microsoft, Alphabet, Amazon, and Meta are spending heavily to build it.
Here’s the list of AI stocks with their year-to-date gains/losses:
Goldman’s warning lands at a point when investors have fewer policy cushions than expected.
For starters, hopes of a rate cut have faded.
Reuters reports that futures still expect the Fed to hold rates at 3.50% to 3.75% in June but now price a 68.4% chance of a December hike after the May jobs surprise.
Moreover, the latest jobs report showed 172,000 jobs added in May, and unemployment held at 4.3%, keeping the Fed from rushing to help growth stocks.
Inflation is a bigger problem.
The BLS said CPI rose 0.5% in May and 4.2% over the year, with energy up 23.5%.
That keeps Treasury yields a valuation problem, even after the 10-year yield fell to 4.46%.
For AI stocks, the tension is sharper.
Nvidia’s massive $81.6 billion quarter and $91 billion guide show demand remains huge, but Goldman says hyperscaler capex could eat up operating cash flow.
The question remains whether AI spending can keep lifting earnings before it eats up returns.
Related: Michael Burry buys another beaten-down forgotten fintech stock