Valor Equity Partners, known for backing Elon Musk’s companies, Atreides Management, Flume Ventures and Resilience Reserve were among the investors who participated in the round.
The funding will enable Positron to scale production of its energy-efficient AI chips. Product is already being shipped to data centres and neoclouds around the US.
The company, which is based in Reno, says that its chips, which are manufactured in Arizona, use less than a third of the power of Nvidia’s top-of-the-line H100 graphical processing units, while maintaining the same performance.
According to Positron, its Atlas systems are presently achieving 3.5x greater power efficiency than Nvidia H100 GPUs for inference.
Leveraging a memory-optimised architecture that hits >93% bandwidth utilisation (vs. 10–30% for GPUs), Positron’s FPGA-powered servers support trillion-parameter models while offering plug-and-play compatibility with Hugging Face and OpenAI APIs - these systems deliver 70% faster inference at 66% lower power consumption than H100/H200 setups, reducing data centre CapEx by as much as 50%.
While Positron’s chips are intended for inference and demand is higher for training chips, analysts have predicted the need for inference chips could surpass them as more AI applications are deployed.
Currently, Nvidia’s chips have a market share of roughly 80%, but rising costs and concern about depending on a single supplier have led customers such as Microsoft, Meta and OpenAI to explore alternatives, both externally and in-house.
