Upscale AI, a new player in high-performance AI networking, has launched with over $100 million in seed funding. The investment round was co-led by Mayfield and Maverick Silicon, with contributions from several other major investors, including Qualcomm Ventures and Stanford University. Incubated by Auradine, the company aims to accelerate AI workloads through open-standard networking solutions.
The company is led by experienced professionals from notable tech companies such as Palo Alto Networks and Google. Upscale AI focuses on offering open-standard technology to enhance flexibility in AI infrastructure. Their solutions are built on open-source initiatives like SONiC, Ultra Accelerator Link, and Ultra Ethernet, promoting a bring-your-own-compute approach that moves away from proprietary systems.
With AI’s rapid growth outpacing traditional networks, Upscale AI is seizing a market opportunity exceeding $20 billion. The company is developing silicon, systems, and software to support ultra-low latency networking, crucial for AI training and cloud-scale deployments.
Founded by successful entrepreneurs Barun Kar and Rajiv Khemani, Upscale AI is set to redefine AI-native infrastructure. Their product lineup includes ultra-low latency network fabrics to optimize xPU performance, a robust Unified Network Operating System for improved scalability, and standards-based rack platforms for comprehensive multi-vendor networking solutions.
Navin Chaddha, Managing Partner of Mayfield, highlights the transformative shift in networking driven by AI’s demands, expressing excitement over partnering with Upscale AI’s founders once more to build a pioneering company for the AI era.
