NVIDIA May Be Cooked
On Friday evening, we covered DeepSeek R1, the Chinese AI breakthrough rattling NVIDIA and broader markets.
I wrote the following:
The implications are massive. In the long run, this could threaten not just American AI software companies, but also NVIDIA. If cutting-edge models can be trained with 5% of the previous hardware requirements, what does that mean for GPU sales going forward?
When I wrote that Friday morning, I considered buying some puts on NVDA. But I didn’t pull the trigger. I thought we had more time.
Of course, today I woke up to NVIDIA crashing in pre-market trading. It was down up to 13% before markets opened, and as I write this at around 11:00 am, NVDA shares are now down 15%.
In full disclosure, I did buy a few NVDA puts just after markets opened today (even though they were about 3x more expensive than Friday).
I have a short position here, so please take what I say about NVIDIA with a pinch of salt. I am now biased.
NVDA Teeters
I bought the NVDA puts as a hedge because of how incredibly important NVIDIA has become to the U.S. economy.
NVIDIA’s market cap, still around $3 trillion as I write this, is about 10% of 2024 U.S. GDP.
So NVDA’s current valuation equates it to roughly 10% of the value of all American goods and services produced over 12 months. That is borderline ludicrous.
The company’s remarkable run has essentially carried markets higher over the past few years.
Wall Street currently expects NVIDIA earnings to more than double from $1.30/share in 2024 to $2.95/share in 2025. By 2028, analysts expect the company to produce $6.15 in earnings per share.
If Wall Street’s 2028 prediction comes true, at current prices NVIDIA would be trading at 20x earnings (a P/E of 20) in 3 years. That’s IF they can grow earnings by 4.7x over the next 36 months.
So the pricing is getting ridiculous. The old adage “priced for perfection” fits this stock well. Yes, NVIDIA is a great company with a special founder/CEO. But it appears they may finally hit a peak for this cycle.
And when these types of bubbles unravel, they tend to do so quickly. Hot stocks take the stairs up, and the elevator down.
If this whole thing about China’s AI breakthrough proves to be overblown, I’ll be thrilled my puts expire worthless. The risk to markets is that significant. If the threat proves to be real, at least I’ll have some downside protection.
Rickards’ Warning
Earlier this month, our own Jim Rickards warned about how fragile this market is in Positioned For a Historic Crash.
Here’s an excerpt:
“We are now positioned for an historic crash. The specific cause does not matter – it could be war, natural disaster, a bank or hedge fund collapse or another unexpected event. What matters is the super-fragility of the market when the trigger is pulled. This is why Warren Buffett has over $300 billion in cash and why central banks are buying gold. Prepare now. Don’t be the last one to know.”
The market may have found Jim’s “unexpected event”: China’s AI breakthroughs (see Friday’s article if you need a primer).
Currently the S&P 500 is only down about 2% from its all-time highs. So if you do want to get prepared for a potential crash, there’s still time.
But with NVIDIA down 15% as I write this, we could easily see a full-blown market panic develop. I sincerely hope it doesn’t happen, but I believe it’s time to prepare for a significant correction. As I mentioned a few weeks ago, cash is increasingly attractive.
Disclosure: Author owns NVIDIA put options.
Comments: