FuriosaAI

FuriosaAI builds high-performance, high-efficiency AI compute for the Inference Era, helping enterprises generate more AI tokens within the power and infrastructure they already have. Founded in 2017 by veteran engineers from AMD, Qualcomm, and Samsung, Furiosa has developed two generations of AI accelerators based on its proprietary Tensor Contraction Processor (TCP) architecture. Its flagship RNGD accelerator is now in mass production, delivering high-performance inference for advanced LLMs and agentic AI while fitting seamlessly into standard, air-cooled data centers. With more tokens per watt, per rack, and per dollar of infrastructure, FuriosaAI is helping enterprises scale AI without waiting for massive new power and data center capacity.
Visit website (opens in a new tab)
Loading

Contact Exhibitor