No Priors · Friday, October 2, 2026
Fractile is focusing on achieving significantly higher bandwidth to memory in their AI chips, moving away from traditional SRAM-based designs. This shift is driven by the need to support increasingly long context lengths in AI models and improve efficiency, especially for tasks like running large Mixture of Experts (MOE) models.
“And I think one of the things that we started to worry about towards the end of 2023 and certainly in 2024 was the scalability of this approach. And I think there are two things that grow with AI today. One is obviously the parameters of the model. But the other, and this was the one that kind of really got us nervous about that architectural approach, was this kind of growing context length that was becoming more and more part of the story for how we saw these models rolling out.”
“So for the past couple of years, we've been involved in these kind of skunkworks projects to move away from SRAM, looking at how we can get, for instance, much, much, much higher bandwidth to DRAM memories.”
“But one of the challenges, one of the headwinds to doing that is actually that it becomes incredibly prohibitive on today's HBM-based GPUs, XPUs, to serve those models efficiently. You end up often bandwidth bottlenecked.”