Goldman Sachs has delivered one of its most aggressive AI outlooks yet, signaling that the artificial intelligence boom is not slowing down—it is accelerating into a multi-trillion-dollar infrastructure cycle that could reshape global equity markets.

The bank’s latest analysis suggests that AI-related capital expenditures by hyperscalers could reach $1.1 trillion by 2027, with a bullish scenario pushing that figure to $1.4 trillion.

This projection is not just a headline—it is a structural thesis about where the global economy is heading.

While early AI hype centered around software models and chatbots, Goldman Sachs argues the next phase is fundamentally physical.

Data centers, chips, power infrastructure, cooling systems, and network expansion are becoming the bottleneck of growth—not algorithms.

Cloud giants like Amazon Web Services and Google Cloud are already reporting surging demand backlogs, signaling that AI adoption is now constrained by hardware availability rather than customer interest.

In earlier phases of the AI rally, semiconductor companies dominated investor attention. Chipmakers such as Nvidia and AMD became the face of the AI trade.

But Goldman Sachs now warns that this phase is maturing.

Instead, the next winners will likely be hyperscalers and deployment platforms—companies that actually integrate AI into enterprise systems rather than simply selling hardware.

This marks a critical rotation in the AI investment cycle.

Despite massive investment flows, Goldman notes a key contradiction: productivity gains from AI remain uneven and difficult to measure at scale.

Enterprise adoption is rising, but monetization lags behind capital deployment.

Some firms report efficiency improvements, while others see minimal financial impact.

This imbalance creates a tension between valuation optimism and real-world economic output.

The most debated question in markets today is whether AI represents a sustainable technological supercycle or a speculative bubble.

Goldman Sachs’ position is nuanced.

The bank does not classify AI as a bubble, but it does acknowledge that valuations are running ahead of realized economic benefits in certain segments.

In other words: fundamentals exist, but expectations may be stretched.

One of the most important insights from Goldman’s analysis is that the limiting factor for AI expansion is no longer capital—it is capacity.

Power grids, data center construction timelines, and semiconductor supply chains are emerging as structural constraints.

This means AI growth will likely be uneven, cyclical, and regionally concentrated.

Goldman Sachs is effectively reframing AI as a “nation-scale infrastructure cycle,” comparable to railroads, electrification, or the internet.

If the forecast is correct, the next five years could see unprecedented capital flows into technology infrastructure—reshaping not just tech stocks, but the entire global equity index composition.