Nvidia FQ3 24 – As good as it gets

Almost perfect with a valuation to match.

• Nvidia reported another stellar set of results, but it will need to increase its bookings at TSMC in 2024 if it wants to continue growing and I am far from certain that TSMC has the space at these nodes to accommodate the demand.
• FQ3 2024 revenues / EPS were $18.1bn / $4.02 well ahead of the average estimate of $16.2bn / $3.37.
• Revenues were up more than 200% YoY driven by the data centre which grew revenues to $14.5bn up 279% YoY.
• This comes as no surprise as Nvidia is the engine of the generative AI craze and the phones are ringing off the hook with customers all wanting to purchase its chips.
• This is how gross margins were able to expand to 74% as Nvidia can pretty much charge what it wants for its chips and if customers balk at the price there are others in line who will pay up.
• The net result of this unbridled demand is that the problems Nvidia is having in China (around 20% of revenues) are little more than a small blip.
• Nvidia’s A800 and H800 processors can no longer be exported to China following the tightening of restrictions in October (see here) but Nvidia appears to have had little trouble selling chips elsewhere.
• This has caused a small problem as FQ4 24 revenue guidance was a little less than the highest of expectations at $19.6bn – $20.4bn ($20bn) but generally, things are going extremely well.
• This shift will have involved capacity at TSMC being switched from H800 and A800 products to other product lines, but Nvidia has so much demand at the moment, that this has not been very difficult.
• Revenues to China will fall heavily this quarter but I expect that when the new products are ready, they will quickly bounce back as China resumes buying everything that it can get its hands on before the goalposts are moved again.
• The net result is that while the generative AI craze rolls on, Nvidia is in an extremely strong position.
• Despite a growing list of competitors, Nvidia remains the place to train one’s generative AI algorithms largely as a result of the CUDA platform that Nvidia has spent almost 20 years building.
• Everyone knows how to use it and likes it combined with its breadth of tools means that nothing can hold a candle to it yet.
• Hence, the biggest competition is coming from OpenAI which is trying to establish an ecosystem (see here) which may shift the control point in AI training up through the technology stack.
• However, while OpenAI is imploding, all of this is on hold and if everyone decamps to Microsoft, the whole effort will have to be recreated from scratch.
• As this is the only credible threat to Nvidia at the moment, the soap opera is working in Nvidia’s favour.
• Nvidia continues to go from strength to strength, but the issue remains the valuation.
• Nvidia currently has an EV of $1.2tn meaning that it is trading on 22x 2024 EV/Sales and 45x 2024 PER which is a very large premium to all of its peers.
• This valuation requires that growth continues at a very healthy clip (probably will in the short term) and that the generative AI bubble does not burst.
• This is where I have concerns as the capabilities of generative AI have been hyped beyond that of which they are capable and so there is disappointment coming.
• This is exactly what happened to autonomous driving and there is no reason to think that it won’t happen to generative AI.
• I would not want to be holding the shares when the bubble pops no matter how much I like the company.