AI Spending Nears Dot-Com Era Levels as Hyperscalers Ramp Up Infrastructure Investment

TLDR:

• Goldman Sachs forecasts hyperscalers will direct about 98% of operating cash flow into CapEx by 2026.
• AI spending is approaching dot-com era levels as firms expand data centers and computing infrastructure.
• Big Tech capital expenditures could reach $920 billion by 2027, with upside estimates near $1.4 trillion.
• Growing concerns remain over whether AI spending can generate returns that justify rising investments.

AI spending is approaching levels last seen during the dot-com era, according to new projections from Goldman Sachs.

The firm expects hyperscalers to allocate nearly all operating cash flow toward capital expenditures by 2026. The forecast has renewed discussions about whether current AI spending can generate returns that match the scale of investment.

Hyperscalers Push AI Spending Toward Historic Highs

Goldman Sachs projects that hyperscalers will allocate about 98% of their cash flow from operations to capital expenditures in 2026. The estimate places current AI spending close to levels seen during the peak of the dot-com boom.

The discussion gained attention after market commentary shared by Global Markets Investor pointed to rising concerns around AI spending.

The post noted that major technology companies could direct almost all internally generated cash toward data centers, computing infrastructure, networking equipment, and AI hardware.

The chart accompanying the analysis tracks capital expenditure as a percentage of cash flow from operations. Historical data shows telecom companies exceeded 120% during the early 2000s infrastructure boom. Meanwhile, the broader technology, media, and telecom sector reached nearly 95% at its peak.

In contrast, hyperscalers maintained much lower spending levels for years. Between 2015 and 2018, the group invested roughly 30% to 40% of operating cash flow.

However, AI spending accelerated sharply after cloud demand expanded and artificial intelligence development intensified.

By 2023, the ratio climbed to around 55%. Goldman Sachs now expects it to reach approximately 68% in 2025 before moving toward 98% in 2026.

The projection suggests that nearly every dollar generated from operations could be redirected into infrastructure growth.

Questions Grow Around Returns on AI Spending

The current investment cycle is largely driven by demand for AI computing capacity. Companies continue expanding data centers while purchasing large volumes of GPUs and networking equipment to support advanced models.

Goldman Sachs expects Big Tech capital expenditures to approach $920 billion by 2027. Under its more aggressive scenario, spending could rise to as much as $1.4 trillion. That figure would represent growth of up to 89% compared with average 2026 projections.

At the same time, some businesses deploying artificial intelligence tools are evaluating whether AI spending is producing sufficient returns. Revenue growth remains a key focus as infrastructure costs continue increasing across the industry.

Competition among model providers has also intensified. Pricing pressure between major AI developers has fueled debate about long-term profitability. Lower prices may support adoption, yet they can also limit revenue growth if usage expansion fails to offset declining costs.

The chart shows hyperscalers moving well above historical averages and approaching the 100% threshold. Such levels indicate that almost all operating cash flow could be committed to future growth projects rather than shareholder distributions or other corporate activities.

Supporters of the investment cycle point to expanding cloud demand and broader AI adoption. Others remain focused on whether AI spending can produce revenue growth capable of supporting the scale of capital commitments now being planned.

With projections extending through 2026 and 2027, AI spending remains one of the most closely watched themes across global financial markets. Investors continue monitoring whether infrastructure demand and commercial adoption advance at a pace that matches rising capital allocation.

The post AI Spending Nears Dot-Com Era Levels as Hyperscalers Ramp Up Infrastructure Investment appeared first on Blockonomi.