All eyes on Nvidia earnings as AI boom reaches historic proportions

Key points:

• Nvidia's Q4 earnings are highly anticipated: Analysts expect record-breaking revenue and EPS growth due to the strong AI demand.
  
  
• Uncertainty surrounds the report: While Nvidia usually beats estimates, geopolitical risks and potential overvaluation raise concerns.
  
  
• AI theme drives markets: The AI sector is experiencing explosive growth, reflected in Nvidia's success and a basket of related stocks.
  
  
• Bubble-like tendencies observed: Similar to past tech booms, there's a sense of FOMO and potentially inflated valuations in AI stocks.

Will Nvidia finally indicate the AI boom is cooling?

There is probably not an earnings release more important for equity market sentiment right now than Nvidia’s FY24 Q4 earnings (ending 31 January 2024) out Wednesday night after the US equity market close. With the AI boom being stronger than ever and spreading to more and more stocks all eyes will be fixated on this earnings release. Analysts expect revenue of $20.4bn up 238% y/y and EPS of $4.60 up 613% y/y in what is probably the most extraordinary increase in a mega cap business ever recorded. We have never seen a $1trn market value company with 12-month revenue of $45bn and a net profit margin of 55% grow at these blistering growth rates.

The chart below shows Nvidia’s quarterly revenue and analyst estimates projecting quarterly revenue to hit more than $25bn within the next year. This level of revenue is around four times the level in late 2022 when Nvidia was hit by a collapsing demand from Bitcoin mining due to falling Bitcoin prices and generally lower technology spending as Silicon Valley was adjusting to higher interest rates.

If we take a look at the geographical revenue split we can see that it is really the US segment that has taken off in the previous two quarters followed by strong growth in Chinese and Taiwanese sales. Singapore was once again separately split out in the previous quarter accounting for 15% of revenue. This is naturally odd and it is clear Singapore is used as a legitimate hub for re-exporting of Nvidia chips to countries were export controls may apply. As we have highlighted in the past a significant amount of Nvidia revenue comes from China and Singapore (37% of revenue) which given the ongoing trade frictions and potential Trump election win later this year could become a key geopolitical risk for Nvidia.

Will Nvidia beat estimates and how is the market positioned? Nvidia has beat on revenue and EPS in the past 7 of 8 earnings releases and the recent signals from Arm and other companies in the AI ecosystem are not suggesting growth is slowing down in the short term. We are leaning towards the view that analysts will once again underestimate Nvidia’s growth, but uncertainty is 2.5x larger than normal around this earnings release. The normal average 1-day move around Nvidia earnings is 4.3% but equity options are pricing 10.5% this time. To get a sense of the explosive growth in cumulative AI performance we are highlighting a slide from Nvidia’s recent investor slide deck (see below).

What about Nvidia’s valuation. Can the outlook justify it? As we pointed out in a recent equity note on Arm shares, both Arm and Nvidia are valued significantly above the market, which is not strange given the underlying growth in AI chips. But is it too much? It depends of course on the underlying assumptions for the future. NYU professor Aswath Damodaran, one of best on equity valuations, recently wrote about the magnificent seven and said this about Nvidia:

“…Nvidia and Tesla, where questions remain about what the end game will look like, in terms of market share. Historically, neither the chip nor car businesses have been winner-take-all businesses, but investors are clearly pricing in the possibility that the changing economics of AI chips and electric cars could alter these businesses.”

Later in the article, Damodaran says explicitly that Nvidia looks overvalued despite factoring significant AI growth into his valuation models before serving a contradiction as he owns shares in Nvidia.