If you don’t know about Nvidia & its bull case, you will find everything you need here.
“The age of AI is here.”
Overview.
Revenue. $43.25B | $44.06B | +1.88% beat
EPS. $0.75 | $0.96 | +28.00% beat
$40.1B of buybacks & $244M of dividends
Business.
It already was impossible not be bullish on Nvidia but is now even more. Starting with the global overview with growth absolutely everywhere, as usual.
Before going onto the main subject, it is important to note the acceleration in gaming which comes from their new retail Blackwell hardware which should continue to scale “personal” computing which means available to individuals & not only for datacenter infrastructure, for personal or professional use.
Now on to the datacenter branch which continues to grow extremely fast, beating market’s & management’s expectations despite not selling its H20 in China for half of the quarter - will talk about this after. This quarter is bringing the first confirmations of my bull case, that inference will be a much bigger deal than training & that Nvidia will be the main provider for inference infrastructures.
“The AI workloads have transitioned strongly to inference […] We are witnessing a sharp jump in inference demand. OpenAI, Microsoft, and Google are seeing a step function leap in token generation. Microsoft processed over 100 trillion tokens in Q1, a fivefold increase on a year-over-year basis.”
And most of it is now running on Nvidia’s most advanced hardware which brings me to my second point, the confirmation that hyperscallers & neoclouds spend to have access to upgrade their hardware & not only work from what they have as most are now migrating from Hopper to Blackwell,.
”Blackwell contributed nearly 70% of data center compute revenue in the quarter, with a transition from Hopper nearly complete.”
They do need the best of the best to stay on top & this trend will certainly continue as long as demand for their services grows - which can be the case for years as models, use cases & global usage grow. And the new hardware is already underway.
“B300 GPUs, with 50% more HBM, will deliver another 50% increase in dense FP4 inference compute performance compared to the B200.”
AI is just starting & we will need more compute to achieve a vision where it helps optimizing the world. And Nvidia is the heart of this revolution.
“The era of robotics is here. Billions of robots, hundreds of millions of autonomous vehicles, and hundreds of thousands of robotic factories and warehouses will be developed.”
On the situation with China. As a reminder, Nvidia has a chip - the H20, tailored to respect the U.S. export controls towards China. The new export controls set up by the Trump administration a few weeks ago now include the H20, forbiding Nvidia to sell it while management confirmed they cannot reduce its capacities to answer the new export controls. In brief & as of today, neither Nvidia nor any U.S. GPU maker can access the Chinese market with their actual hardware.
This translated to a $4.5B revenue miss between the inventory which couldn’t be sold post April 9 & the orders that should have been fulfilled later on, and an $8B revenue guidance evaporated for the next quarter - more than $20B revenue evaporated for the full year.
Two important things to discuss about this. First, this gives China the opportunity to catch up on U.S. tech, which is one of Jensen’s arguments I agree with & have shared many times. Until today, China had no incentive to manufacture their own. It isn’t the case anymore & China will without any doubts try to build their own, locally. And it is ridiculous to believe that China doesn’t have the capacities do it - intellectually or practical.
Jenssen explains this better than anyone.
“The question is not whether China will have AI, it already does. The question is whether one of the world's largest AI markets will run on American platforms. Shielding Chinese chipmakers from US competition only strengthens them abroad and weakens America's position. Export restrictions have spurred China's innovation and scale. The AI race is not just about chips. It's about which stack the world runs on. As that stack grows to include 6G and quantum, US global infrastructure leadership is at stake. The US has based its policy on the assumption that China cannot make AI chips. That assumption was always questionable and now it's clearly wrong. China has enormous manufacturing capability. In the end, the platform that wins the AI developers wins AI. Export controls should strengthen US platforms, not drive half of the world's AI talent to rivals.”
Second, this is a huge loss from a business perspective. Management expects the AI Chinese market to reach $50B in a few years. Nvidia’s fundamentals won’t change, but the financials & market appreciation for its stock might without an access to this massive & rapidly growing market.
Although we could also see the other side of the coin & believe that export controls will change if/when the administration understands the need for Nvidia to be sold everywhere, even to rivals, so their infrastructures run on American technology.
We might also see an increase from Singapore’s purchases as Chinese companies could try to bypass the export controls. A dangerous game but which is certainly already happening in small volume.
“For Q2, we expect a meaningful decrease in China data center revenue. As a reminder, while Singapore represented nearly 20% of our Q1 build revenue as many of our large customers use Singapore for centralized invoicing, our products are almost always shelved elsewhere. Note that over 99% of H100, H200, and Blackwell data center compute revenue billed to Singapore was for orders from US-based customers.”
Interesting that management chose to disclose this number this quarter.
In terms of geography, to be even more bullish, the world is finally waking up after the U.S. led the way & proved that AI innovations would bring massive returns. As usual, everyone is late.
“Last week, I was in Sweden to launch its first national AI infrastructure. Japan, Korea, India, Canada, France, the UK, Germany, Italy, Spain, and more are now building national AI factories to empower startups, industries, and societies.”
Better late than never.
Last point is about the Trump administration’s wish to manufacture onshore, which is an objective shared by Nvidia as the company is partnering to fund four robotized fabs with different partners.
“Process qualification is underway, with volume production expected by year-end. Our goal from chip to supercomputer, built in America, within a year”
This is a nice goal but it will also raise the price of their hardware, without any doubt. For now, demand seems to follow but it’ll be interesting to see how things go when those hardware will hit the market.
I’d like to share a few words on AMD & on why I personally do not believe in the bull case many share lately - being that they would grow massively “once” they are able to take market shares from Nvidia, which to me, won’t happen.
Why? Because the sector isn’t about benchmarks anymore, nor about which hardware is the best, it is about network effect & switching costs. Why would neoclouds or hyperscalers migrate to AMD? Because their GPUs are a bit cheaper?
What about the cost of the infrastructures, meant to optimize Nvidia’s architectures? What about their engineers’ competencies, which are focused on Nvidia & CUDA? What about robotic products which are built to integrate Nvidia’s GPUs? What about training & inference models optimized on Nvidia’s logic? Infrastructures aren't as simple as unplugging something to plug in something else.
Nvidia has the crown and it's glued to its head.
Financials.
As good as usual, even better than we could expect considering the China situation.
A 79% YoY & 12% QoQ revenue growth, while not selling to China for a month… As said, above market’s & management’s expectations.
The rest is harshly impacted by the H20 inventory not sold, as the hardware was built, hence shows up in the costs without generating any revenues. Management shared revenues excluding those, which are what Nvidia usually shares, with 71.3% gross margins, slightly above what we could have expected considering Blackwell is still ramping up. They confirmed during the call that the next quarter will be back above 70% gross margins.
In terms of cash, the balance sheet remains very strong with $43B of net debt & $26B of FCF with no shares dilution. Cash machine.
Guidance.
Strong impact without the Chinese market but still remarkable growth.
The company still expects sequential growth while not selling any hardware to China the next quarter… We would have been talking about 20% sequential growth without the new export controls.
Nvidia is a beast.
My Take & Valuation.
I won’t conclude myself & will simply use Jenssen’s last words. I got nothing better to add to them.
“This is the start of a powerful new wave of growth. Grace Blackwell is in full production. We're off to the races. We now have multiple significant growth engines. Inference, once the light of workload, is surging with revenue-generating AI services. AI is growing faster and will be larger than any platform shifts before, including the Internet, mobile, and cloud. Blackwell is built to power the full AI lifecycle, from training frontier models to running complex inference and reasoning agents at scale. Training demand continues to rise with breakthroughs in post-training and reinforcement learning and synthetic data generation. But inference is exploding. Reasoning AI agents require orders of magnitude more compute. The foundations of our next growth platforms are in place and ready to scale. Sovereign AI, nations are investing in AI infrastructure like they once did for electricity and the Internet. Enterprise AI must be deployable on-prem and integrated with existing IT. Our RTX Pro, DGX Spark, and DGX Station Enterprise AI systems are ready to modernize the $500 billion IT infrastructure on-prem or in the cloud. Every major IT provider is partnering with us. Industrial AI, from training to digital twin simulation to deployment, NVIDIA Corporation's Omniverse and Isaac Groot are powering next-generation factories and humanoid robotic systems worldwide.
The age of AI is here.”
I’ll just add my piece for the valuation.
This model assumes a 30% & 25% CAGR growth until FY26 & FY29 respectively, 53.5% net margins, 1% of return to shareholders & P/S & P/E at 7.5x & 35x respectively.
Nvidia is still trading in its range between $150 & $90, so I would continue to buy the low of the range any time it comes - hopefully many in the next months. It is one, if not the most transformative company at the moment, and it won’t stop any time soon with AI accelerating all around the world and competition being pushed away every time Nvidia sells a GPU & the world becomes more & more dependent.
It is a company to own.
Can you explain what inference means in this context..? Google translate doesn't make sense
Thanks for the review. Need to be patient for a bargained stock price to arrive.