Nvidia Q4 earnings: Can it push the AI trade higher — again?

Wall Street used to ask companies for a good quarter and a nice sentence about “momentum.” Then, it started treating good quarters as background noise. Then, it demanded great quarters, decided those were predictable, and moved the bar again — faster than most companies can move product. Now, it wants more.

Nvidia $NVDA reports Wednesday after the close into a market that has already assumed brilliance, already priced in fireworks, and already moved on to the next question: Can the company narrate acceleration, not just scale, in guidance, margins, and the cadence of the next quarter?

Nvidia matters so much that the market has turned it into a referendum. The quarter is expected to be enormous: Zacks Consensus Estimates have Nvidia’s fiscal Q4 2026 sales reaching $65.56 billion (up 66.7% year over year) and EPS at $1.52 (up 70.8%). Nvidia’s own official outlook is: revenue $65 billion ±2%; gross margin 74.8% GAAP/5% non-GAAP ±50 bps, opex $6.7 billion GAAP/$5. billion non-GAAP. And if you want to see where the quarter “lives” inside that $65-ish billion, the Street has been modeling a Data Center number around $58.7 billion, with roughly $9 billion in networking and about $51.1 billion in compute, plus gaming around $4.3 billion and automotive around $663 million.

That’s all huge. Obviously. But Nvidia is heading into earnings week in a very modern bind: the quarter can be massive, the numbers can be pristine, and the stock can still respond with the emotional range of a spreadsheet.

Options traders are leaning into a big reaction anyway. Investopedia’s options read pegs the implied move around 6% by the end of the week, even though the median day-after move across the past 10 earnings reports is 3.2%, according to MarketWatch. And the setup is extra touchy because whisper numbers could be even higher, while short bets have been growing. So the reaction will hinge on whether the forward story feels like acceleration rather than continuation — and whether that acceleration looks durable in an AI trade that is getting more anxious, more crowded, and more expensive quarter by quarter.

The company can clear consensus and still disappoint a crowd that paid for a bigger and louder fireworks show. Meet the expectations treadmill: A good quarter gets a shrug. A great quarter gets a shrug. Now, the market’s daring Nvidia to deliver a monster quarter — and wondering whether even that’s enough if the forward story doesn’t feel like acceleration.

Priced for perfection — and then some

Nvidia’s stock has become a market object. It’s big enough to tug the major indexes, with a 7.8% weight in the S&P 500. When it moves, it drags money and sentiment with it, which helps explain why people talk about its quarterly reports as if they’re a major weather event.

Analysts may be looking for $65-ish billion now, but the fixation has shifted to what Nvidia says about April; UBS analyst Timothy Arcuri said in a recent note that “investor expectations for Nvidia’s April quarter likely demand revenue somewhere in the $74 billion-$75 billion range.” And the April obsession is a Blackwell obsession. CEO Jensen Huang handed the market a tagline last quarter: “Blackwell sales are off the charts, and cloud GPUs are sold out.”

The EPS dispersion tells you how nervous the market is about Nvidia’s forward curve. For the coming fiscal year, analysts average $7.76 in earnings per share, but the range runs from $6.28 to $9.68 — a spread that shows the market admitting it’s arguing with itself about whether this is a long runway or a late-cycle sprint.

Marta Norton, a strategist at Empower, told Reuters that “the expectation for outsized results for Nvidia has been a persistent theme over the past few years, and so it’s hard for Nvidia to surprise when everyone expects it to surprise.” The baseline has moved so far upstream that a beat can feel like confirmation, not upside.

The stock has been trading like expectations have outrun imaginations; it’s spinning in circles. The consensus bull case has been so widely adopted that it has stopped being a catalyst, which leaves the stock drifting until management delivers new information that changes the slope — the next-quarter guide, the cadence of the ramp, the margin outlook.

And as skepticism about AI mounts, Wall Street is increasingly worried Nvidia’s report could weigh on the shares regardless of what the results say. When the market is skeptical about the AI trade, the reaction becomes all out of sorts. It doesn’t take a disaster to sell off; it can take a guide that looks merely excellent, a margin comment that sounds careful, or a ramp update that feels like a constraint story rather than a demand story. The upside requires a lot to go right — all at once: pipeline confidence, margins that don’t blink, and a forward guide that reads like acceleration rather than continuation.

The market isn’t asking Nvidia to be good. It’s asking Nvidia to be inevitable.

The market’s AI checkpoint

The AI trade used to be about belief. Now, it’s increasingly about invoices — and whether anyone is getting paid enough, fast enough, to justify the bill.

Bridgewater’s Greg Jensen has described the AI boom as having entered a “more dangerous phase” that is “marked by exponentially rising investments in physical infrastructure and growing reliance on outside capital.” Alphabet $GOOGL, Amazon $AMZN, Meta $META, and Microsoft $MSFT will invest about $650 billion in AI infrastructure in 2026, up from $410 billion in 2025, which is the capex curve investors have been both cheering — and increasingly side-eyeing.

“Compute demand continues to significantly outpace supply, driving hyperscalers to invest even more rapidly to try to someday get ahead of the demand,” Jensen added, powering both the bull case and the anxiety. If Nvidia’s guide comes in clean and strong, it supports the idea that demand is still outrunning supply and the buildout still has urgency. If Nvidia sounds anything less than sure-footed, the market starts whispering about a slower gear.

The market is now debating the slope — how much bigger the next quarter gets, how clean the ramp looks, how defensible margins are, how long hyperscalers keep spending at this pace, and how much of that spend eventually gets negotiated away via custom silicon. Until one of those questions gets answered with enough force to reset the forward curve, the stock can grind sideways even while the fundamentals look spectacular.

Alpine Macro’s Nick Giorgi told Reuters that Nvidia CEO Jensen Huang “has to come out and show his confidence in his own customers.” That’s a remarkable ask for a chip company. It’s also a logical one. Hyperscaler shares have been pressured by worries about returns on capital spending. Nvidia’s commentary gets treated as a proxy read on whether those customers are still spending aggressively — and whether they sound convinced (enough) that the spend is worth it.

The retail-facing version of this argument is even more direct. Zacks stock strategist Andrew Rocco said in an email Monday that “regardless of whether an investor has a position in the stock, Nvidia’s earnings will likely be the most important and watched report this earnings season.” He added that Nvidia, with its $4.6 trillion market cap, “remains the undisputed king of the AI revolution.” While the company’s sheer scale “may give some investors pause,” he wrote, “the underlying fundamentals — characterized by a reasonable 46 P/E ratio and a consistent history of crushing estimates — suggest that the growth story is far from over.”

The post-earnings call will offer, Rocco said, “clues into the health of AI spending in general” and “valuable clues” about the health of Nvidia-adjacent partners. Zacks’ “what to watch” section is basically the expectations treadmill in list form: the demand is already understood, so investors want forward reassurance that hyperscaler spending will continue — plus a narrative about what comes next.

Even the more mainstream “how much will the stock move” coverage has started baking bubble language into the setup. The stock is down about 8% since a late-October peak, amid concerns about an AI bubble — will it burst? won’t it burst? — and a loop-de-loop economy with other Big Tech players. This market is: jumpy, over-positioned, and looking for an excuse to hit the sell button on something that just sounds solid.

The choke point is getting crowded

Demand can be strong and still feel fragile when three forces hit at once: supply, customer leverage, and the trade broadening beyond the headline winner.

Nvidia is tethered to components that don’t have its logo on the box, including high-bandwidth memory. SK Group chairman Chey Tae-won has pledged to boost production of AI-specific HBM chips, alongside a significant increase in SK Hynix’s capital spending. When HBM supply loosens, the forward guide gets easier. When it doesn’t, the guide starts feeling like a logistics forecast.

Customer leverage is the other pressure point, and it’s getting harder to ignore. Nvidia’s multiyear deal to sell Meta “millions” of AI chips is a demand receipt — and a reminder that hyperscalers are building their own Plan B. The deal includes Blackwell now and Rubin later, plus standalone installations of Nvidia’s Grace and Vera CPUs. But Meta is also developing its own AI chips and is in discussions with Google about using TPUs for AI work. So Nvidia’s pitch is to make the escape hatch less attractive by owning more of the stack. 

“Meta has already had a chance to get on Vera and run some of those workloads,” said Ian Buck, who runs Nvidia’s hyperscale and high-performance computing unit, “and the results look very promising.”

But concentration turns that into a market risk, not just a competitive subplot. Meta is widely believed to be among four customers that made up 61% of Nvidia’s revenue in its most recent fiscal quarter. When the spend cycle broadens, that concentration looks like a temporary phase on the way to a wider base. When it concentrates, it starts looking like leverage — a few procurement calendars, a few internal chip teams, and a lot of negotiating power.

And there’s been a good deal of investor discomfort with opaque, circular deals between model builders and infrastructure providers, as well as a rising reliance on outside capital. In a market that’s already jumpy about ROI, tight loops can read as strength or as strain, depending on the week.

Meanwhile, the AI trade itself has started to widen. D.A. Davidson’s Gil Luria wrote in a Monday note that Nvidia’s “may not represent as much of a bellwether for the market” as investor attention shifts toward other AI leaders and adjacent beneficiaries — Alphabet, Broadcom $AVGO, and memory and optical chipmakers — while competition from custom chips such as Google’s TPUs grows more salient. He also ties part of Nvidia’s discount to a specific fear: that 2026 could mark the peak of investments into AI.

But “peak” doesn’t have to mean “demand falls off a cliff.” It can mean “slower,” and “slower” is enough to change how the market prices Nvidia’s next guide.

And then, of course, there’s still China — the overhang that doesn’t care how clean the demand curve looks. Beijing has moved to pause H200 orders, and Chinese customs authorities have reportedly instructed agents to block Nvidia’s H200 chips from entering the country, despite U.S. approval (under certain conditions), which is a reminder that “demand” doesn’t help if the shipment can’t land. China remains a live variable and any commentary about what can ship, what can’t, and where demand is being rerouted will land hard.

Put together, Wednesday’s earnings become less about whether the company prints another blockbuster quarter (it likely will) and more about whether it can tell a forward story that clears all the hurdles at once: a guide that reads like acceleration, a ramp narrative that sounds repeatable, margins that look defensible, and demand that feels durable even as the biggest buyers keep building alternatives. And if Wednesday’s post-earnings call doesn’t settle any arguments, the next reset is already on the calendar: Nvidia’s mid-March GTC conference in San Jose, California, where the company can refresh the roadmap — and the market can decide, again, what (they think) everything means.

A clean upward guide in the range that investors have started treating as the new baseline for April strengthens Nvidia’s “yes, demand still outruns supply” story. Anything that sounds just excellent gives the market room to do what it has learned to do with that excellence: nod, shrug, and start debating where the AI trade’s next winner sits in the stack.