General Compute Splits AI Inference Between GPUs and Positron Hardware

AI12 hours ago

General Compute is purchasing Atlas and Titan systems from Positron AI to scale its heterogeneous cloud infrastructure. The partnership enables the company to run the prefill stage of AI inference on GPUs and hand off the decode stage, or token-by-token generation, to Positron's memory-first systems. This hardware will go live in Q1 2027 and will be sold to customers through a single endpoint under one contract and set of SLAs.

The design targets memory speed bottlenecks by avoiding dependence on the high-bandwidth memory (HBM) that currently limits GPUs. General Compute believes the demand for fast inference for open-source models will exceed supply for the next 10 years, making this approach critical for the development of always-on AI agents and multi-agent systems.

Mapped to your portfolio

Possible scenarios

45%Hybrid Inference Normalizes+0.3%
25%Positron Breakout+0.2%
20%Nvidia Ecosystem Holds+0.6%

Each figure is the modeled move of the S&P 500 if the outcome occurs. For information only. Not investment advice.

Looking for more context or analysis?

Tilt chat provides reliable, tailored insights, alerts, and automations.

General Compute Splits AI Inference Between GPUs and Positron Hardware

In this story

General Compute Positron AI

Portfolio

SPY
State Street SPDR S&P 500 ETF Trust
100%

Portfolio

SPY
State Street SPDR S&P 500 ETF Trust
100%

Mapped to your portfolio

Possible scenarios

45%Hybrid Inference Normalizes+0.3%
25%Positron Breakout+0.2%
20%Nvidia Ecosystem Holds+0.6%

Each figure is the modeled move of the S&P 500 if the outcome occurs. For information only. Not investment advice.