Goldman Sachs projects $1.2 trillion Big Tech AI spend by 2027

By Splitfeed EditorialLast updated Sep 27, 2026, 08:20 UTC

Goldman Sachs published a projection in late September 2026 estimating that five major technology companies, Amazon, Alphabet, Microsoft, Oracle, and Meta, will collectively spend $1.2 trillion on AI infrastructure in 2027. That figure represents an increase of more than 50 percent above 2025 spending levels and sits well above what Wall Street had previously estimated. The bank framed the scale of the investment by comparing it, in GDP terms, to the railroad construction boom of the 19th century, one of the largest capital mobilizations in American economic history. Goldman also identified three structural bottlenecks that could constrain or delay the buildout: power supply, availability of skilled labor, and capacity in the memory chip market.

The projection matters for several reasons beyond the headline number. It covers only five companies, meaning the broader industry figure would be larger still. The comparison to railroad construction carries historical weight because that era produced both transformative infrastructure and a series of financial crashes driven by overbuilding and speculation. Whether AI infrastructure follows a similar arc is a genuinely open question, and the bottleneck warnings Goldman included suggest even the bank sees the path as uncertain. The gap between this projection and prior Wall Street estimates also raises questions about how quickly analyst consensus has shifted and whether the new figures are pricing in realistic returns.

Because no specific reactions from the Pro-AI, Anti-AI, or Middle Ground camps have been published in response to this story yet, their likely positions can only be sketched in general terms. The Pro-AI camp would typically treat a projection of this scale as validation that AI is entering a phase of industrial buildout comparable to past transformative technologies, pointing to the investment as evidence of genuine commercial confidence. The Anti-AI camp would likely focus on the bottlenecks Goldman named, arguing that constraints in power, labor, and chips reveal how fragile the expansion is, and might draw on the railroad analogy to warn of a speculative bubble. The Middle Ground camp would probably accept the investment as real while pressing on whether the returns justify the outlay and whether the bottlenecks will compress timelines or trigger a reallocation of capital.

The arguments on all sides will sharpen as the named companies release their capital expenditure disclosures through 2026 and into 2027. Quarterly earnings calls from Amazon, Alphabet, Microsoft, Oracle, and Meta will show whether actual spending tracks Goldman's projection, and any sign of slowdown tied to power constraints or chip shortages would become a concrete test of the bottleneck warnings the bank included in its analysis.