AMD working to catch Nvidia with latest tech at Advancing AI 2025 event

AMD working to catch Nvidia with latest tech at Advancing AI 2025 event

AI-Generated Summary

AMD unveiled its MI400 AI chips at its Advancing AI event, directly challenging Nvidia. Set to ship next year, these chips are integrated into Helios Think server racks, offering hyperscale capabilities crucial for training large language models. AMD aims to undercut Nvidia on both price and power consumption, though specific costs were not disclosed. The company also claims superiority in AI inference performance. CEO Lisa Su highlighted partnerships with major AI players like Meta, Oracle, and OpenAI. Despite these advancements, AMD’s stock remains stagnant, lagging behind Nvidia’s growth. AMD’s open ecosystem, ROCm, contrasts with Nvidia’s CUDA, as both companies adapt to the industry’s shift towards interoperability. Analysts debate whether AMD’s aggressive strategy can disrupt Nvidia’s dominance, with some favoring AMD for its growth potential.

📜 Full Transcript

Partsinevelos joins us with the key readouts from AMD’s advancing AI event. Kristina. >> Well, AMD really just threw down the Gantlet against Nvidia today with its mi 400 AI chips, which are set to ship next year. These chips are stacked into massive server racks called Helios Think. Hundreds of processors that work as one giant brain. This matters because AI companies need exactly this kind of hyper scale setup to train their large language models. AMD is betting it can also undercut Nvidia in terms of price as well as power consumption, with executives promising, quote, aggressive pricing, although they didn’t share the actual numbers in terms of costs. The company also claims its new hardware beats Nvidia on offerings for AI inference, which is the part where a trained models actually answer questions and generate responses. CEO Lisa Su just spoke with our Jon Fortt just within the last hour or so about the major partners on board. Listen in. >> If you look today, you know, seven out of the top ten, you know, model builders and AI companies are using AMD products. You know, that includes meta, Oracle, OpenAI. Sam Altman was here with us today. Meta and Oracle were here. You know Tesla to name a few. >> Despite that, though, AMD’s share price stock really hasn’t gone anywhere in the last year. What, down almost 2% versus Nvidia up 5% year to date, suggesting investors really aren’t convinced so far that this David can take on the Goliath of AI chips. But with cloud providers and AI companies burning through billions of dollars on computing power, AMD’s timing couldn’t be better. The question is whether customers will be willing to risk switching from Nvidia’s proven ecosystem for potential savings from Nvidia AMD. I should say Melissa. >> Christine, Lisa Sue was also talking about an open system yielding the best, you know, innovation. And I guess that’s sort of she’s talking her book. She wants people to be out of the Nvidia ecosystem and be willing to use AMD chips and technology. >> Right. So it’s the seventh generation Rocm. That’s her software platform that they also announced the seventh generation. That’s an open ecosystem, which differs from Cuda, Nvidia’s closed system. But if you recall, perhaps, maybe just was it a month ago? Not even Nvidia is finally opening up their closed ecosystem specifically with the product called NVLink. So without getting too technical, Nvidia is realizing that the future is open. You need to be able to take different parts from different companies, even though both companies really just want to own the entire stack and have everybody just stuck on their products. But it seems like that’s not going to be the future people. Chips need to either work together or else you lose market share. >> All right, Christina. Thank you. Kristina. Partsinevelos. What do you make of AMD? >> Only because you asked me, would I rather Nvidia or AMD? Because I know that’s what you’re thinking, right? I didn’t think. That. Either. No. >> One answer the question. Anyway. >> AMD I would rather have AMD. I think AMD is the wolf on the hill versus Nvidia is the wolf at the top of the hill. I think Nvidia has more to lose. AMD has more to gain. >> So many times. >> You could have, right? So AMD was in a declining trend line since March of 24, convincingly broke out of that declining trend line in May this year. So I would rather if I’m putting new money to work, I’d rather buy AMD Nvidia way ahead. But AMD is poised to be the biggest opportunity for them. >> Well, I don’t want to take the other side of the birthday boy, but I’m going to anyway. So I’d rather. And the way I’m positioned is far more dollars in Nvidia than I was going to say no, but we were just talking about AMD. I do like Lisa Sue a lot, but the valuation differential is not very big, right? So that makes me far more interested in Nvidia. Over. >> Yeah. Well I mentioned this. So the gross margins in Nvidia, this has been something we’ve all been tracking. The company guided them down. And you know, since they did a couple quarters ago from the high 70s to low 70s, they said it’s going to pick up back at the end of the year. The stock really stalled out. I mean it’s really gone sideways. And you know if you look at AMD and you talk about how they’re going to compete with Nvidia, it’s going to be on price. AMD has already this built in kind of ramp as far as margins are concerned. Over year over year. They wanted 53% last year, 51.5% expected this year up to like 55. If they’re going to be competing on price, that’s not happening. So the stock

[ad_1]
[ad_2]