This is Goldman Sachs Research’s second upward move in as many months. In late April, the bank had already raised its year-end forecast to 7,600 after the S&P 500 surged 13% from its March lows, its sharpest rally since April 2020, Goldman Sachs said in late April.

The May revision sharpened the same argument rather than changing it. Goldman now sees the index ending 2026 at 8,000 and says the market’s strength is still showing up in corporate profits, not in investors paying more for each dollar of earnings, Goldman Sachs said last month.

Ben Snider, Goldman’s chief US equity strategist, put it plainly: the 2026 rally has been powered by earnings growth, not multiple expansion, Goldman Sachs said last month. That matters because it leaves the bank making a fairly narrow call. If earnings wobble, the rest of the case gets thinner fast.

The change in Goldman’s forecast came down to profits, and not in a subtle way. In late April, the bank projected 12% earnings-per-share growth for 2026, Goldman Sachs said. By May 28, strategists had raised that estimate to $340 per share, which they said would mean 24% annual growth, and to $385 for 2027, or 13% growth, Goldman Sachs said last month.

Earnings season pushed them in that direction. S&P 500 profits rose 18% year over year in the first quarter, and the median company in the index was on track for its strongest quarterly growth rate in a decade outside the 2018 tax-cut surge and the post-pandemic reopening, Goldman Sachs said last month. MarketScreener, citing LSEG data, reported last week that consensus expectations for 2026 S&P 500 earnings growth had jumped from 16% in early January to almost 25% by late May.

Valuation is the other half of the story, and it is not exactly cheap. Goldman expects the S&P 500’s price-to-earnings multiple to stay roughly flat at around 21 times forward earnings through year-end, a level it says sits in the 88th percentile of the past 40 years, Goldman Sachs said last month. Modestly lower Treasury yields help at the margin, but slowing growth, geopolitical uncertainty and skepticism about AI profit durability all lean the other way.

The core of Goldman’s case is that AI spending has become large enough to matter at the index level. AI-infrastructure beneficiaries are expected to account for roughly half of S&P 500 earnings growth in 2026 and 2027, Goldman Sachs said last month, up from around 40% of this year’s earnings growth in Goldman’s April note.

The spending behind that forecast is enormous. Analyst consensus now puts hyperscale tech capital expenditures at $754 billion this year, an 83% increase from 2025, rising to $905 billion in 2027, Goldman Sachs said last month. Goldman thinks 2027 estimates may still be too low, noting that analysts have been too conservative for three straight years.

The beneficiaries are not just the biggest tech names. Goldman says semiconductor companies are the primary direct winners, with tech hardware, industrials and utilities also getting a sizable earnings lift from the AI buildout, Goldman Sachs said last month. Goldman also said in late April that a basket of AI data center construction stocks had returned nearly 60% so far this year, Goldman Sachs said.

The catch is that Goldman’s productivity assumptions are still modest. Its forecasts embed a 0.4 percentage point boost to S&P 500 EPS growth from AI-driven productivity this year and a 1.5 percentage point boost in 2027, Goldman Sachs said last month. That means the near-term case depends mostly on companies selling the picks and shovels for the AI buildout, while enterprise adoption is still early.

Goldman is bullish, but it is not pretending the path is clean. Market breadth, the share of stocks participating in the rally, has fallen to one of its narrowest readings since the dotcom era, Goldman Sachs said in late April. Narrow rallies can keep going. They also tend to leave a mess when they stop.

The macro risks are more immediate than usual. Goldman warns that a sustained disruption to flows through the Strait of Hormuz would weaken consumer spending, pressure profit margins, keep inflation elevated and leave the Federal Reserve with less room to ease than strategists had expected earlier this year, Goldman Sachs said last month. Goldman Sachs Asset Management also said this week that if energy flows resume in the next month, Brent futures could hover near $90 a barrel by year-end, GS Asset Management said on June 2.

There is a more structural risk too. Goldman says sustaining the earnings momentum in its forecast will require hyperscale companies to generate enough return on AI investment, which ultimately depends on enterprise users realizing productivity gains that justify continued spending, Goldman Sachs said last month. That is a long chain of dependencies. All the links have to hold.

Goldman’s wider macro view is a version of slow, but not broken. Its base case is 2.1% US GDP growth in 2026, with recession still not the base case because a resilient private sector keeps growth intact even as the second half slows under higher energy prices, GS Asset Management said on June 2.

Inflation should cool as well. Goldman sees core PCE falling to 2.8% by December as tariff effects fade and wage and shelter inflation continue to cool, though higher oil prices and AI-related price pressures may slow that move, GS Asset Management said on June 2.

For rates, Goldman thinks the Bank of England and the FOMC can stay on hold this summer, while the European Central Bank is expected to deliver two 25bp hikes this year in June and September before cutting back to 2% in 2027. The Bank of Japan is expected to keep hiking, GS Asset Management said on June 2.

The message for portfolios is straightforward enough. Goldman says the speed and source of rate moves matter more for equities than the absolute level, and that a jump of 40bp or more in a month would likely cause some indigestion, GS Asset Management said on June 2. Sovereign benchmarks also offer attractive income, and rates would have to rise significantly to produce negative returns over the next year, with central banks still able to cut if the macro backdrop deteriorates, GS Asset Management said on June 2.

Goldman’s 8,000 target is meaningful, but the mechanism matters more than the number. The bank is betting that earnings can keep doing the work while valuations stay rich and growth slows. That leaves little room for disappointment, especially with the S&P 500 trading at about 21 times forward earnings, a level Goldman says ranks in the 88th percentile of the past 40 years, Goldman Sachs said last month.

The wider market is still less convinced. Goldman’s forecast sits roughly 380 points above the 7,620 median from the strategist poll reported by MarketScreener, a useful measure of how much optimism the bank is carrying relative to the pack.

So the question hanging over the Goldman Sachs 2026 market outlook is not complicated, just awkward. Can AI spending keep turning into durable profit growth? Goldman thinks the answer is yes, and says the productivity side of the story should become more visible in coming years, Goldman Sachs said last month. For now, it is still a wager on a handful of companies that are selling the future and trying to get paid for it twice.