AMD’s first-quarter results sailed past Wall Street’s estimates.
The company’s Q2 outlook also outpaced expectations.
The results show that AI adoption continues to ramp higher.
The company’s gross profit margin expanded by 300 basis points year over year to 53%, while operating expenses grew more slowly than revenue, driving higher profits to the bottom line. On a GAAP basis, net income surged 95% to $1.38 billion, driven higher by its operating margin that expanded 300 basis points to 14%.
AMD chair and CEO Dr. Lisa Su called it an “outstanding quarter,” noting that the results were driven by “accelerating demand for AI infrastructure, with data center now the primary driver of our revenue and earnings growth.” The chief executive went on to say, “We are seeing strong momentum as inferencing and agentic AI drive increasing demand for high-performance CPUs and accelerators.”
Indeed, the data center segment results tell the tale, as revenue jumped 57% year over year to $5.8 billion. AMD saw strong demand for its Epyc CPUs and Instinct GPUs – which are seeing strong data center demand.
Foolish me, I sold a non-trivial chunk yesterday (maybe 20%) figuring there would be a post-earnings plunge and also thinking to chase other opportunities.
My wife is so risk averse as to verge on paranoia. I got out of AMD when the gains seemed to guarantee our happy retirement, and went more into dividend stocks. Nothing against AMD, but Taiwan vs. China IMO means they are still a risky stock.
Four years later, I could have NINE times my guaranteed money. That would put us from our inherited townhouse in a decent community back into a private house where my wife controls all the gardening and we don’t need to deal with the architecture committee to install solar panels. And have an extra $2M in the bank. But we’d have lost four years of secure living.
But I invested in AMD last century. My gains were enormous. For others with enormous gains, it should be worth considering that the macro environment is scary as hell. The more major powers get involved in wars, the likelier that China decides nobody will stop them from retaking Taiwan (and thus TSMC and destroy AMD’s ability to make chips for a while, maybe permanently). TSMC is doing its best to get off Taiwan. We have to start reading tea leaves to know what preparations have been made, but it would not surprise me if TSMC has all its foundries rigged to detonate, and leaked that quietly to China to deter an invasion hoping to use TSMC to kickstart China’s own chip manufacturing.
I still don’t know where else to put my money. By the standards of retired people, we are in the top 10% in income in America. My best investments were always work related: computers, electronics, and software, especially computer games. But for other old timers like Antonio, there can be great comfort in taking ones ten bagger and getting out while it’s still a ten bagger. That might not be so for those who got into AMD more recently, but the majority of my basis was under 10. But if things get really bad, your double might drop to a fraction.
As long as I am predicting doom, without TSMC, Intel again becomes an interesting proposition. There will still be demand for silicon, and they do have fabs.
I’ve just taken out half of my AMD (at 460) but left the rest in there. Why not all? dunno! it just seems weird to too much “Real money” hanging around. We’re going to be retiring in about 5 years most likely.
We’ll have enough assets to retire “comfortably”, after selling our main house and rental, and moving to smaller home at the beach. But definately not splashing any money around to our kids with our situaton.
The money from the sale hasn’t even come into my account yet, but thinking I’ll put it in some term investments.
Hey RP, It has been a while! I continue to enjoy my retirement and spending much more time with my grandkids, which certainly saves on the expensive vacations. Medicare dropped our health care expenses dramatically. Upcoming SS at age 70 will improve cash flow.
Luckily I stayed well above the recommended percentage to equities over the past decade, but I am finally moving toward a MUCH more conservative allocation. Bonds are still pretty stinky, but I went anyway. Currently at about 60% equities, 30% bonds, and cash heavy at 10%. I’m not spending any time picking stocks anymore.
I too followed my own investing path. Our grandchildren live in Sweden, so vacations are expensive, but our expenses are low and our savings have grown every year since I retired even while we’re spending whatever we like. I am about 20% cash, but plan to start withdrawing from 401(k) this year before I am forced to start in three years. I have a new kidney, our daughter’s family are coming in in July, and then there’s the World Science Fiction convention in August, so I’m staying busy. My wife’s a little older than we are so she has a pension from back when they still gave them out to non-executives, which supplements our two socials well.
My Portfolio has never been anywhere close to being as high in value as of today June 2nd.
I have very significant unrealised gains (ranging between 200%- 620%) in AMD, ARM, ANET, CRWD and some others.
With so much uncertainty in the geo-political space (Middle East), I feel that a big readjustment is almost certain to happen. So, should I sell some of my portfolio now, bank the cash, wait for the downturn and then buy back in?
I am 73 yrs young and consider my investment window to be +/- 5 years.
I suspect you are in the same boat as many of us. Every year I increase my percent to conservative investments (like CD’s and Bonds). This results in a “double whammy” on the need to sell equities: they are significantly outperforming, and I want to reduce exposure to them. The issue over the past few years is selling the equities can result in a very high tax burden so my goal has been to trim equities, but attempt to minimize the realized gains. Rather than making any drastic moves I just make one or two moves every year.
I am not sure this is very helpful, but it is my approach.
Very interesting. I presume it will run the Windows ARM OS like Qualcomm Snapdragon Elite laptops. The NVIDIA site mentions “up to” 1 petaflops (FP4) but doesn’t mention the usual TOPS (trillions of OPS) spec. A peta is a 1000 trillion, so does this mean the Spark can do 1000 TOPS? Are FP4 ops the standard TOPS measure? If so, the best TOPS specs I’ve seen max out at 150 or so, so a true 1000 TOPS would be a 7X speed up!
Actually it didn’t have much information like I would have expected from a Toms article. It seemed to me to be an introduction article but my thought was that AMD would also have a gpu/video card coming soon as well…doc
It is not just a GPU, but rather a complete PC SOC. It is mostly a competitor to the QUALCOMM ARM based PC ecosystem. Real x86 compatibility issues coupled with perception have held ARM back as a niche player. I think NVIDIA has the software chops to create a serious threat to the x86 duopoly. It appears there is a belief that somehow AI accelerates the transition away from X86… we will see.
In your position I’d feel compelled to lighten up on AMD and other hyperbolic stocks… but then again, man it’s hard. I’m turning 61 momentarily and while I’m not retired I’m not exactly working gainfully either (need to fix that!) and while the last 2 years of hiatus have left me actually up on paper, I am pushing my luck.
It’s more aimed at the mobile AI workstation segment, if I had to guess, and built on an assumption that local heavyweight AI processing is going to become a must-have for high-end business use cases. It’s got A LOT of horsepower in ways that completely separate it from today’s so-called “AI PCs”.
I’m running an AMD Strix Halo workstation with unified memory and 128GB of RAM but it’s barely sufficient for any powerful local model work because of limited memory bandwidth. NVidia’s machine will smack around the current Strix Halo generation and I fear the successor as well. And it runs real CUDA.
I’m not really tempted because I’ve determined that for most of what I’m doing I really will need true frontier model level smarts for a while (I use Claude a lot and I can really, really tell the difference for some of what I’m working on when I use Opus vs. Sonnet, and anything less than Sonnet just falls on its face). I can see how this machine might start to close a gap, though.