Positron develops purpose-built hardware for generative AI inference, including the Atlas inference appliance and Titan systems.

Products & media

Atlas

A production inference appliance that runs transformer models with up to 500 billion parameters using an OpenAI-compatible API.

Titan

An announced inference system planned for 2027, combining four or eight Asimov chips for large-model deployment.

Asimov

An announced AI inference accelerator planned for 2027, designed with up to 2.3 terabytes of memory per chip.

Cosigns

Dylan PatelFounder, CEO, and Chief Analyst at SemiAnalysis
“Positron is taking a unique approach to the memory scaling problem…”
Source

Team

2 people across company lists

Current Affiliates

1 person

List

Jobs

Funding

All fundraising announcements are based on public sources. Please contact Cosign if you believe there are any discrepancies, or claim the profile to edit.

Leads

Participating Firms

Angels