AI Capex Boom Outpaces Dotcom Era, But Risks Are Manageable

The surge in capital expenditure (capex) on artificial intelligence (AI) by U.S. Hyperscalers is reaching historic proportions, raising questions about whether this boom will lead to another tech bubble similar to the dotcom mania of the late 1990s. According to analysts at Goldman Sachs and Morgan Stanley, AI-related capex is projected to hit $800 billion in 2026, with a further increase to an eye-popping $1.12 trillion by 2027. However, while the parallels are striking, there are key differences that suggest a bubble burst may be less likely this time around.

The Scale and Speed of AI Investment

The current AI capex boom is not just the largest since the dotcom era; it's the biggest in history. Analysts at Goldman Sachs and Morgan Stanley have revised their forecasts upward, with Morgan Stanley raising its 2027 outlook by 17% to $1.12 trillion. For context, hyperscalers' capex bill was only $260 billion in 2024, according to Morgan Stanley. This massive investment is rapidly depleting Big Tech's cash reserves, with PIMCO analysts estimating that capex will absorb 94% of hyperscalers' operating cash flows over the next two years, up from 40% in 2023.

The scale and speed of this investment are driven by a "we can't afford to fall behind" mentality, reminiscent of the late 1990s. However, there are crucial differences. In the dotcom era, aggressive competition led to significant overcapacity in fiber optic cables, which contributed to the bubble's burst. Today, while AI investments are similarly intense, they are more focused on developing and deploying technologies with clear commercial applications, such as data centers, power, equipment, and software.

The Bottom Line

The current AI capex boom is unprecedented, but it also comes with a different set of risks and opportunities compared to the dotcom era. While the rapid pace of investment and increasing debt levels are cause for concern, the focus on practical applications and long-term productivity gains suggests that this boom may be more sustainable. Investors should remain vigilant and monitor key metrics such as capex absorption rates and return on investment to ensure that the AI revolution delivers on its promise without leading to another tech bubble.

As James Sullivan, co-head of JPMorgan's global investor conferences, notes, "AI headlined our recent global investor conferences, the core message: experiment early, productivity gains are just getting started." This sentiment underscores the belief that while the risks are real, the potential rewards from AI investments could be transformative for both tech giants and the broader economy.