Goldman Sachs kicked off 2026 with a confident forecast: global equities are likely to climb about 11% over the next 12 months (including dividends, in US dollars). That projection, the bank says, is driven largely by earnings growth rather than another multiple expansion — a subtle but important distinction for investors who’ve ridden a powerful rally.
A rally built on earnings — and roomy valuations
The research team, led in part by Peter Oppenheimer, frames the market as moving through the late-stage optimism phase of a prolonged cycle that began with the pandemic-related bear market in 2020. In plain terms: corporate profits are expected to do the heavy lifting in 2026. Goldman’s regional, market-cap weighted model points to roughly 9% price appreciation and an 11% total return including dividends.
That’s not a call for a repeat of 2025’s explosive run. Valuations sit at historically elevated levels across the US, Europe, Japan and emerging markets. Goldman’s analysts flag that, absent a recession, it would be unusual to see a full-blown bear market — but high starting multiples make returns more dependent on real profit growth than easy multiple gains.
Goldman also stresses diversification. In 2025, investors who spread risk across regions did unusually well; the firm expects that geographic, sector and style diversification can continue to help in 2026 as growth-adjusted valuation gaps narrow.
The AI capex wrinkle
Here’s the fly in the ointment. Massive capital spending on AI infrastructure — data centers, chips, networking — has been a major engine for tech earnings and stock returns. But Goldman analysts including Ben Snider warn that capex growth may slow, and that the profits needed to justify the scale of investment may not materialize for all firms.
Goldman estimates hyperscaler capex around $400 billion in 2025, up sharply year over year. Maintaining investors’ accustomed returns on that spending would require an eventual annual profit run-rate north of $1 trillion — more than double consensus estimates for 2026. The implication is blunt: some of today’s market leaders may struggle to generate sufficient long-term profits to match the valuation baked into their stock prices.
That dynamic can spark rotation. If AI-related returns on capital decelerate, traders and portfolio managers will get choosier — rewarding companies that convert capex into sustained profits and punishing those that don’t. The result could be two-way risk for the aggregate index, even if the broader market keeps rising.
When we talk about capex for AI, we’re not just talking about racks and servers. Big projects and experiments — from hyperscaler campus builds to more novel ideas — are reshaping where and how compute gets deployed. Some of that ambition even spills into out-there concepts like orbital data centers, underscoring how infrastructure strategy has become part of the investment story (see how firms are exploring new forms of AI infrastructure in Google’s Project Suncatcher).
Meanwhile, corporate adoption of AI services and capabilities is accelerating on the software side, changing revenue mix and productivity dynamics — developments that show up in research tools and product rollouts such as Gemini’s deeper integration into productivity workflows.
What investors might do (without sounding like a checklist)
Goldman’s view nudges toward a few practical moves: broaden geographic exposure, balance growth and value factors, and consider sector themes beyond headline tech names. With correlations falling and alpha opportunities rising, stock-picking matters again — especially as markets may reward companies that translate AI spending into real, repeatable profit.
Non-tech sectors could benefit from spillovers of AI capex (supply-chain winners, industrial automation, enterprise software), so the opportunity set is wider than just the megacap AI leaders. Commodities may also see selective gains — precious metals offsetting energy weakness in Goldman’s commodity outlook.
Finally, remember that the market’s present optimism can feel fragile. Elevated valuations mean the path to that 11% return is narrower: outperforming will likely require earnings beats and steady macro conditions (Goldman expects modest Fed easing), rather than another surge in valuation multiples.
Markets are rarely as tidy as forecasts. Goldman’s 11% base-case is plausible — but it arrives with a clear rider about AI spending and the uneven probability that today’s biggest spenders will deliver the profits investors hope for. That makes 2026 less a simple continuation of the last year’s party and more a test of which companies can turn big bets on AI into lasting returns.