The hype surrounding the "soon-to-be" popping of the AI bubble has resulted in a buying opportunity of a wonderful company at a wonderful price. Concerns of the AI bubble being "worse than the dotcomm bubble" is overstated. Yes, i hate AI as much as you do, but lets not let that cloud our judgement.
Unfortunately, AI will be a staple to come after the LLM craze busts. Whether the consumers want it or not, it is certain it will continue to be used in company technologies like it has over the past few decades (AI isn't anything new!) Most tech scalers have no choice but to choose nvidia for their chip needs. Why? They can't afford to because of competition, and the fact nvidia is years ahead of the pack in terms of raw compute power. Companies will pay these prices not because they want to, but because they are forced to.
Google's TPUs and the other mag 7 chips are only used for deploying ai models, not training them. GPUs are the most efficient in training AI models.
As for AMD, they are 1-2 years behind nvidia. If AMD catches up, nvidia just releases their next lineup that is 2x faster with their robust capex spend.
Nvidia also sells proprietary server racks, unlike amd's customers which are forced to use ethernet. Ethernet can reduce the efficiency of clusters and can lag out in times of high compute (which is 24/7 for these chips).
If we assume a modest 20% EPS growth going forward, that places the stock at around fair value. Currently, estimates are at 48% growth for the next 5 years, and the industry is projected to grow 20% until 2031.
Usually, the mag 7 have beaten expectations despite all odds, and if this isn't the case with nvidia, you have a 50%+ MOS at these price levels.
Even if nvidia's price collapses due to the ai bubble popping (which, it's price has been suppressed to reflect that), so long as the fundamentals remain intact, it is a strong buy. The price is irrelevant.
Wallstreet wants to keep this stock low and stagnating (like it has in the past 6 months) before the next leg up.
"But what about the law of large numbers, they can't sustain that growth forever!"
This is currently reflected in its price. Wallstreet already is pricing in an earnings deceleration. If earnings cool off and the stock crashes, this is a buying opportunity (you're still buying a growing company below fair value)
Theoretically, there's nothing stopping nvidia from growing its market cap. Market cap is an arbitrary number. Gold has a 30+ trillion market cap, does that mean there are 30 trillion physical dollars in gold? No.
"But the PE ratio is too high! It must be overvalued"
PE ratio to determine a stock's value is relative, it depends on how fast a company is growing.
"Nvidia will be the next cisco." This is what I've been hearing a lot lately, and it's a nonsense comparison for 2 reasons.
1) Cisco's products were very simple to produce, so their moat (monopoly) was easy to dismantle. To create the next best AI chip, you have to pour billions in capex and hire top talent. By the time a competitor does this, nvidia is a year ahead.
2) And, Nvidia's valuation is lower than Cisco's.