AI circular investments raise dot-com bubble concerns as market peaks

Growing concerns about circular investments in AI deals are sparking debate over whether the artificial intelligence boom resembles the dot-com bubble or represents sustainable growth. OpenAI’s recent financing arrangements with major tech companies mirror the investment patterns that contributed to the early 2000s crash, raising red flags among seasoned investors about potential market overheating.

The big picture: Circular investing—where companies invest in each other in exchange for products rather than cash—is becoming prevalent in AI infrastructure buildouts, echoing dangerous patterns from the dot-com era.

• OpenAI, the maker of ChatGPT, has announced multiple deals with Nvidia, AMD, Broadcom, and Oracle that are financed without the company paying cash or taking debt backed by proven consumer demand.
• During the dot-com bubble, similar circular investments worked only as long as valuations remained elevated, but collapsed when the market corrected.

Why some companies would survive a crash: Unlike pure-play AI companies, major tech hyperscalers have diversified revenue streams that would cushion them from an AI market correction.

• Alphabet generates most revenue through Google Search, Amazon through e-commerce, and Meta through social media—core businesses that would remain intact even if AI spending halted.
• Microsoft’s enterprise software and cloud services would continue generating cash flows independent of AI infrastructure investments.
• However, companies like Nvidia would face significant impact since demand for their graphics processing units (the specialized chips that power AI systems) depends heavily on continued AI buildout.

Key differences from the dot-com era: Current AI adoption shows more immediate practical value than early internet investments demonstrated.

• AI hyperscalers are already growing revenue in the mid-double digits without fully realizing AI benefits.
• Business and consumer adoption of generative AI is occurring at a faster pace than internet adoption in the early 2000s.
• The AI data center buildout trend is projected to reach multitrillion-dollar levels by 2030, suggesting substantial long-term economic value.

What investors should know: Market valuations are reaching frothy levels, particularly in the private sector, but the fundamentals may support continued growth.

• The market is trading at all-time highs, which could limit return potential in the near term.
• Investors should maintain balanced cash positions and carefully manage exposure to AI stocks rather than making dramatic portfolio changes.
• The internet buildout eventually delivered massive economic value in the decade following the dot-com crash, though not at the pace initially envisioned.