Nvidia FQ1 26 – China? What China?

Game still on.

• Another great set of results confirms that the AI boom is still on and that any weakness from Chinese restrictions can easily be resold to other customers.
• FQ1 26 revenues / adj-EPS were $44.1bn / $0.81 comfortably ahead of estimates of $43.3bn / $0.74 which did not come as much of a surprise.
• This is because Nvidia still has more demand than it knows what to do with, meaning that its revenues are pretty much a function of how much capacity it has booked at TSMC and the orders it has placed with other hardware suppliers.
• Nvidia will have good knowledge of these factors well ahead of time, meaning that its ability to predict its revenues up to 12 months away is the best it has ever been.
• This is why we are seeing performance just enough ahead of expectations to maintain the valuation of the shares, but not so much that irrational exuberance sets in.
• This is careful management of expectorations, which is precisely what Nvidia needs if it is to hold onto its now gold-plated reputation and not be dismissed as a flash in the pan.
• Currently, the market is giving Nvidia the benefit of the doubt when it comes to wobbles like the China FQ2 26 guidance, and this is crucial as many of its peers can barely put a foot wrong without getting walloped.
• FQ1 26 data centre revenues dominated the story once again, with $39.1bn in revenues up 10% QoQ and 73% YoY.
• Gaming was up 42% YoY, Professional Visualization 19% YoY and Automotive and Robotics 72% YoY, but together these made up just 11% of revenues.
• Guidance predicted lower growth as it contains a loss of $8bn in revenues from the H20 ban but still managed to just about meet expectations.
• FQ2 26 revenues / gross margins are expected to be $44.1bn – $45.9bn ($45bn) / 71.3% – 72.3%, only slightly adrift of revenue expectations of $45.9bn.
• If one includes the $8bn in lost revenues, then forecasts would have been handily beaten, and this is the benefit of the doubt that the market is affording Nvidia as the shares were up 4.8% in after-hours trading.
• This is why revenue growth in FQ2 26 will slow to 50% YoY, while adding back the $8bn would have given 76% YoY for FQ2 26.
• However, I think that the full $8bn is not going to hit the top line as Nvidia should be able to redirect some sales that were destined for China elsewhere.
• I continue to think that Nvidia would be best served by giving up on the China market in its entirety, as China will replace Nvidia as soon as it can, and the restrictions are now so tight that it is difficult for Nvidia to sell a useful product there.
• Furthermore, I continue to believe that Jensen is not thinking long-term enough about his China business (see here) as the $50bn market he thinks he is going to lose is very likely to be made up elsewhere.
• This is because the restrictions are having the effect of making China’s product uncompetitive in terms of cost, meaning that they are unlikely to fare very well in markets outside of China.
• We are already seeing this as the big economies in the Middle East are pivoting towards the USA and the West, and away from China with whom they have had a close relationship over the last 10 years.
• I think more countries will follow when they start to realise what the economic realities of Chinese AI powered with Chinese silicon really are without subsidisation from the CCP.
• The bad news is that the Chinese market is effectively lost, but the good news is that I think Nvidia will more than make up for this loss in other markets.
• AI is becoming a strategic imperative for many countries, and so on top of every company wanting access, all the governments do too.
• If they all choose Nvidia, then there is still a long and steady trajectory for revenue and the shares, although nothing like the rally we have seen in the last 18 months.
• This is why, of all the direct ways to play AI, Nvidia remains my favourite, although I still prefer the adjacent themes of nuclear power and inference at the edge.