US Stock Selloff Spreads Beyond Tech Sector, Echoing Dot-Com Bubble Warnings

nashnova research
今天发布阅读约 9 分钟

The S&P 500 is heading for its worst week since April 2025 as AI-driven selling spills from tech into the broader market — TS Lombard warns the pattern mirrors the run-up to the 2000 dot-com crash, and today's far heavier tech weighting means the next shock wave could hit harder.

01

The 2000 playbook — where are we in the script?

TS Lombard traces the sequence: after the Nasdaq peaked on March 10, 2000, it fell more than a third in roughly five weeks, then rallied 35% as investors bought the dip and rotated into profitable big-tech names.
The real turn came six months after the peak: tightening liquidity, widening credit spreads, and earnings downgrades began dragging down companies whose fundamentals had been healthy — the Nasdaq then dropped another 40%-plus in three months.
This means → the first leg down was not the killer; the first bounce created the illusion that "the worst is over," and the second wave spread the damage across the entire market.
02

Where does the current sell-off stand?

The Nasdaq's correction began from its October all-time high, triggered by fears of excessive AI capital spending combined with the Democrats' congressional-election sweep.
The S&P 500 initially held up, but selling pressure has now visibly spread to non-tech sectors — the hallmark of the "late first stage" in 2000.
TS Lombard macro strategist Dario Perkins put it bluntly: "This bears an uncomfortable resemblance to the summer of 2000 — just one step away from the real crash."
03

Tech weighting is higher than in 2000 — what does that change?

Last month, information technology's weight in the S&P 500 topped 40%, above the 2000 peak of 35%.
Add Meta Platforms and Alphabet — classified under communication services — and the broad tech weight briefly exceeded 46%.
In plain terms = in 2000, other sectors could still cushion the index when tech collapsed; today tech accounts for nearly half the index, so the buffer barely exists — every extra point of tech decline drags the whole market harder.
04

Valuations are saner than in 2000 — but is that reassuring?

The clearest difference from the dot-com era: many companies back then had no earnings at all, and those that did traded at forward P/Es near 60×. Today's major tech firms are solidly profitable, valuation premiums are more moderate, and much of the froth appears to have been wrung out during this correction.
Yet Perkins notes that the arguments his own team makes for AI expansion are "strikingly similar" to those circulating before the 2000 bubble burst.
This reflects a deeper truth: reasonable valuations ≠ safety. Every bubble had a "this time is different" narrative before it popped. Perkins' own words: "That should at least keep us humble."
05

Order books are still full — so why worry?

AI infrastructure demand remains strong: major cloud providers have not cut capex plans, and compute-demand metrics stay elevated.
Perkins warns: "Bottlenecks are not a reason to be bullish — they may be a harbinger of massive overcapacity."
In plain terms = a full order book does not mean the turning point won't come. The lesson of 2000 is that healthy orders masked the inflection — when demand finally rolled over, the capacity built up beforehand became the weight that crushed the market.

市场有风险,内容仅供研究参考,不构成投资建议。