Twenty-five years after the dot-com boom peaked, its shadow still hangs over markets. Cisco Systems (CSCO) briefly reclaimed its 2000 valuation this week, a reminder of how far investor enthusiasm can drift from business fundamentals. And as AI stocks surge, the comparisons to the late 1990s are unavoidable, Wall Street Journal columnist James Mackintosh wrote Sunday.
Valuations are stretched, again
By nearly every metric, including forward earnings, cash flow, equity risk premium and cyclically adjusted valuations, U.S. stocks are as expensive as they’ve been since the dot-com bubble. Investors are once more paying up on the assumption that a breakthrough technology will deliver years of outsize profit growth.
AI’s promise, and its uncertainty
The hope surrounding AI today mirrors the optimism surrounding the early internet. Generative AI systems can produce seemingly magical outputs, yet most are still drastically unprofitable. Unlike many dot-com startups, however, today’s AI leaders at least have revenue streams. The question is whether those revenues can scale fast enough to justify their price tags.
Infrastructure boom returns
Just as telecom companies binged on debt to lay fiber-optic networks in the late 1990s, today’s AI developers are pouring staggering sums into data centers and computing power. The spending is so large that economists say it’s propping up GDP. And just as Cisco (CSCO) benefited then, Nvidia (NVDA) and other chipmakers are the “picks and shovels” powering the current boom, posting extraordinary revenue growth that excites investors.
Narrow, single-theme market
The parallels don’t end there. In 1999, most S&P 500 (SP500) stocks lagged while anything related to the internet soared. This year, roughly a third of the index is down, but AI-adjacent names such as chips, power producers and data-center equipment are sharply higher. If a company has an AI angle, it rallies; if not, it’s ignored.
Retail traders are back in force
Individual investors are again piling into speculative small-cap names, echoing patterns seen in the dot-com mania, the SPAC frenzy and crypto’s boom. Loss-making firms have outperformed profitable small caps, a classic late-cycle sign. The rally has helped brokerage platforms such as Robinhood stage eye-popping rebounds.
But there are differences
One major distinction: today’s stock gains are modest compared with 1999. Nvidia’s (NVDA) meteoric rise doesn’t approach Qualcomm’s 2,600% explosion back then, nor the extremes reached by Cisco (CSCO), Apple (AAPL) or Intel (INTC) during the bubble’s final stretch. The speed and magnitude of the late-90s melt-up still dwarf today’s advances.
Bubble or not?
Debating whether we’re in a bubble won’t resolve much, the arguments raged just as loudly in 1999. The real test comes later. If AI fails to deliver both the promised productivity leap and strong profits for its creators, investors may be revisiting the dot-com playbook, complete with a painful aftermath. If it succeeds, today’s valuations could yet prove rational, according to the Journal.