Nvidia (NVDA) posted a beat and raised its Q3 guidance, which backs up the AI demand story. The market still has a problem. The gross margin guidance came in a bit light, below 75%. The likely cause is heavy AI capital spending, data center obligations, and the cost of supply commitments tied to memory procurement, which is expensive right now.
Numbers no longer drive the stock
The plain numbers in the report barely matter anymore. What matters is the color commentary and nuance from the call. Nvidia (NVDA) has now beaten 15 reports in a row, and this is the fifth straight time the stock dropped right after. People expect the objective numbers to be a beat, so the beat is already priced in. On this report the stock was down about 1% after an initial dip and some chop.
"Compute is revenue"
Jensen Huang put a quote in the front of the report: compute is now revenue. He is trying to change the narrative. A few weeks earlier he called chips an investable asset class. The pitch: with the tokenization of everything, the hardware can be rented out, so investors should worry less about the capital spending and less about the bottlenecks. The message is that what he is building is bigger than anyone can predict, which is meant to ease those fears. For now Nvidia (NVDA) will keep reporting the same way, because the numbers keep coming in; the real test is whether Jensen can paint the buildout as still happening, still relevant, and only in its first inning.
What matters on the call
I want to hear how Jensen talks about the growth path, how Blackwell transitions to Rubin, and how he handles the bottlenecks. The known bottlenecks are electricity supply and the politics around data centers. Any of these, whether a real risk or just a perceived one, can act as a small poke in a bubble and move things. There are open questions and posturing around circular financing and around the new Rubin architecture and its on-ramp - that is what people most want to hear about.
Many Mag 7 CEOs can talk up AI, but Jensen is the captain of the ship. If he is not defensive, sounds optimistic about the transition and buildout, and explains what "compute is revenue" really means, that could turn investors his way. Josh Brown calls Jensen the "Fed chair of AI," because what he says has ripple effects across the whole market. He has also been called the pied piper of the AI trade, which has been struggling recently.
Jensen vs. the other AI voices
Mark Zuckerberg (META) a few weeks ago was lamenting that he is not in first place anymore and said he wants an open system. Sam Altman of OpenAI has talked about wanting to outsource parenting to GPT. These comments do not move the needle. Every Mag 7 CEO is on the AI rocket ship, but Jensen Huang is its captain. He is more reserved, and what he says carries far more weight. Nvidia (NVDA) has gone from a gaming company 10 years ago, to the center of AI, and now toward being the backstop of AI - the "bank of AI." Jensen believes he can be a rising tide that lifts all ships. Not everyone bought that pitch the day he appeared with Wall Street figures and said chips could be an investable asset class with "new check writers in town." Whether he is actually backstopping the trade remains to be seen.
Depreciation, circular financing, and the two AI markets
Expect a lot of questions on depreciation and on the circularity of financing. I am not worried about depreciation. The Blackwell chips are still very strong, the transition should be smooth, and even the quieter over-the-air software updates Nvidia (NVDA) pushed to past chips keep them powerful.
Split the buildout in two. There is the B2B AI buildout, which Nvidia leads, and the B2C side, which is a separate conversation and could be a bubble - that side involves OpenAI and Anthropic. Jensen feels very good about the B2B side. This is not like the internet bubble; the numbers and the revenue being generated are very strong. On hardware he is best in class, and because his chips are best in class there will be a premium. That premium lets him skirt around "RAMageddon" and rising component prices, because he is the leader with the best hardware.
China and shiny objects
China still matters. Jensen has a strong relationship there and should handle it in a smart way. Ideally he will not make it a focal point on the call and will keep it managed in the background, but it stays a threat. Robotics and similar projects get a lot of social feed attention and look like they are advancing fast, but a lot of that is shiny object syndrome. Jensen is truly focused on building out the infrastructure.
Working with many Fortune 500 companies putting AI in place shows it is not easy to implement and it is not a magic silver bullet or a light switch. Once it is actually integrated, real revenue shows up. That is why Jensen is confident: he knows he is building the new infrastructure of the internet.
