Nvidia is not Cisco, but the semiconductor trade still needs a reset
The easy bubble analogy misses the real point: **NVDA** is operating from a far stronger position than **CSCO** ever did at the end of the dot-com cycle. But this selloff still matters, because the semiconductor trade had become too indiscriminate and now looks set for a selective de-rating across the stack rather than an outright AI bust.
The market keeps asking whether NVDA is the next CSCO
, but that comparison is too neat to be useful. Nvidia is still putting up growth and profitability numbers that Cisco never had when the dot-com cycle rolled over, which is why the straight-line bubble call feels lazy. The better read is that semis had become a one-way AI trade, and now the market is finally separating the genuine platform winners from the rest of the stack. That is a reset, not a collapse — and investors should treat it that way.
Start with the basic operating reality. Nvidia is not a story stock trading on distant promises; it is a cash machine scaling at absurd size. The company just posted Q1 FY2027 revenue of $81.6 billion, up 85% year over year, with gross margin of 74.9%. Cisco’s late-cycle profile was nowhere close: by fiscal 2001, net product revenue growth had slowed to 15% from fiscal 2000. That does not mean Nvidia is immune to multiple compression, but it does mean the popular analogy confuses valuation psychology with business quality.
The valuation picture also argues for nuance, not panic. On the supplied market data, NVDA trades at 35.43x earnings with 65.5% revenue growth and a 63.0% net margin. CSCO, by contrast, sits at 27.65x earnings with just 5.3% revenue growth and 19.7% net margin. If the bear case is that Nvidia is simply another Cisco, the numbers do not back it up. A premium multiple attached to elite growth and elite margins is not the same thing as a mature infrastructure name getting bid up into a mania.
What the selloff is really exposing is that investors had started treating the entire semiconductor chain as one trade. That was always too simplistic. TSM still reported Q2 2026 revenue of $40.2 billion and continues to describe AI demand as strong and multi-year, yet the stock still got hit because the market is now more focused on capex intensity and margin risk than on whether orders exist at all. That is the key distinction. The market is not saying AI demand disappears tomorrow; it is saying not every company tied to AI deserves to hold peak multiples while spending ramps and expectations get stretched.
The spread inside the group makes that point clearly:
NVDA: 35.43x P/E, 65.5% revenue growth, 63.0% net margin
TSM: 35.06x P/E, 33.0% revenue growth, 50.4% net margin
ASML: 56.17x P/E, 15.6% revenue growth, 29.7% net margin
AVGO: 47.45x P/E, 23.9% revenue growth, 38.8% net margin
MU: 21.58x P/E, 48.9% revenue growth, 55.9% net margin
That is not one valuation regime. It is a stack with very different growth durability, margin structures, and capex exposure, yet for months it traded as if AI would lift every layer equally. A reset was overdue.
Yes, the bulls can fairly argue that hyperscaler demand remains real and that foundry and equipment spending still reflects a structural compute buildout, not a fad. They are not wrong. But that argument skips over the market’s actual problem, which is not whether AI exists, but how much future perfection was already priced into every adjacent name. When even strong updates from ASML and TSM are met with selling, the message is not that demand has vanished; it is that investors are repricing duration, capex burden, and the odds of over-earning at the cycle peak.
That is why we think the current pullback looks more like a selective de-rating than an AI bust. The Philadelphia Semiconductor Index has already taken a sharp hit from its highs, including an 8.5% weekly drop cited in recent market coverage, and the weakness has been broad enough to catch both obvious winners and second-order beneficiaries. But broad price damage does not automatically mean broad fundamental damage. In fact, the supplied TickerSpark Score framework would likely force the same conclusion investors are reaching the hard way: the group is no longer a momentum bundle. It is a sorting exercise.
The names most exposed to that sorting are the ones where valuation still asks investors to underwrite a long, clean capex cycle without much room for execution slippage. ASML at 56.17x earnings with 15.6% revenue growth is a very different proposition from NVDA at 35.43x with 65.5% growth. AVGO has stronger fundamentals than many peers, but at 47.45x earnings it also carries less room for disappointment than the market was pretending a month ago. Even MU, despite explosive recovery metrics, is still a memory name in a business that history says can turn quickly. The point is not that these are bad companies. The point is that the market had stopped charging different prices for different risks.
So no, Nvidia is not Cisco in the way the internet likes to frame it. Cisco became the symbol of a busted era because the business slowed into a market that had priced endless expansion; Nvidia is still delivering hypergrowth, extraordinary margins, and unusually strong demand visibility. That matters, and it is why calling this an AI bust is too blunt.
But the second half of the argument matters just as much: semis still needed a reset. What we would watch from here is not another round of bubble metaphors, but whether capex-heavy names can defend returns as spending rises and whether the market keeps rewarding actual earnings power over AI adjacency. If demand starts cracking at NVDA or TSM, the bear case gets much stronger. Until then, this looks less like 2000 all over again and more like a healthier, overdue repricing inside a trade that had become far too crowded.
Our take, not advice. This is opinion commentary — informational only, not personalized investment recommendations. Markets carry risk. Do your own research and consider your own situation before any trade.
▌The Daily Briefing · Free
A new stock idea, every evening.
One stock worth watching each weekday, plus the analysis behind it. Free, in your inbox.
