Nvidia at 15 Times Earnings: Wall Street's New AI Safe Haven?

Nvidia shares have gained just 14% this year — barely ahead of the S&P 500 — yet some on Wall Street now argue the chip giant could serve as a defensive play inside an AI trade that investors are re-examining.

The math behind that argument is valuation. Nvidia trades at roughly 15 times estimated earnings per share over the next 12 months, making it the third-cheapest stock in the PHLX Semiconductor Index, according to Dow Jones Market Data. That multiple, paired with the company's dominant share of the AI chip market, has analysts asking whether the stock can behave like a safe haven while riskier AI plays sell off.

Seaport Research analyst Jay Goldberg posed the question directly on Thursday: could Nvidia become "a safe haven in a sea of uncertainty" as the sector grows more volatile? His answer, notably, is no — or at least not yet.

What do the bulls see?

Goldberg is the only analyst tracked by FactSet who rates Nvidia at sell, and the bull case comes from elsewhere. Jamie Meyers, a senior analyst at Laffer Tengler Investments, told MarketWatch that Nvidia "is trading like a value stock, but the growth is astronomical."

Meyers expects the transition from Nvidia's Blackwell platform to its upcoming Rubin platform to drive "ridiculous growth in sales," with strong underlying AI demand continuing to improve. In his view, that combination makes Nvidia one of the "more defensive tech plays" heading into a period of expected market volatility.

Jake Behan, head of capital markets at Direxion, frames the defensive case around Nvidia's position in the stack. The stock, he told MarketWatch, "can be a clean place to hide, as it's selling the infrastructure that underlies competing strategies." While frontier AI labs and major cloud providers compete to build and serve the best models, their heavy spending ultimately benefits Nvidia, Behan said in emailed comments.

Hendi Susanto, a portfolio manager at Gabelli Funds, cited Nvidia's dominant chip-market position, visibility into future demand and strong long-term earnings potential as reasons its current valuation looks attractive.

The caveat: Nvidia is no Treasury bond. A traditional safe haven it is not — only relatively safer within a risky trade.

Why is the lone sell-rated analyst worried?

Goldberg's concerns center on the neoclouds — companies that buy Nvidia's graphics processing units and rent out computing power — and the debt they have taken on to do it.

"Any valuation-modeling exercise for the neoclouds shows that their terminal value is almost entirely driven by long-term capex," he said. "At some point, that needs to taper down for the neoclouds to have any value."

Nvidia maintains a tight relationship with CoreWeave, a leading neocloud. To keep demand flowing, Goldberg argues, Nvidia must keep rolling out new products, raising prices and offering attractive financing packages. That pushes neoclouds deeper into debt and invites more competition.

"If the neoclouds fail to lift off, that doubles back on Nvidia with reduced sales and even more reduced margins," Goldberg said.

He flags two additional risks. Labor and power shortages could limit next year's planned 55 gigawatts of compute power to somewhere between 20 and 25 gigawatts — a potential shortfall of more than half. Nvidia also "risks overtaxing its balance sheet," he said, pointing to roughly $250 billion in commitments made in the last quarter alone. The company can afford the spending, he acknowledged, but "it is also clear they are just [getting started]."

Goldberg expects the compute market to saturate eventually, with the neoclouds Nvidia now supports coming to rely on the company as a backstop when demand slows. Waning interest could dry up Nvidia's cash flow — not a near-term risk, he noted, but one investors should keep in mind.

What happens next?

Nvidia's stock could benefit over the next six months as chipmaking rivals run into supply constraints, Goldberg concedes. But he reads the company's financing deals with neoclouds as a sign that competitive pressure is already building — a warning that even a defensive AI trade carries leverage of its own.