The most important number from the quarter is 109%. That’s Nvidia’s April quarter revenue growth after adjusting out China revenue from a year ago to get a cleaner comparison. That compares to 89% adjusted growth in December. In other words, growth accelerated by 20% in front of incredibly difficult comps.

Put get a sense about how tough those comps are, in calendar 2022, Nvidia’s total revenue was $27B. This year, calendar 2026, revenue is likely to exceed $400B. That is the kind of growth pattern that tells me something bigger is happening, and that’s large tech companies making a bet that the substance of AI will exceed the hype over the next decade.

The July guide implies 95% adjusted growth. My sense is they will beat that, which means the quarter probably lands closer to 100% growth. That essentially means core growth was 89% in January, 109% in April, and likely around 100% in July. That is why I think the AI infrastructure trade remains early.

The distinction between Nvidia the company and Nvidia the stock matters more now.

Above we talked about Nvidia the company. As for the stock, the impressive April quarter and better than expected July guide does not change the stock narrative. The issue is the law of large numbers. The Street is looking for about 40% growth in CY27, which is a major step down from roughly 90% growth we will likely see this year.

The difficult reality is that even if Nvidia surprises next year and grows 70%, investors will likely just shift the slowdown debate to calendar 2028, when expectations may again settle around 30%. That dynamic makes it harder for NVDA to outperform, even if the company keeps beating expectations.

This is why I believe there are better ways to invest in AI from here. Nvidia remains the brain of AI, but the stock is increasingly fighting the expectation that growth must eventually normalize.

China is not a meaningful catalyst for Nvidia.

The July guidance does not include China even though they have approval to sell in China. The problem is the Chinese government has thus far told Chinese company’s not to buy Nvidia chips.

Stepping back and looking at the bigger picture. At most, China could represent 10% of the business, and I see a more realistic best case is closer to 5%. Even with approval to sell chips into China, Nvidia is not yet seeing revenue there.

If China revenue does return, I don’t think investors will give the company much credit for it. They will likely treat it as one-time or fragile because the approval environment can change.

In the end, China is noise around the NVDA story, because while any incremental revenue is welcome, Nvidia doesn’t need China. The core business is doing just fine without it.

One new piece from the call was Jensen’s comment around the standalone CPU opportunity tied to an emerging CPU business, Vera. He suggested Vera could represent $20B in revenue this year. If “this year” means calendar 2026, and the product starts shipping in the back half of CY26, that implies a roughly $40B annualized run rate. That would be big enough to add about 10% to this year’s growth and around 8% to next year’s revenue base.

If instead the $20B refers to the first four quarters of shipments, the impact is closer to half that or adds 4% to overall revenue in CY27. In fairness, the CPU opportunity matters because it helps Nvidia build more novel hardware and keep its overall compute platform at the cutting edge. My point is will it be enough to ease fears of a big growth slow down next year.

On the other hand, Physical AI will be a meaningful driver.  Jensen once again reiterated that physical AI has not started. I agree. Ride sharing this year is likely only about 2% autonomous miles. That means the opportunity around robotics, autonomy, and real-world AI remains largely in front of Nvidia.

As for the impact of Physical AI on Nvidia’s business, I believe a third of overall AI infrastructure spend long term will be related to physical AI. The catch; the theme won’t likely ramp until CY28 or CY29.

Disclaimer