Nvidia's Data Center Machine Aims at $109.3 Billion Quarter

Analysts expect Nvidia to post $109.3 billion in revenue for the third quarter of its 2027 fiscal year, with the data center segment carrying nearly all of that figure, according to the Visible Alpha consensus.

The estimate range for data center revenue alone spans $99.7 billion to $110.8 billion for the quarter, a spread of more than $11 billion that reflects how much uncertainty still surrounds the pace of AI infrastructure spending. Nvidia trades under the ticker NVDA.

The numbers frame a company in transition. Blackwell, Nvidia's current flagship GPU architecture, is expected to generate $148.9 billion in revenue this fiscal year, up from roughly $98.0 billion last year. That is a 52 percent increase. But the Visible Alpha consensus also projects Blackwell revenue to decline next fiscal year as Rubin, the successor architecture, ramps into volume. Rubin is expected to contribute $64.2 billion in full-year revenue.

What does the Blackwell-to-Rubin handoff mean for revenue?

The architecture transition is the central mechanical story in the consensus estimates. Blackwell's rise from $98.0 billion to $148.9 billion within a single fiscal year documents the steepest part of the adoption curve. The projected decline afterward is not a demand signal. It is a product-cycle signal, the standard pattern when a new GPU generation begins absorbing orders that would otherwise have flowed to the prior one.

Rubin's $64.2 billion full-year contribution sits well below Blackwell's peak figure. That gap matters for how investors model the next two fiscal years. If Rubin ramps faster than the consensus assumes, the blended total could exceed current projections. If it ramps slower, Nvidia faces a period where neither architecture is at full run rate.

The $109.3 billion quarterly consensus for fiscal Q3 2027 lands squarely in this transition window. The $99.7 billion to $110.8 billion data center range means the low-end scenario sits nearly 10 percent below the high end. That spread is the market's honest admission that it cannot yet pin down the Rubin ramp.

Why are gross margins expected to fall?

The Visible Alpha consensus projects gross margins declining to 73.7 percent in fiscal 2027, down from 78 percent previously. That compression, roughly 4.3 percentage points, arrives alongside record revenue. The mechanics are consistent with an architecture transition: new products typically carry early production costs, and a competitive AI accelerator market gives cloud buyers more leverage than they had two years ago.

Operating margins are expected to move in the opposite direction. According to the consensus, they should rise because expenses grow more slowly than revenue. In plain terms, Nvidia spends more dollars on research, sales and administration in absolute terms, but revenue grows faster, so the operating margin percentage still climbs.

This combination, falling gross margin and rising operating margin, tells a specific story. Nvidia is trading some pricing power at the product level for scale efficiency at the company level. Whether that trade works depends on Rubin hitting its $64.2 billion mark.

What changed in Nvidia's valuation?

Consensus has Nvidia trading at a price-to-earnings ratio of 11x for fiscal 2027, down from 17x last quarter. The P/E ratio measures a stock's price against its earnings per share; a falling multiple means the price has not kept pace with expected earnings growth.

A 35 percent multiple compression in one quarter is unusual for a company still expected to grow revenue at the rates embedded in these estimates. It signals that investors are applying a discount, and the source identifies two reasons for it.

The first is key-person risk from Jensen Huang's resignation. Huang has been Nvidia's chief executive since 1993 and the public face of its AI dominance. Any departure of a founder-CEO at this scale forces investors to reprice execution risk, particularly during an architecture transition as expensive as the one from Blackwell to Rubin.

The second is ongoing investigations. The source does not specify the bodies conducting them or their subject matter, but their mere existence functions as an overhang, a persistent drag on the multiple that persists regardless of quarterly results.

Why this quarter matters beyond Nvidia

Nvidia's data center segment has become the single largest revenue line in the AI infrastructure economy. When the consensus range for one quarter spans $99.7 billion to $110.8 billion, that spread itself is a barometer for the entire AI capex cycle, because Nvidia's bookings reflect what hyperscalers, cloud providers and sovereign buyers expect to spend on computing capacity.

The stakes run in both directions. The numbers assume the AI buildout continues at a pace that supports a $148.9 billion Blackwell year and a $64.2 billion Rubin follow-on. Each of those figures represents purchases by the largest technology buyers in the world. A quarter that lands at the top of the range confirms the spending thesis. A landing near $99.7 billion would mark the first hard data point that the cycle is bending.

The margin projection sharpens that test. At 73.7 percent gross margin, Nvidia remains one of the most profitable hardware companies in history. The direction, not the level, is what analysts will watch, because continued compression would indicate that pricing power is eroding faster than scale can offset.

What to watch next

The forward question embedded in the consensus is sequencing. Blackwell must still deliver its $148.9 billion year, Rubin must hit its $64.2 billion mark, and the company must do both while absorbing a CEO transition and unresolved investigations at a compressed 11x multiple.

The Visible Alpha consensus of $109.3 billion for fiscal Q3 2027 is the market's current best estimate of where that sequencing stands. S&P Global Market Intelligence, which provides data, news, research and analysis on global public and private companies, supplies the underlying consensus framework. The next earnings report will test whether the Blackwell peak, the Rubin ramp and the margin reset are tracking the numbers the Street has already written down.