Draft:Positron AI
From Wikipedia, the free encyclopedia
Positron AI is an American artificial intelligence hardware company that develops computing systems designed for machine-learning inference workloads. Founded in 2023 and headquartered in Reno, Nevada, the company develops accelerator hardware and software intended to run transformer-based models such as large language models in data-center environments.
 | Draft article not currently submitted for review.
This is a draft Articles for creation (AfC) submission. It is not currently pending review. While there are no deadlines, abandoned drafts may be deleted after six months. To edit the draft click on the "Edit" tab at the top of the window. To be accepted, a draft should:
It is strongly discouraged to write about either yourself or your business or employer. If you do so, you must declare it. Where to get help
How to improve a draft
You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed article. Improving your odds of a speedy review To improve your odds of a faster review, tag your draft with relevant WikiProject tags using the button below. This will let reviewers know a new draft has been submitted in their area of interest. For instance, if you wrote about a female astronomer, you would want to add the Biography, Astronomy, and Women scientists tags. Editor resources
Last edited by Jarvis635 (talk | contribs) 6 days ago. (Update) |
History
Positron AI was founded in 2023 by software engineer Thomas Sohmers and computer scientist Edward Kmett. Technology executive Mitesh Agrawal later joined the company as chief executive officer.
Early development work focused on running large language models on FPGA-based accelerator systems. These early systems were used to prototype inference-optimized architectures before the development of dedicated semiconductor designs. The company subsequently announced dedicated inference hardware platforms designed for deployment in data centers.
Technology
Positron AI develops hardware and software systems designed primarily for artificial intelligence inference rather than model training. The company’s architecture focuses on memory capacity and memory bandwidth, which are significant constraints when running transformer-based neural networks.
Its systems are designed to support large language models and other generative AI systems in production environments.
Atlas
Atlas is Positron AI’s first-generation inference accelerator system. The platform consists of accelerator cards integrated into server systems designed for standard data-center racks.
Atlas systems are designed to run transformer-based models and are optimized for inference workloads with an emphasis on memory bandwidth and power efficiency. The platform supports integration with commonly used machine-learning frameworks and inference APIs.
Asimov
Asimov is a custom application-specific integrated circuit (ASIC) architecture announced by Positron AI as a successor to its earlier FPGA-based platforms. The architecture is designed to support large-scale transformer models while increasing memory capacity and computational efficiency for inference workloads.
Systems based on the Asimov architecture are intended for deployment in large-scale data-center environments.
Software
Positron AI provides software designed to integrate its accelerator hardware with existing machine-learning ecosystems. The software stack is intended to support widely used model frameworks and inference interfaces, allowing transformer models to run on Positron hardware without extensive modification.
Funding
In 2025 the company raised $23.5 million in early funding and later secured a $51.6 million Series A round from venture investors. In 2026 the company announced a $230 million Series B financing that valued the company at approximately $1 billion.