Nvidia is the enabler of the modern-day gold rush.
- Nvidia reported stellar results as the intensifying mania around generative AI is driving unprecedented demand for its products which has driven both the share price and valuation into the stratosphere.
- FQ1 24 revenues / Adj-EPS were $7.2bn / $1.09 way ahead of estimates at $6.53bn / $0.92 and the fun did not stop there.
- Nvidia is swimming in a tsunami of orders for its products which has meant that FQ2 guidance was also way ahead of expectations and the company also hinted that even more is to come in the second half of the fiscal year.
- FQ2 24 revenue / gross margins are expected to be $11.0bn (+/- 2%) / 69.3% (+/- 0.7%) trouncing expectations of $7.6bn / 65%.
- The company also stated on the call that it was ramping up sourcing of manufacturing capacity for the second half in a clear sign that it expects demand to go up again from the rapid recovery it is already enjoying.
- Nvidia shares responded to these results with a staggering 24.6% rally adding close to $200bn in market capitalisation with the company now nearing $1tn in market value.
- Nvidia is currently by far the market leader in the supply of silicon chips for AI training as it is more than 5 years ahead of its peers and as such has locked down the developer community.
- AI developers love using the CUDA platform which is only available for Nvidia silicon which is exactly how Nvidia’s business model works.
- Consequently, Nvidia is being driven by rapidly increasing demand for GPU accelerated compute for large language models (LLMs) and the fact that it has a lock on the developer community.
- Nvidia can lay claim to being the picks and shovels of the AI gold rush and its competitors are really struggling to make any dent in its dominance meaning that the short to medium-term outlook is excellent.
- However, if either of these two things come unstuck, then the Nvidia growth story will quickly fall apart.
- The good news is that in the short-term there is no sign of this but looking out into 2024 and 2025, there are some clouds on the horizon.
- RFM research has picked up on research which indicates that the current fine-tuning of LLMs is very inefficient and that very large savings can be made.
- This combined with the reality that these machines remain as stupid as ever despite having incredible data retention and regurgitation skills means that the hype cycle of super-intelligent AI will hit the buffers sooner or later.
- This could very quickly turn a boom into a bust through no fault of Nvidia’s.
- Furthermore, there is also scope for a shift of the control point in AI away from Nvidia further up the value chain meaning that it could lose its ability to control the market for AI training silicon.
- Finally, there is the valuation which I find to be as difficult to internalise as ever.
- It looks like Nvidia will earn something like $6.50 in EPS this fiscal year which puts on a 12-month forward PER of 58.5x which is much too high for me to get comfortable with.
- This is why I don’t own the shares, but I continue to think that while generative AI dominates public discourse, they will continue to rise and Nvidia will do very well.
- Instead, I own Palantir which I bought back in January on a DCF-based valuation which has also now caught some of the AI tailwind following its excellent Q1 23 results.
- Despite the fact that the shares have almost doubled so far this year, the valuation on a fundamental basis looks much more reasonable than Nvidia’s and I do not have similar concerns about the long-term other than the rampant dilution from employee share compensation.
