By Tobias Mann
Publication Date: 2026-02-04 22:06:00
On paper, Positron’s next-gen Asimov accelerators, no doubt named for the beloved science fiction author, don’t look like much of a match for Nvidia’s Rubin GPUs.
Yet, the Arm-backed AI startup boasts its inference chip will churn out five times as many tokens per dollar while using one-fifth the power of Nvidia’s latest accelerators to do it.
Those are certainly some bold claims, which the company contends are possible because the chip was designed to support large-scale inference workloads. Another $230 million of fresh capital probably doesn’t hurt either.
Positron’s Asimov couldn’t be more different from the GPUs popularized by Nvidia and Arm.
Unlike its prior generation Atlas systems, which used high-bandwidth memory (HBM), the Asimov uses LPDDR5x memory, which can be expanded using Compute Express Link (CXL) from 864GB to 2.3TB per chip. Higher memory capacity means more room for LLM parameters and the key-value caches used to keep track of the model state.
But while LPDDR5x is both cheaper and higher capacity than HBM, it’s also glacially slow by comparison.
