4 ms·
My unpopular opinion (well to people heavily invested in AMD anyway) is that their valuation is bordering on the delusional based on a conjecture that they will
by belval 2y ago
My unpopular opinion (well to people heavily invested in AMD anyway) is that their valuation is bordering on the delusional based on a conjecture that they will somehow completely kill off Intel in the datacenter and eat some of Nvidia's lunch for GPU-based workloads.
They are seriously hurting Intel in the datacenter space, but Intel is not dead yet and their accelerator have so far seen very small adoption. They are fabless meaning that their margins are capped, x86 mobile CPUs are under attack from ARM and all hyperscalers are working on their own accelerator with some degree of success. They won't move to AMD GPUs.
They are a good company putting out great hardware, but at the current price I just don't get it.
- qzw 2y agoI think it’s because AMD has been the “if only…” company for decades. It used to be in relation to Intel’s x86 dominance, and now it’s relative to Nvidia. The x86 fairy tale did eventually come true, which made a lot of money for investors who held on long enough or got in early in the Ryzen era. So now a lot of people are hoping the same will happen in GPUs, and that accounts for the valuation/speculation. And it does make a certain amount of sense. When you have basically a market duopoly but instead of two nearly equal sized competitors you have a 90/10 split, then there’s always a chance the smaller firm will somehow gain significant market share at some point. Waiting for that pop and selling high is a gamble a lot of investors are willing to make.
- JonChesterfield 2y agoOne could rephrase your thesis slightly. Conjecture that Intel will continue killing themselves and that some of the companies spending billions on AI that doesn't care what hardware it's running on will continue to choose the much cheaper hardware. Fabless means the capex is tiny. The hyperscalars have been trying to follow Google's TPU approach for a decade now. Whole load of silicon startups came to life, set fire to a lot of VC money, died off in the same time period. I believe AMD is the most likely winner of the silicon age. Progress is slower than one would like but everything is heading in the right direction.
- hakfoo 2y agoI think the case for AMD over nVidia is that it's somewhat less bubble-driven. A lot of the crypto boom is over. If the AI boom fizzles, and nVidia goes back to selling GPUs for gamers, an admittedly niche market, that's going to be a hell of a haircut. If AMD has to retreat to that, well, they've lost far less, and they have a successful CPU product line to lean on.