AI: Nvidia Reigns on sans China Q1 2027. AI-RTZ #1093

Yes, another Nvidia earnings post with ‘Nvidia Reigns’ in the title. Regular readers may notice that this has been quite the tradition for some time here at AI-RTZ now. This Q1 2027 is not different. Founder/CEO Jensen Huang remains in paranoid peak form, covering all the flanks and corners of AI opportunity ahead.

And yes, it’s another strong and notable quarter without (sans) China in the results, just like quarter. It’s now an unfortunate feature of Nvidia’s quarters this AI Tech Wave. Hopefully there’ll be a turn in that story, perhaps after the second US/China trade meeting this September, this time at the White House. In another Universe, China likely added tens of billions to Nvidia’s quarter just reported.

But because of the secular global demand for AI training, inference, and everything in between to drive AI applications and services, the Nvidia of this universe did not seem to miss China that much. Both in their results, and in the fundamental and financial markets. For now.

Let’s go through the strong results.

The Information summarizes it well at a high level in “Nvidia Projects 95% Sales Growth in the Current Quarter”:

“The AI boom is far from peaking, Nvidia results suggest.”

“Nvidia’s revenue grew 85% to $81.6 billion in the three months that ended April 26, the company said Wednesday. That’s 12 percentage points higher than the blistering growth the AI chip designer reported in the January quarter. It also projected revenue would grow 95% in the current fiscal quarter, suggesting that the AI boom is far from peaking and its Blackwell-series chips are hitting their stride. The projected current-quarter results would mark the fourth consecutive quarter of sales-growth acceleration.”

So China could have added another $10+ billion in revenues potentially in another Universe, but the results were really great around the world without China. That the growth is accelerating is of course further evidence of fundamental, secular momentum. With rising GAAP gross margins, at 74.9% this quarter vs 60.5% last year due to inventory charges.

It helps to put these numbers in perspective relative to peers, like comparing an orange to a grape.

“To put the growth in perspective, Advanced Micro Devices, an AI chip rival about seven times smaller than Nvidia in revenue, said earlier this month that its revenue rose 38% in the first quarter of the year. Both companies sell graphics processing unit for AI but are also benefiting from a boom in central processing unit sales; Nvidia said it is on pace for nearly $20 billion of standalone CPU sales this year—separate from the CPUs attached to its GPU chip systems—to power the rise of AI Agents like Anthropic’s Claude Code and OpenAI’s Codex coding agents.”

The CPU callout is an important one for Nvidia. It’s a new category for Nvidia as the world moves from AI chatbots to AI Reasoning and AI Agents on the way to ‘AGI’.

The last two need far more CPUs for on the spot inference. Resulting in an opportunity for Nvidia’s product line after the current Blackwell to Vera Rubin.

Jensen Huang called out the CPUs as a $200 billion+ new addressable market opportuity for the company.

And it grows fast with more AI Inference. Which grows even faster with more AI Reasoning and Agents.

Vera represents Nvidia’s own ARM based CPUS, and Rubin the next generation AI GPUS. Just to fill out the tech product descriptions, LPX represents new growth in AI low-latency Inference products, bolstered by Nvidia’s recent acquihire of Groq a few months ago. And Spectrum X is Nvidia’s Ethernet connectivity for it all, which this quarter surpassed regular Ethernet markets globally.

CFO Colette Kress highlighted how Vera Rubin will start to ramp in shipments in the second half this year, getting to meaningful numbers by Q1 2028. So this time next year.

Going back to the Q1 2027

“Nvidia’s free cash flow rose 86% to an astonishing $48.6 billion compared to the same period a year ago, and it projected steady gross profit margins in the current quarter. Nvidia rewarded shareholders by increasing its stock dividend to 25 cents a share from a penny a share. It also continues to buy back more of its stock every quarter, including $20 billion in the April quarter. The company said it had authorized another $80 billion of buybacks.”

Investors tend to like those kind of moves. Which brings us to current valuations.

“Despite the booming results, the stock is trading at 26.7 times its forward earnings, according to S&P, down slightly from its P/E ratio three months ago. Shares fell slightly in after-hours trading Wednesday after rising 18% so far this year, giving Nvidia a $5.4 trillion market capitalization.”

All this despite the ‘Frenemies’ nature of the business, particularly with the major LLM AI model and cloud companies (aka Hyperscalers), who represent roughly half of a globally accelerating business.

Jensen specifically called out Anthropic as the most recent major customer in that bucket, that is ramping its use of Nvidia infrastructure, despite its use of Amazon and Google AI chips.

“As Nvidia’s profit machine gets bigger, its biggest customers are trying to lessen their reliance on its chips, but that hasn’t eaten much into its share of the industry’s profits yet. Meantime, some of those customers are likely generating better gross profit margins from renting out Nvidia chips as demand rises, leading to a growing shortage among startups and other rental customers, The Information has reported. Underscoring that demand, Nvidia CFO Colette Kress said Wednesday that prices to rent Nvidia chips that launched three to six years ago had risen 15% to 20% so far this year.”

That last point of rising prics for older Nvidia chips is a point worth pausing on. Depreciation accounting on AI GPU chips, has been a key concern for bears on this AI wave thus far, arguing that the depreciation estimates for AI Data Center infrastructure and chips should be shorter than currently assumed. So say 2-3 years vs 4-5 or longer. Nvidia is highlighting demand for AI infrastructure up to six years old.

The overall numbers ahead are far from zero sum, in a globally expanding market for AI infrastructure.

“The company said its biggest customers are on track to triple or quadruple their capital expenditures to between $3 trillion and $4 trillion, including chip purchases, between 2027 and the end of the decade. Kress said Nvidia’s share of the market for powering the most advanced, “frontier” AI models “will grow significantly,” and CEO Jensen Huang called out Anthropic as an important, relatively new customer of Nvidia chips at a large scale. He reiterated his confidence that the upcoming Vera Rubin-series chip systems will be even more successful than the current Blackwell systems.”

And of course the ‘glue gear’ for all these chips also did well for Nvidia in the quarter.

“Nvidia also boosted revenue by tripling its sale of networking equipment that connects AI servers powered by its ships, showing how the company is maintaining its dominance by getting cloud providers and other data center hardware customers to buy a broader suite of products.”

All this underlines the fundamental, secular AI momentum this AI Tech Wave not just for Nvidia, but for its customers. And the customers’ customers. Up and down the AI tech stack below.

WIthout China for now. Which has its own dynamics as the second largest AI market. Perhaps soon Nvidia will get a chance to be back there again. In this Universe. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)