Flume Ventures
Assembly Line
Positron Secures $23.5M to Design And Manufacture Energy-Efficient, Made-In-America AI Chips
Positron, a rising force in AI inference technology, has secured $23.5 million in funding to scale production of its energy-efficient, made-in-America AI chips. The company’s Atlas systems achieve 3.5x better performance per dollar and 3.5x greater power efficiency than Nvidia H100 GPUs for inference, leveraging a memory-optimized architecture that hits >93% bandwidth utilization. Positron’s FPGA-powered servers support trillion-parameter models and offer plug-and-play compatibility with Hugging Face and OpenAI APIs, delivering 70% faster inference at 66% lower power consumption than H100/H200 setups. With its domestic manufacturing capabilities, Positron is reducing reliance on foreign supply chains and providing a cost-effective alternative to Nvidia’s AI hardware. The company’s leadership, including newly appointed CEO Mitesh Agrawal, is driving the next generation of AI compute, with a mission to deliver domestically made AI compute solutions that outperform conventional GPUs in both cost and energy efficiency.