Positron raises $230M Series B to challenge Nvidia in AI inference chips

What’s the deal? Positron, a US-based AI chip startup, has raised $230M in a Series B round that values the company at more than $1B. The round was backed by investors including Arm Holdings and the Qatar Investment Authority, with co-leads Arena Private WealthDealroom has a profile for this one. Try Dealroom → and Jump TradingDealroom has a profile for this one. Try Dealroom →.

Based in Reno, Nevada, Positron is developing energy-efficient chips designed for AI inference, the process of running trained models. It is already selling a first-generation product and plans to complete a second, purpose-built chip later this year.

Why now? Demand for inference hardware is rising fast as AI models move from training into production. Buyers are increasingly looking for alternatives to Nvidia that can deliver better performance per watt and lower power costs.

Positron says interest accelerated after Jump Trading tested its first chip and pushed the company to raise capital. The timing also reflects a broader shift toward mixed hardware setups in AI data centres.

What could go wrong? Nvidia still dominates both training and inference, with deep customer lock-in and rapid product cycles. Matching that pace will be difficult for a younger company with fewer resources.

Hardware development also carries execution risk. Delays in delivering the next-generation chip or weaker-than-expected performance could slow adoption.

The signal: The funding highlights growing investor confidence in specialised AI chipmakers focused on inference rather than training. As energy use becomes a constraint, efficiency is turning into a competitive advantage.

It also suggests the AI hardware market is opening up. Large buyers are becoming more comfortable with heterogeneous architectures, creating room for challengers that can carve out focused use cases alongside Nvidia.

Sources:
TechCrunch
Bloomberg
Business Wire
eWeek
EE Times
Pitchdeck

J.V.