Making AMD GPUs competitive for LLM inference

Re:Making AMD GPUs competitive for LLM inference by SixtyHurtz on Wednesday December 25, 2024 @08:33AM (#42508703)

That's just BOM. When you factor in R&D they are clearly still losing money on B580. There's no way they can recoup R&D this generation with a 10% gross margin.

Still, that's to be expected considering this is still only the second generation of Arc. If they can break even on the next gen, that would be an accomplishment.

[ Parent | on HN ]

Re:Making AMD GPUs competitive for LLM inference by ryao on Wednesday December 25, 2024 @10:48AM (#42509335)

To be fair, the R&D is shared with Intel’s integrated graphics as they use the same IP blocks, so they really only need to recoup the R&D that was needed to turn that into a discrete GPU. If that was $50 million and they sell 2 million of these, they would probably recoup it. Even if they fail to recoup their R&D funds, they would be losing more money by not selling these at all, since no sales means 0 dollars of R&D would be recouped.

While this is not an ideal situation, it is a decent foundation on which to build the next generation, which should be able to improve profitability.

[ Parent | on HN ]