Goldman Sachs pegs global AI investment at $1 trillion in 2026, sees ‘robust near-term growth outlook’

The $794 billion figure that is very often cited for artificial intelligence (AI) spending this year may be missing a much bigger part of the investment story.

Goldman Sachs Research estimates that companies will invest about $1.019 trillion globally in AI-related capital expenditure (capex) in 2026, once spending beyond US hyperscalers is taken into account.

The estimate puts US AI investment at about $581 billion, however investment outside the US accounts for roughly $438 billion. The US figure includes about $529 billion from public hyperscalers, around $35 billion from private US companies and other AI-exposed businesses, and a smaller contribution from non-US public companies tied to investment in the US. Outside the US, the estimate includes about $233 billion from US-linked public hyperscaler investment, $68 billion from private companies and $138 billion from non-US public companies.

The $1 trillion AI capex picture

The commonly used measure focuses largely on the capital spending plans of US hyperscalers. Analyst consensus puts their 2026 spending at about $800 billion, or $794 billion in the specific estimate examined by Goldman Sachs Research.

However that number leaves out spending by private companies and businesses outside the US, including major Asian companies. It also assumes that all hyperscaler capex is connected to AI, even though that is not necessarily the case. Because US technology companies operate globally, some of their investment is made outside the US.

What goes into the estimate

Goldman Sachs Research made several adjustments to the traditional hyperscaler-capex measure. It included capex projections for US public companies outside the hyperscaler group that feature in Goldman Sachs’ baskets of AI-related stocks. The economists also gathered media reports on capital spending by key private companies considered pivotal to the AI ecosystem.

The calculation also adds capex projections for AI-exposed companies outside the US. There is another important adjustment. Goldman Sachs Research subtracts 2022 capex levels from US hyperscaler spending. The reasoning is based on the view of its equity analysts that nearly all incremental capex since 2022 has been directed towards AI projects. For the private companies included in the estimate, all capex is similarly attributed to AI. For less directly exposed and non-US companies, the economists assume that capex above 2022 levels is AI-related.

Financial leases are excluded wherever possible to reduce the risk of counting the same hardware investment twice across companies.

The researchers also looked at where the money is actually being invested. They used data on the reported locations of announced investment projects, particularly those from US hyperscalers. For most other non-hyperscalers, including private and global companies, they assume that the geographical pattern of AI investment is broadly similar. Investment by major Chinese and Korean AI-exposed companies, however, is assumed to occur outside the US.

The numbers come with some uncertainty

Goldman Sachs Research acknowledges that parts of the calculation depend on assumptions that are difficult to verify. There is also a risk of double-counting in cases where companies do not separately report property, plant and equipment investment and financial leases in their capex figures.

To test whether the preferred estimate is reasonable, Goldman Sachs Research used two additional approaches. The first looked at gross-profit realisations and revisions to forecasts relative to projections made in 2022 for public companies exposed to the AI buildout. The second used official government data to track the rise in nominal AI investment, alongside global trade data and the observed relationship between US imports and total AI investment to estimate investment in other economies.

Both checks produced a result remarkably close to the preferred measure, around $1 trillion of AI investment globally and just under $600 billion in the US in 2026.

Looking further back, the methodologies differ somewhat in their estimates of cumulative investment since 2022. But taken together, they point to about $1.8 trillion in cumulative AI investment by the end of 2026, Briggs writes.

AI spending could become a bigger part of GDP

The scale of the investment is important not just for technology companies but also for the broader economy and financial markets. “The AI capex growth outlook, including how high AI investment ultimately rises as a share of GDP and when capex growth slows, is a key source of uncertainty for macro markets right now,” Briggs writes in the note.

Goldman Sachs Research expects AI investment to keep increasing as a share of economic output over the next two years. In the US, AI capex is estimated at 1.8% of GDP in 2026, rising to 2.5% in 2027 and 2.8% in 2028. Globally, AI investment is estimated at 0.9% of GDP this year, increasing to 1.3% in 2027 and 1.4% in 2028.

Those levels are not unprecedented for a major technological shift. “These levels are consistent with the 2%-5% of GDP peak investment impulses observed in prior general-purpose technology buildouts,” Briggs writes. “And while our US portfolio strategy team has flagged that consensus capex projections for 2027 are likely too conservative, even significant upward revisions would leave the level of AI investment as a share of GDP comfortably within the historical range observed in prior technology cycles.”

Is the AI spending boom close to slowing?

The next question is not simply how much companies are spending, but how long the surge can continue. Goldman Sachs Research says there is no single indicator that can reliably signal when AI capex growth will slow. Instead, its economists favour a “dashboard approach”, combining several leading indicators.

The dashboard includes semiconductor manufacturing equipment imports in Taiwan and South Korea, relevant Purchasing Managers’ Index (PMI) indicators and components, import prices, and prices for memory purchases and Graphics Processing Unit (GPU) rentals.

These indicators have historically provided advance signals on the pace of US AI capex growth. So far, they are not flashing a slowdown. “The good news for the capex outlook is that all leading indicators rank near the top end of their range since 2022,” Briggs writes. “This pattern suggests a robust near-term growth outlook.”