TLDR: Positron AI, a company focused on American-made semiconductors and inference hardware, has successfully closed an oversubscribed Series A funding round, raising $51.6 million. This brings their total capital raised this year to over $75 million. The funding will be used to further deploy their first-generation product, Atlas, and accelerate the development and rollout of their second-generation products in 2026, addressing the growing demand for AI infrastructure and the challenges of high costs and GPU shortages.
RENO, Nev. – Positron AI, a leading innovator in American-made semiconductors and inference hardware, announced today the successful closure of an oversubscribed Series A funding round, securing $51.6 million. This latest investment boosts the company’s total capital raised this year to more than $75 million. The funding round saw significant participation from key investors including Valor Equity Partners, Atreides Management, and DFJ Growth. Additional support came from Flume Ventures, which notably includes tech luminary Scott McNealy, along with Resilience Reserve, 1517 Fund, and Unless.
This substantial capital injection is earmarked for the continued deployment of Positron AI’s first-generation product, Atlas, and to expedite the launch of its second-generation products, anticipated in 2026. The move comes as global technology firms are projected to invest over $320 billion in AI infrastructure in 2025, highlighting the urgent need for efficient and cost-effective solutions amidst escalating cost pressures, power consumption limits, and persistent shortages of NVIDIA GPUs.
Positron AI’s Atlas, a purpose-built alternative to general-purpose GPUs, is designed specifically to accelerate and serve generative AI applications. The company reports that Atlas delivers 3.5 times better performance-per-dollar and up to 66% lower power consumption compared to NVIDIA’s H100. Randy Glein, co-founder and managing partner at DFJ Growth, emphasized the market opportunity, stating, “The early benefits of AI are coming at a very high cost – it is expensive and energy-intensive to train AI models and to deliver curated results, or inference, to end users. Improving the cost and energy efficiency of AI inference is where the greatest market opportunity lies, and this is where Positron is focused.” He added, “By generating 3x more tokens per watt than existing GPUs, Positron multiplies the revenue potential of data centers. Positron’s innovative approach to AI inference chip and memory architecture removes existing bottlenecks on performance and democratizes access to the world’s information and knowledge.”
Atlas features a memory-optimized FPGA-based architecture that achieves an impressive 93% bandwidth utilization, significantly higher than the typical 10-30% seen in GPU-based systems. This architecture enables support for models with up to half a trillion parameters within a single 2-kilowatt server. Furthermore, Atlas is fully compatible with Hugging Face transformer models and provides inference requests through an OpenAI API compatible endpoint. The chips powering Atlas are fabricated in the U.S. and are already in production environments, facilitating LLM hosting, generative agents, and enterprise copilots with reduced latency and hardware overhead.
Also Read:
- E2B Secures $21 Million to Advance Cloud Infrastructure for AI Agents
- Salient Secures $60 Million to Advance AI-Driven Loan Servicing
Dylan Patel, founder and CEO of SemiAnalysis and an advisor and investor in Positron, highlighted the critical role of memory in scaling AI workloads. “Memory bandwidth and capacity are two of the key limiters for scaling AI inference workloads for next-generation models,” said Patel. He further noted, “Positron is taking a unique approach to the memory scaling problem, and with its next-generation chip, can deliver more than an order of magnitude greater high-speed memory capacity per chip than incumbent or upstart silicon providers.”
