Nvidia

Nvidia’s April quarter was another reminder that the AI infrastructure trade is still in the second inning. While shares trading down 1.7% in trading the following day, hides the bigger point. Nvidia’s core revenue growth accelerated to 109% y/y in April from 89% in December, adjusting for China, and July guidance implies about 95% growth before what will likely be another beat. The harder question is not whether Nvidia’s business is exceptional, it is whether the stock can keep outperforming given the law of large numbers.

The U.S. approval for Nvidia to sell H200 GPUs into China creates a call option for an incremental 3-5% of revenue in CY26 and CY27. At the high end, I believe China could add 10% to CY27 estimates. I caution that, even though China can drive some upside, the unpredictable nature of future China revenue likely means NVDA won't get credit for that upside. In the end, Nvidia doesn't need China because growth has accelerated without China. In the April quarter that will be reported this Wednesday, May 20, the Street is expecting revenue growth of 79%. Removing the China revenue from April of last year to get an apples-to-apples growth rate, sales are expected to be up 100% y/y, up from 89% growth reported in the recent January quarter, adjusting for China.

I believe we're still in the second inning of AI, and shares of NVDA will underperform the broader silicon trade over the next six months. Nvidia reports on May 20th, and expectations began to get amped up two months ago at the March GTC event, where Jensen projected the company would see at least $1T in cumulative data center revenue through CY27, which implies about 40% growth, vs. the Street currently at 32%. Since those March comments, NVDA is up 18%, vs. AMD up 129% and INTC up 183%. The issue is that investors know the April quarter will be strong and that numbers for CY26 will likely increase by 10%, but they struggle with the the law of large numbers resulting in slowing growth next year and the risk of custom silicon. Unfortunately for NVDA, the company likely won't be able to ease those fears on the April call.

Jensen’s keynote reinforced a simple point: demand is tracking well above even high expectations, while investors remain concerned that growth beyond 2027 could slow sharply or even decline. Shares of NVDA fell 3% in the three days following GTC, compared to a 1.4% decline for the Nasdaq.

At next Monday's GTC 2026, Jensen will make it clear that demand continues to run ahead of investor expectations. This will likely be viewed by investors as a non-event, given it's consistent with their earnings comments three weeks ago. Investors will be keyed into any impact the conflict in the Middle East is having on business; however, I believe the impact is not measurable. The stock has a bigger challenge than a read on the next six months: the investors' "wall of worry" that growth will drop off in CY27. That reality is evidenced by shares being down 7% since they reported earnings on Feb 24th, compared to the Nasdaq being down 4%.

It's been two days since Nvidia delivered January results and April-quarter guidance that cleared both Street and whisper expectations, yet the stock has traded down 9% since earnings compared to the Nasdaq being down 2%. Investors are already looking through what will be another monster year in 2026 and landing on anxiety over what 2027 will hold. The biggest risk to the stock is that fundamentals continue to exceed expectations and the "anxiety year" gets pushed out to 2028, resulting in shares remaining rangebound. As for the broader AI takeaway, Nvidia’s guide is the latest evidence that we're still in the early innings of the AI buildout.

Heading into Nvidia’s earnings next Wednesday, expectations are being shaped by two developments over the past month: AI models are showing clearer utility, and the hyperscalers guided to a bigger infrastructure buildout. During that time, Nvidia growth estimates for CY26 have increased from 50% to 55%. The demand story is well understood, yet NVDA shares are up only about 5% over the past month versus the Nasdaq roughly flat. The disconnect between the bullish updates and modest share price increase highlights that the real debate is what growth looks like in 2027 and 2028. Ultimately, investors have to decide what inning of the AI buildout we are in, if it's the 5th inning, 2027 growth should look more modest, and if it is the 2nd inning, which I believe, Nvidia’s growth outlook over the next several years remains robust.

Nvidia gave three incremental updates at CES that pointed to upside this year. The first was Jensen’s comments in the keynote twice describing demand as “skyrocketing.” The second was CFO Colette Kress saying at the JPMorgan fireside that their previous target calling for a $500B data center revenue outlook through 2026 has “definitely gotten larger.” The third was Jensen’s comment that demand for the relaunch of the H20 in China was “quite high.” Shares of NVDA shrugged off the good news, trading down 4% in the three days following the updates, compared to the Nasdaq which was flat, underscoring investor concern that it's a function of time before growth stalls. I believe growth will be higher for longer, driven near term by the hyperscalers and long term by Physical AI.

The 2% decline in NVDA shares following October results and January guidance, both of which were above whisper expectations, reflects investor concerns that, when it comes to AI, it can’t get any better. While the concern is understandable, most signs suggest that growth for Nvidia and the broader AI infrastructure buildout is still in its infancy.

I remain positive on shares of NVDA and believe over the next two years growth will be higher for longer. That said, the cross currents around next week’s earnings set up a Catch-22 for the AI complex, because stronger guidance can amplify worries about overspending, while a modest raise can be read as the first sign that growth is normalizing faster than expected. As for changes to Street estimates, Jensen’s comments on October 28 suggest the consensus CY26 growth estimates will increase from the current 39% y/y growth to 45%.

Over the past two months, Nvidia has made five major investments totaling $15B in near-term commitments and $100B in long-term commitments. During that same period, the company has also invested in approximately seven private companies and established twelve new partnerships. In total, Nvidia has made about twenty-five strategic announcements. The common thread that ties them together: Jensen Huang’s belief that the future of work is shifting from tools to bots.

Here are the five pressure points for this Wednesday and Thursday’s earnings from Google, Meta, Microsoft, Amazon, and Apple (and an Nvidia bonus). Big picture: The setup carries more risk than it did three months ago, as investor confidence is notably higher heading into this round of reports. That said, the key takeaway will likely remain that we are still in the early stages of AI adoption, and the largest technology companies continue to hold their ground.