I want to open with a confession after last AMD analysis article. I sat down to write a clean, decisive “here’s the winner” piece for Bestock readers about Nvidia versus AMD, and about forty minutes in I found myself talking out loud to an empty room saying “okay but WHY is the P/E that different” like the spreadsheet owed me an explanation. It doesn’t. Spreadsheets never apologize. Let’s do this properly.
Both companies just posted numbers that would’ve been unthinkable five years ago, and both stocks are now priced like the market has already decided who wins the AI hardware war, except the market apparently decided two completely different things depending on which day you check.
The Headline Numbers, Which Are Genuinely Absurd
Nvidia’s fiscal Q1 2027 results, reported in May, showed revenue of $81.615 billion, up 85% year over year. Data Center revenue alone hit $75.2 billion, up 92%. Management guided the next quarter to $91.0 billion, and here’s the part that made me actually sit back in my chair, that guidance assumes zero Data Center compute revenue from China. Zero. As in, if China policy ever loosens up, that $91 billion number has room to go higher, not lower. Net income expanded 211% year over year. I had to read that percentage three times because my brain kept autocorrecting it to something more reasonable.
AMD’s Q1 2026 revenue came in at $10.253 billion, up 38% year over year, with Data Center revenue of $5.775 billion, up 57%. Client and Gaming hit $3.6 billion, up 23%, driven by Ryzen AI processor demand. These are genuinely strong numbers for basically any other company on earth. Next to Nvidia’s quarter, they read like a very talented little sibling who keeps getting compared to the sibling who apparently invented fire.
The Valuation Situation Is Where My Sanity Left The Building
Here’s the number that broke me. Depending on exactly which week you check, AMD trades somewhere between 58 and 68 times forward earnings. Nvidia, the company growing revenue faster, growing Data Center revenue faster, and posting the higher margins, trades at somewhere between 24 and 31 times forward earnings.
Let that sit for a second, because I stared at it for way longer than a professional adult probably should. The company growing 85% trades cheaper than the company growing 38-57%. I checked this twice because I assumed I’d swapped the tickers by accident. I had not.
| Metric | Nvidia (NVDA) | AMD (AMD) |
|---|---|---|
| Latest Quarterly Revenue | $81.6B (Q1 FY2027) | $10.25B (Q1 2026) |
| YoY Revenue Growth | 85% | 38% |
| Data Center Revenue | $75.2B | $5.78B |
| Data Center YoY Growth | 92% | 57% |
| Forward P/E | ~24 to 31x | ~58 to 68x |
| Dividend | Raised to $0.25/share (May 2026) | None |
| Recent Buyback Authorization | $80B additional (on top of $38.5B remaining) | $1.316B total FY2025 |
| Next Earnings Date | Aug. 26, 2026 | Reported Aug. 4, 2026 |
I want to point at that dividend and buyback row specifically because it changed how I think about this comparison entirely. Nvidia just raised its quarterly dividend from a symbolic one cent to twenty-five cents a share, declared May 18 and paid out in June, and tacked on an $80 billion buyback authorization on top of what was already left over from the prior one. This is a company returning roughly $20 billion to shareholders in a single quarter while still growing Data Center revenue 92%. That’s not really a “growth stock” behavior pattern anymore, that’s a company that’s figured out how to be a cash machine and a rocket ship at the same time, which I did not think was allowed.
AMD, meanwhile, pays no dividend and bought back a comparatively modest $1.316 billion worth of stock across all of FY2025. Not a criticism exactly, plenty of growth companies reinvest instead of returning cash, but it does mean you’re not getting paid to wait around while the MI400 story plays out.
The Actual Bull Case For AMD, Because I’m Not Trying To Just Dunk On It For 2000 Words
Okay, deep breath, let me be fair here, because AMD’s story genuinely has legs and I don’t want this to read like I’m just here to make fun of a company that’s up something like 267% over the trailing twelve months, which, again, is a completely bananas number for basically any stock that isn’t Nvidia.
The MI400 series, specifically the flagship MI455X, delivers 40 PFLOPS of FP4 compute with 432 GB of HBM4 memory, and the Helios rack systems built around it start shipping in Q3 2026. AMD has landed two genuinely massive hyperscaler commitments here. OpenAI committed 6 gigawatts. Meta committed up to 6 gigawatts as well, worth roughly $60 billion across a multi-year deployment. Those aren’t rumors or hopeful analyst extrapolations, those are named, dollar-figured commitments from two of the biggest AI compute buyers on the planet.
The bull case, distilled, is this: AMD’s base is small enough that these two deals alone could reshape its Data Center revenue mix dramatically over the next two years, and the market is pricing in exactly that kind of transformation. A reasonable framework I came across while researching this, and one I actually like, suggests holding AMD alongside Nvidia rather than instead of it, something like 50% Nvidia, 25% AMD, and the remaining quarter split across other AI infrastructure names like Broadcom and TSM. That captures the established leader and the up-and-coming challenger without doubling down on the exact same bet twice.
The Actual Bear Case For Both, Because Nobody Warns You About This Part
Here’s what nobody puts in the headline. Both companies face the exact same China export headwind, and it’s not a small asterisk, it’s baked directly into Nvidia’s own forward guidance now. When your $91 billion quarterly guide already assumes zero China Data Center revenue, you’ve essentially already priced in the worst case on that specific front, which somewhat ironically makes Nvidia’s number look more conservative and more trustworthy than it would if China were still fully in the mix.
For AMD, the bear case is more about execution risk than policy risk. The MI450 deployment timeline needs to hold. If Helios shipments slip, or if hyperscaler capex flattens out even slightly, a 58 to 68 times forward earnings multiple has essentially no room for disappointment. I don’t say this to be dramatic, I say it because I’ve watched enough high-multiple growth stocks get repriced overnight on a single soft data point to know that valuation compression is not a hypothetical risk here, it’s the single biggest variable in whether AMD’s current price makes sense a year from now.
My Actual, Personal, Slightly Exhausted Take
If you’re asking me to pick one, and I know that’s ultimately why you’re reading a Bestock article with this title, here’s my honest answer after wrestling with this spreadsheet for far longer than my Friday evening plans should have allowed. Nvidia is not the “safe boring pick” the way legacy blue chips used to be. It’s growing faster than AMD on every metric that actually matters, trading at a meaningfully lower multiple, paying you a dividend now, and buying back stock aggressively, all while guiding conservatively enough that China policy can only be a positive surprise from here. That is, frankly, a strange and wonderful combination to find in a single stock, and it’s the main reason my own money leans heavier there.
AMD earns its place too, but as the higher-beta, higher-conviction-required piece of the puzzle. The Meta and OpenAI commitments are real and they’re big, but you’re paying a genuinely steep premium for execution that hasn’t fully shown up in the numbers yet. I don’t think that makes AMD a bad stock. I think it makes AMD a stock where you need to actually believe in the MI400 shipping story, not just the AI theme broadly, because the valuation has already priced in quite a lot of that belief for you.
Both companies report again soon, AMD already reported August 4, Nvidia’s up next on August 26, and I’ll be honest, I’m going to be refreshing that page the same way I refresh everything else in this industry lately, half excited and half bracing for my spreadsheet to yell back at me again.