It seems like neither P/E ratio nor EV/EBITDA is a perfect measure. EV/EBITDA excludes capital expenses, but it also excludes amortization costs, which for AMD means the Xilinx acquisition.
It is very hard to have a discussion about under vs. overvalued, because it is all predicated on guesses of future performance if you are not stripping a company for its components in the short term.
To try to play the anti-nvidia side, they are in a very good position now, but at 65 P/E where do they grow beyond the current frothy expectations so new investors can see stock price appreciation? Especially with companies seeing the profits desiring to compete.
With AMD, you can see past performance with higher earnings and be optimistic that they can hit past earnings bringing the P/E ratio into alignment.
Nvidia is $3.42T market cap. AMD is $246B market cap. Nvidia makes up over 5% of the ETF I use. AMD is less than 0.5%
There's a lot more wealth tied up in Nvidia, so there is more interest and worry. They are also more associated with hyped up tech like Bitcoin and AI. Whether you believe in those technologies in the long term determines whether you're likely to think Nvidia can maintain.
With everyone dumping on Intel lately, AMD also looks like the primary CPU manufacturer.
It really is annoying me how much Nvidia is now tied up into almost every ETF, which means that it’s becoming “too big to fail” but also will drag things down horribly if it doesn’t keep delivering the huge earnings growth or suffers competition or AI doesn’t really deliver the goods in a meaningful way.
Right! AMD could double its share of desktop / server CPUs without a huge change in the industry but its much harder for NVIDIA to double its revenue of GPU cards, AI chips.
I have no idea what will happen, or what either is worth. But like Apple, at some point the market is the limiting factor, not the company.
Trailing P/E is about as informative as the tea leaves in your cup. You should ignore it.
Looking at Morgan Stanley's estimates for NVDA, their base case values them at $150.00 12 months from now, which is ~42x CY25 EPS of $3.57. According to the same report, the mean consensus estimate is ~$137.88 which if you assume the same EPS, equates to ~38.6x (this is from an Aug 29, 2024 report, which came out after NVDA's latest earnings, as usual)
For AMD, in the report that came out yesterday, the analysts values them at $169.00 which is ~43x FY2025 EPS of $3.94, below consensus estimates of $188.77 which imply a ~47.9x forward P/E ratio
Now this comes with a bunch of caveats like the different dates, the fact that each analyst has their own view on forward EPS, the fact that I picked just a couple of reports instead of looking at all analysts, the different periods from AMD's FY 2025 vs. NVDA's CY 2025... but two things appear to remain true: (1) trailing PE multiples are useless and the sooner we stop mentioning them the better and (2) the discrepancy in multiples you pointed out doesn't hold up to scrutiny
Diligencing the above caveats is left as an exercise to the reader
But it does make the opposition's point: while many will never acknowledge trouble in hindsight,it's much less common to acknowledge trouble in foresight ;)
Re: trouble in foresight, analysts and practitioners try to account for that by creating scenarios. Note I cited the base case for the Morgan Stanley reports, but those also include a bear case and a bull case. Naturally those still depend on assumptions...
Trailing P/E is not an accounting fact, just the EPS bit is an accounting fact. The numerator and the denominator relate to totally different points in time and should not ever be used in unison.
What can go up so quickly can also come down quickly.
It's impressive that Nvidia has been delivering explosive growth in their earnings to match their valuation, but it is valid to be concerned whether the world can keep investing in GPUs like this sustainably.
Yeah, that was Xilinx. Bought with stock. I remember thinking the rise from $30 to $140 or so was way too fast so using the gain to acquire Xilinx looked great. It turned a huge gain on paper into a new division of the company with a track record of steadily printing money. It was totally lost on me that it would come at the price of literally years of people looking at P/E and declaring it too expensive.
Company just put up ~7B in a quarter at 50% gross margin. Market cap is 240B. So about 9x a year's revenue, for a fabless semiconductor company. Looks crazy cheap to me, '132' P/E ratio and all.
It guarantees that Nvidia's margins will be shifting downward by painting the biggest target anyone has painted on his own back since the US in Desert Storm.
Nvidia P/E ratio is ~65
Yet I constantly hear how Nvidia is overvalued.