Global artificial intelligence capital expenditure is projected to reach approximately $1 trillion in 2026, with Goldman Sachs forecasting sustained growth over the next two years. According to Goldman Sachs Research, the investment will climb from 0.9% of global GDP in 2026 to 1.3% in 2027 and 1.4% by 2028. The projection highlights the ongoing expansion of AI technology and infrastructure across global markets, with the firm noting that AI's economic impact is only getting bigger. Goldman Sachs points to clear growth signals including semiconductor manufacturing equipment imports in Taiwan and South Korea, and higher GPU rental costs as evidence of sustained momentum. The bank's projections indicate that AI investment in the US could increase from 1.8% of GDP in 2026 to 2.5% in 2027 and 2.8% in 2028, while global AI investment would rise to 1.4% of GDP by 2028. Since 2022, cumulative AI investment could reach about $1.8 trillion by the end of 2026, according to the latest Goldman Sachs data. The firm noted that these spending levels align with historical general-purpose technology booms, which typically peaked between 2% and 5% of GDP. Goldman Sachs emphasized that the pace and eventual scale of AI investment remain key uncertainties for macroeconomic and financial markets, particularly as investors assess how high AI capital expenditure could rise as a share of GDP and when its growth could begin to slow.
The United States will drive approximately $581 billion in AI investment this year, as reported by ANI. This total offers a wider picture than the standard $794 billion US hyperscaler capex projection, as it incorporates spending from non-hyperscalers, private companies and international investments. US AI investment is expected to expand from 1.8% of GDP in 2026 to 2.5% in 2027 and 2.8% in 2028. Goldman Sachs emphasizes that this estimate covers not just tech giants but also private firms and international players, making the US the top spender globally. The research estimates that the estimate is broader than the commonly cited projection of around $794 billion in capital expenditure by US hyperscalers, as it also incorporates AI investment by private companies, non-hyperscaler firms and companies outside the US. Goldman Sachs verified this momentum by analyzing key indicators, including manufacturing equipment imports in South Korea and Taiwan, purchasing managers' indices, import costs, memory prices and GPU rental rates. The bank noted that near-term indicators point to continued strength in AI-related spending, with these indicators currently near the upper end of their ranges since 2022, providing a positive signal for AI capital expenditure growth.
The Federal Reserve has examined AI-related investment and concluded that it made a meaningful contribution to U.S. economic growth through the first quarter of 2026. According to the Wall Street Journal, software, computer equipment and related investment have become significant enough that the AI buildout is showing up in actual GDP numbers. This means the boom is now feeding construction jobs, electricians, cooling-equipment manufacturers, power companies, utilities, semiconductor manufacturers, real estate developers, engineering firms, cloud companies and financial institutions. The spending has become so concentrated that just four major hyperscalers represented roughly 17% of the S&P 500 while preparing to spend more than $600 billion on AI-related infrastructure during 2026. However, AI critic Ed Zitron warns that OpenAI and Anthropic may account for 70% or more of AI-related revenue at some of the largest cloud providers, creating a heavily concentrated spending cycle. Goldman Sachs noted that consensus expectations for 2027 capital expenditure could prove conservative, leaving scope for upward revisions, as the AI buildout continues to support capital spending across multiple sectors.
The AI boom is being propped up by massive spending, debt and Wall Street financing, creating significant financial risks. The Wall Street Journal reports that Nvidia has struck agreements with Apollo, Blackstone, BlackRock, Brookfield, Goldman Sachs and KKR to create compute financing platforms targeting more than $500 billion of outside capital to help companies finance AI computing infrastructure. This financing system addresses the problem that many Nvidia customers cannot afford the massive quantity of chips and computing infrastructure required to maintain current growth rates. However, this setup creates vulnerability, as Zitron warns that "the moment they stop spending, they crash the market." The concentration risk becomes evident when Nvidia's proposed guarantee for a gigantic data-center campus in Ohio was reduced from around $250 billion to less than $120 billion due to investor concerns over the company taking too much financial risk while simultaneously supporting demand for its own chips. Goldman Sachs cautioned that its estimates involve assumptions that are difficult to verify and carry some risk of double-counting, particularly where companies do not separately report financial leases and hardware capital expenditure. The bank therefore cross-checked its estimates using corporate earnings revisions, government investment data and global trade flows, with these approaches producing broadly similar estimates of around $1 trillion in global AI investment for 2026.
