stock research / valuation / ai stocks / risk management
Is NVIDIA Undervalued? My Fundamental Analysis After the $440B Drop
NVIDIA lost nearly $440 billion of market cap in two days after the Jensen Huang affair news, and I owned NVIDIA calls, so I did a full fundamental analysis.
NVIDIA trades at about 21 times earnings, close to McDonald's at 22, but NVIDIA grew revenue 106% last quarter and McDonald's grew 1%. Microsoft, Google, Amazon, Meta and Oracle plan to spend about $844 billion on capex this year, and for every dollar they spend, NVIDIA has brought in 47 to 57 cents every quarter since 2023. Working backwards, if Big Tech spends 30% more next year, today's $168.11 only makes sense at 20 times earnings if NVIDIA's cut falls to 38 cents, lower than any quarter since 2023.
At 20 times earnings my 2027 base case is $223 a share (+33%), the bull case $300, and $420 at 28 times. If capex stays flat, the bear case is $124; in a full downcycle at 15 times earnings it's $54 (-68%). I also cover why cheaper AI means more AI (the cotton gin), and why I treat the affair as a headline risk: in my research on 2.27 million company-days of news, good and bad news showed no reliable difference in returns. This is not financial advice.
Transcript
0:00An ABG caused the world's largest company to lose nearly $440 billion of market cap in just two days.
0:09I owned NVIDIA calls before the drop so I did a full fundamental analysis to figure out what to do next.
0:17It's titled F*** That Spy NVIDIA is still massively undervalued And here's what I found.
0:26Number one NVIDIA isn't priced for the growth it's achieving it's priced like a burger chain.
0:33It trades at around 21 times earnings and McDonald's trades at 22.
0:40To put that into perspective McDonald's grew its revenue 1% last quarter and NVIDIA grew it 106%.
0:51Number two NVIDIA gets paid out of big tech's AI spending.
0:57Microsoft Google Amazon Meta and Oracle plan to spend around 840 billion dollars on capex this year and Google says its spending will increase significantly.
1:11For every dollar they spend NVIDIA brings in around 50 cents and it's been that way every single quarter since 2023.
1:22So I decided to work backwards from the stock price and if big tech spends 30% more next year today's price only makes sense if NVIDIA's cut drops to 38 cents on the dollar And it hasn't been this low in any quarter in the past three years.
1:42Number three I looked at what NVIDIA is actually worth on its earnings.
1:47At 20 times earnings if capex grows 30% my 2027 base case is about $223 per share.
1:58That's a 33% gain from today's stock price.
2:03If capex grows 50% then the bull case is even more outstanding at $300 per share And if the market pays 28 times earnings then it goes to $420 per share.
2:20Now let's look at the bear case.
2:23Let's say capex stays flat even though Big Tech confirmed that it won't Then NVIDIA might be worth $124 per share And if Big Tech spends even less if they cut its spending and if NVIDIA is only getting paid 15 times earnings then NVIDIA should be worth $54 per share a 68% drop from where it is right now.
2:53So the whole bet on NVIDIA is if you think Big Tech is going to continue to spend like how they are.
3:02Now I want to address the elephant in the room and that's the fact that tokens get cheaper every single day.
3:12Many people think that cheaper AI means that Big Tech is going to spend less.
3:19I highly disagree and I'm going to give a concrete example from 10th grade where I learned about Eli Whitney and the cotton gin At the time many economists thought that if cotton was cheaper to produce then farmers could just have more of it so the demand for slavery was going to decrease.
3:44But what happened was the exact opposite.
3:48As cotton became cheaper to produce the demand for slavery exploded beyond what even the best economists ever thought And that's exactly what's happening with AI.
4:02Cheaper AI does not mean companies are going to spend less it just means they're going to do more.
4:11And about Jensen Huang's affair I model it as a headline risk and in my research on 2.27 million days of company news good news and bad news shows no real difference in long term returns Now this situation is a little different.
4:32There could be dramatic consequences for Jensen Huang's affair including increased government regulations including forcing Jensen Huang to resign and decreased spending by the AI companies that rely on NVIDIA's GPUs.
4:49So this model accounts for this within the full paper.
4:54You can read it with the link in my bio and you can tell me exactly where I'm wrong.
5:01Follow for part 4 where I tell you exactly what I plan to do with my NVIDIA calls.
Join the conversation
Loading conversation…