Skip to content
Cutting Through the Noise: Nvidia’s Outlook Signals We’re Still Early in the AI Buildout
Nvidia
Shares of NVDA are down 3% in after-hours on a fractional miss in July datacenter revenue and a October revenue outlook that fell below the whisper. Adjusting for the $2B–$5B in China H20 revenue excluded from the October guide, the outlook lands about 2%-7% above the October revenue whisper. Additionally, Jensen dropped a nugget on the call hinting that growth in CY26 could be closer to 50%, well above the Street's 29% expectation. The bottom line is we're still early in the AI buildout, and growth next year will be higher than most expect.

Key Takeaways

The less than 1% miss in July datacenter revenue is irrelevant. The October guidance included noise that initially appeared below the revenue whisper. Adjusting for China H20, guidance was roughly 2% above the whisper. Taken together, this suggests the business is performing better than investors had expected.
Jensen dropped a first read on CY26 growth, suggesting it could be up around 50%, above the Street's 29% growth outlook. While he stopped short of endorsing that exact number, the cat is out of the bag.
1

The October Guide

For starters, I’m going to throw out the less than 1% Datacenter miss in July. It’s too small to matter and guidance is always more important.

At first glance, the October guide fell flat, calling for $54B in revenue, plus or minus 2%. That compared to the $55B whisper number. However, the Street’s published estimate was $53.4B, meaning the $54B guide was actually above the official number. Still, in the near term, the whisper carries more weight.

Importantly, that $54B expectation does not factor in any contribution from China. Trying to back into how much China was in the $55B whisper is more of an art, but here goes. Back in mid July, when it was announced that Nvidia could start to sell the H20 in China, the Street was looking for $51.5B in sales for October. Following that announcement, along with the hyperscalers reporting, the numbers increased by $2B. How much of that was China related is hard to say.

Additionally, only 18 of the 42 analysts updated their numbers, which means the average analyst who factored in events since mid July increased their estimates by about $4B. That $4B increase is what drove the whisper to $55B. We also don’t know how many analysts had China H20 revenue for October in their model prior to the curbs coming off. Some may have anticipated this playing out just as it has. All of that said, we can conservatively say there was $2B in China expectations in October estimates. Backing that out implies the whisper for October was $53B, and the $54B actual guide was 2% higher than the whisper.

On the call things got better. They said they could see between $2B and $5B in China revenue in the quarter if the licensing comes through as expected. That means the high end could be $59B, 7% ahead of the $55B whisper that included China.

If they come in at $59B, growth would be up 68% y/y. That compares to growth in July of up 82% when adding back the $8B in China revenue that they did not reel in. In other words, growth rates are slowing, but not as fast as many had feared.

2

CY26 Growth Should Be Higher Than the Street

On the call, Jensen dropped a nugget on next year’s growth. He said China could be $50B in revenue this year if fully sold in, and could grow 50% annually. In a follow up question, Jensen was asked if that 50% China growth (per year, so more of a CAGR) is a reasonable bogey for how much datacenter revenue should grow next year.

He would not outright endorse the 50% number, but his language seemed to support it. Three comments from Jensen stuck out to me:

  1. “I think the best way to look at it is we have reasonable forecasts from our large customers for next year, a very, very significant forecast.” I assume he is referring to significant growth.

  2. “And we still have a lot of businesses that we’re still winning and a lot of start-ups that are still being created.”

  3. “AI native start-ups (OpenAI, xAI) are really scrambling to get capacity, so that they could train their reasoning models. And so the demand is really, really high.”

My take: reading between the lines, if China grows a bit faster than the rest of the world, Jensen is implying the overall business will grow around 40–50% next year. The Street is at 29%. And getting comfort on next year’s growth is the key to getting the Nvidia investment right over the next six months.

Disclaimer

Back To Top