Positron AI raised $230 million in an oversubscribed Series B financing at a post-money valuation exceeding $1 billion, as investors continue to back alternative architectures for AI inference. The round was co-led by ARENA Private Wealth, Jump Trading, and Unless, with strategic participation from Qatar Investment Authority, Arm, and Helena. Existing investors including Valor Equity Partners, Atreides Management, DFJ Growth, Resilience Reserve, Flume Ventures, and 1517 also joined the round. The financing was announced at Web Summit Qatar.
The new capital supports Positron’s roadmap from its shipping Atlas inference systems to its next-generation custom silicon, Asimov. The company plans to tape out Asimov in late 2026, with production targeted for early 2027. Positron positions Asimov as a memory-first accelerator, designed to address inference workloads increasingly constrained by memory bandwidth and capacity rather than raw compute. The company says Asimov will support roughly 2 TB of memory per accelerator and up to 8 TB per Titan system, with rack-scale memory exceeding 100 TB.
Positron reports early commercial traction with Atlas, including deployment by Jump Trading, which later chose to co-lead the Series B. Atlas systems are marketed as air-cooled, production-ready inference platforms manufactured entirely in the U.S., a point Positron emphasizes for supply-chain predictability. The company expects strong revenue growth in 2026 as it expands deployments across cloud, trading, video, and other memory-intensive inference workloads.
- $230M Series B financing at a valuation above $1B
- Co-leads: ARENA Private Wealth, Jump Trading, Unless
- Strategic investors include Qatar Investment Authority and Arm
- Atlas inference systems shipping today; Asimov silicon targets tape-out in late 2026
- Memory-centric design aimed at long-context, video, and agentic AI inference workloads
“We’re grateful for this investor enthusiasm, which itself is a reflection of what the market is demanding,” said Mitesh Agrawal, CEO of Positron AI. “Energy availability has emerged as a key bottleneck for AI deployment, and our roadmap is built to address power and memory constraints head-on for real-world inference.”
🌐 Analysis
The Series B places Positron among a growing group of AI hardware startups targeting inference economics as hyperscalers and enterprises confront power and memory limits. Its memory-first positioning contrasts with GPU-centric roadmaps from incumbents such as Nvidia, while aligning with broader industry exploration of alternative inference architectures. Strategic participation from Arm and early validation from customers-turned-investors highlight continued diversification in the AI infrastructure ecosystem.
Positron AI, which is based in Reno, Nevada, was founded by CEO Mitesh Agrawal and CTO Thomas Sohmers. The company is developing a vertically integrated platform spanning custom silicon, system design, and software, targeting deployment of high-performance inference workloads in data centers with significantly lower power consumption and total cost of ownership. Positron’s current Atlas systems are shipping today, while its next-generation Asimov inference silicon is on the roadmap for tape-out in late 2026 and production in early 2027.
🌐 We’re tracking the latest developments in networking silicon. Follow our ongoing coverage at: https://convergedigest.com/category/semiconductors/
