This article first appeared on GuruFocus.
AI’s most crowded trade is starting to feel the weight of its own history. Investors keep drawing parallels between today’s boom and the dot-com era, and the comparison has taken on new urgency after reminders of how long recovery once took. Amazon needed more than eight years to regain its 1999 peak, while Cisco only completed its post-crash climb this week. That backdrop has pressed on Nvidia (NASDAQ:NVDA), which has slipped more than 7% over the past seven days even as expectations point to another forceful quarter ahead. Analysts still see the company delivering 57% year-on-year revenue growth and possibly reaching fiscal 2026 sales of $208 billion, supported by a steady ramp-up of next-generation hardware. But with access to selling top chips into China now appearing closed off, a growth engine many investors counted on may be narrowing at the same moment the 1999 narrative is returning to the conversation.
Still, demand signals around AI infrastructure continue to show strength, even as sentiment has turned more cautious. Recent high-profile exits have sharpened the debate: SoftBank sold its entire $5.8 billion Nvidia stake to redirect capital toward OpenAI, while a fund tied to Peter Thiel disclosed that it had unloaded roughly $100 million worth of Nvidia shares after expressing doubts about AI’s ultimate payoff earlier this year. At the same time, CoreWeave’s more than 40% slide over the past 30 days has highlighted the financial strains that could complicate the broader buildout. Margins have tightened, guidance has eased, and questions around whether today’s tidal wave of spending might be bumping up against real-world constraints are getting louder. Blue Owl Capital co-founder Marc Lipshultz captured the mood, noting that when the commitments run into the trillions, the question becomes whether counting still even matters.
Even with the market’s new caution, the industry’s leaders continue to frame AI’s trajectory as a multi-year transformation that could echo the internet’s impact. Alphabet’s Sundar Pichai said in a BBC interview that the early web saw heavy excess investment but still reshaped how society operates, a pattern he expects could repeat. Nvidia’s position inside the AI supply chain remains central, although the ecosystem around it is broad and includes players that may be far more vulnerable. The latest example came Tuesday when Nvidia committed $10 billion to Anthropic, which said it would use Microsoft’s cloud platform to access Nvidia chips, underscoring how circular and interdependent these partnerships have become. That strategy may diversify Nvidia’s exposure away from a handful of hyperscalers, yet it cannot fully protect the company from a potential cooling of confidence across the sector as 2025 turns toward 2026.