Nvidia's AI Slowdown Debate Is Asking the Wrong Question

Nvidia shares fall 5.6% over seven days as AI chiefs call for restraint, yet Q2 fiscal 2027 revenue rose over 105% to $96.22 billion.

Something unusual is happening at the top of the artificial intelligence industry: its loudest voices are suddenly talking about restraint. Over the weekend, Anthropic chief Dario Amodei published an essay titled "We Must Pace the Frontier," calling for a slowdown in model development — and within days, Sam Altman, Elon Musk and Demis Hassabis were all signaling agreement. For a sector that has known only one direction for years — faster, bigger, more compute — it marks a striking reversal.

Nvidia, sitting at the center of that buildout, is absorbing the whiplash. In Frankfurt the stock trades at 183.40 euros, up 0.4% on the day, a calm surface over a turbulent stretch. Over seven days the shares are down 5.6%; over 30 days, 5.7%. The 52-week high of 202.50 euros, set only in May, now sits 9.4% out of reach. An RSI of 44.4 describes a stock that has cooled off without tipping into oversold territory.

A Selloff With Many Fathers

No single event drove this week's pressure. On Monday the Philadelphia Semiconductor Index dropped 5.9%, with Micron and Intel each losing more than 5%. At the same time, the yield on ten-year US Treasuries climbed past 5% — a threshold Barclays regards as a historical turning point at which bond yields become a genuine drag on equities.

Tuesday brought the broad market down with it: the Dow fell 0.89%, the S&P 500 0.49% and the Nasdaq 0.75%. Rising oil prices, an approaching Fed rate decision on Wednesday under Chair Warsh, and now the AI industry's own second thoughts — a rare blend of macro jitters and sector introspection.

The more interesting question is whether the slowdown talk actually means what the market seems to fear. As one MarketWatch commentary put it, a slower pace of model development does not automatically translate into less spending on chips and data centers. That is the twist in this story: Nvidia doesn't earn from how quickly models get smarter, but from how much computing power that process consumes.

The Numbers Refuse to Cooperate With the Panic

The company's operating results tell no story of deceleration. In the second quarter of fiscal 2027, revenue reached $96.22 billion, up more than 105% year over year. The data center business alone grew 117% to $89.02 billion, with gross margin at 75%. For the current quarter, management has guided to $108 billion — and that figure explicitly assumes no data center revenue from China.

Those growth rates put the recent price swings in perspective, even if they don't erase them. What's equally notable is how deeply Nvidia is now woven into the broader AI ecosystem. Reuters reported over the weekend that the company is in talks to anchor Anthropic's planned IPO with as much as $10 billion, in a listing that could value the company at roughly $2 trillion — the same Anthropic whose chief executive is now publicly urging caution.

That headline deserves context. Nvidia had already signaled an investment commitment of up to $10 billion to Anthropic back in November 2025, which makes the reported anchor role more a continuation than a doubling of its exposure. Neither side has confirmed the talks. The same caution applies to the revenue figure circulating for Anthropic: a reported annualized run rate of about $65 billion by the end of July 2026, up from roughly $9 billion at the end of 2025. That number comes from investor documents and media reporting, not from Anthropic itself — treating it as a confirmed growth miracle overreads an unverified metric.

Meanwhile, the technological footprint keeps widening. At the Dreamforce conference, Nvidia and Salesforce unveiled the reasoning model Koa, built on Nvidia's open Nemotron model — evidence of how far the company's technology has spread into other vendors' software, well beyond its own chips. Nvidia also extended its CUDA-Q platform with an orchestration layer for fault-tolerant quantum computing, underscoring that the growth story rests on a broad infrastructure bet rather than graphics chips for language models alone.

The Real Bet

What lingers is the worry about a crowded trade. One fund manager recently cautioned that Nvidia isn't expensive on a one-to-three-year view, yet the concentration in the AI trade carries the risk of a rapid unwind should demand for data centers ever stall. Investors such as Steve Eisman point out that a substantial share of Nvidia's receivables is owed by a handful of large customers. The stock therefore sits on a fault line — between a company delivering operationally like few others and an industry publicly arguing about its own pace for the first time.

Chart-wise, the shares are consolidating rather than euphoric: an RSI of 44.7 leaves them in neutral territory, just below the 50-day moving average of 185.49 euros. On the substance of the quarterly numbers, the picture remains one of exceptional operating momentum. Anyone dazzled by unconfirmed billion-dollar commitments risks overestimating how much of that story is already real. And anyone watching Nvidia is ultimately watching a larger question: whether AI progress and AI caution are truly two different things — or whether, in the end, both draw on the same current of computing power.