Nvidia's First Quarter Without Him, and the Questions the Numbers Cannot Answer

The numbers are the easy part. Nvidia guided to $108 billion in revenue for its fiscal third quarter, give or take two percent. The street sits between $108.7 billion and $109.3 billion, which would be roughly ninety percent growth year over year. Adjusted earnings per share should land near $2.47, up from $1.30 a year earlier. Data center revenue is expected around $101.8 billion. Gross margin, 74 percent.
None of that is in doubt. The company has beaten and raised for two straight quarters, and the order book is not the constraint. Supply is. So when Nvidia reports on November 17 after the close — the first earnings release in thirty-three years without Jensen Huang at the helm — the interesting part is not the quarter. It is what interim CEO Colette Kress says about the next four.
Three questions, one microphone
Kress has co-hosted every one of these calls for years. She knows the format, she knows the analysts, and she has spent thirteen years as CFO explaining this business to the same people. What she has never done is open the call.
Huang rarely gave prepared remarks. He would take the questions and treat them as a conversation, wandering into product philosophy, supply chain war stories, and the occasional digression about the price of electricity. Morgan Stanley's note ahead of the print warns that this — the freewheeling part — will probably shrink. That is not a criticism. It is just what happens when the founder leaves the room.
The three things worth listening for are specific. First, the fourth-quarter guide. Consensus is $124.3 billion to $126.4 billion, with Wells Fargo at $126.4 billion and Wedbush floating a range up to $130 billion. A new interim CEO has every incentive to guide conservatively, and a conservative guide can swallow a beat. Second, Vera Rubin. Management said production shipments began in August and expected Rubin to be about twenty percent of data center revenue in the third quarter. KeyBanc has already cut its 2026 Rubin GPU shipment estimate from 1.7–1.8 million units to 1.5 million on HBM4 supply, with GB300 filling the gap. Third, and least discussed: whether Nvidia still stands behind its fiscal 2028 framework of roughly seventy percent revenue growth.

The guidance question nobody wants to ask
In August, management framed fiscal 2028 growth at about seventy percent, justified by customer demand growing closer to a hundred. Morgan Stanley now models $623 billion for the year against a consensus of $694.8 billion. That gap — around $70 billion — is the sound of analysts quietly derisking a number the company has not withdrawn.
Morgan Stanley's note says the quiet part: with an interim CEO in place and a lot of open questions, they are "not sure the company will give guidance" for next year at all.
Think about what that means. Nvidia's entire valuation rests on visibility — the promise that orders placed today become revenue in four quarters. If the company stops giving a framework, the market has to build its own, and the market's version will be more conservative than the company's. That is how a multiple compresses without a single bad number.
There is a defensible reason to withhold. The demand side has real physical constraints now: land, power, substations, the time it takes to energize a hall. SoftBank, a partner on one of Nvidia's projects, took a $3.7 billion impairment on power equipment tied to data center buildouts. You cannot ship GPUs into a building that has no power. Morgan Stanley's read is that the 2028–2029 window carries more risk than 2027, because halls already under construction will not be stopped by new restrictions, but the next wave might be delayed.
The margin is the tell
Gross margin is where this quarter gets honest. Nvidia did 75 percent in the second quarter and guided the third to 74 percent, plus or minus fifty basis points. Management has already said the fourth quarter lands at 71 to 72 percent, and that the recovery to 72–73 percent comes in fiscal 2028 as price increases take effect.
Why does a single point matter? Because on roughly $303 billion of trailing twelve-month revenue, one point of gross margin is about $3 billion a year. Three points is $9 billion, which is more than most semiconductor companies earn in total.
The cause is memory. HBM4 and conventional DRAM have both been climbing, and Nvidia is not immune just because it buys in enormous volume. There is a second pressure arriving next year: chip designers say they have received wafer pricing notices from TSMC and others in the ten to twenty percent range. If those costs are not passed through, the impact is four to five points of gross margin.
So the sequence to watch is unglamorous. Does the fourth-quarter margin guide come in at 71 percent or below? If yes, the story is that input costs are eating more than management admitted in August. If it holds at 72, the price increases are working, and the fiscal 2028 recovery story survives.
Two numbers that are bigger than the quarter
The first is $145 billion. That is the total supply Nvidia had committed to as of the August call, including inventory purchase commitments and prepayments. It is a bet that demand shows up on schedule. If a hyperscaler pushes a hall by two quarters, that inventory does not disappear; it sits there, aging, while the next platform ramps.
The second is north of $99 billion in equity investments. Nvidia has been putting money into the customers and labs that buy its hardware — strategic positions in AI labs and funding support for neocloud providers. In a boom, this is a flywheel: you finance the buyer, the buyer buys your product, the revenue compounds. In a slowdown, it is a circularity that analysts will pick apart line by line.
Neither number is new. What is new is who has to defend them. Huang could explain the flywheel as strategy, and the room would nod. Kress has to explain it as accounting, and the room will ask about impairment.
What a good call looks like
A good call does not require a beat. It requires three sentences: the fourth quarter is on track, Rubin is ramping on schedule, and the fiscal 2028 framework still holds. That is it. If Kress says those three things clearly and takes six questions without a headline about the investigation, the stock can start rebuilding the 25 percent it lost.
A bad call is easier to imagine. A vague answer on guidance. A deferral on Rubin timing. A question about the board's review that gets answered with "we can't comment on an ongoing matter" three times in a row. None of that is a business problem, but all of it is a discount problem, and the discount is already 25 percent deep.
The strange thing about this quarter is that the company is reporting into a vacuum it did not create. The demand is real, the backlog is real, the roadmap is real. What is not real yet is the answer to a question no financial statement contains: who decides, and will the market trust that person the way it trusted the last one?
Kress gets one call to start answering it. The numbers, for once, will not do the work for her.
Related articles
AMD Got 19 Percent for Being the Other Option
Money did not leave AI. It left the one name with a governance problem and went to the two without one.
The $99 Billion Loop Nvidia Built Around Its Own Customers
Selling shovels is a good business. Lending money to the people buying shovels is a different one.
The 74 Percent Gross Margin Has a Shelf Life
Three forces arrive in the same four quarters, and each one costs about $3 billion per point.
175,000 Signatures Will Not Put Him Back. Here Is What Would.
The longer the campaign takes to win, the weaker its case becomes. That is the trap built into it.