Latency Optimized AI Processor

Watch CEO Jay Hu Announce Trex AI Processor  @ Alchemist  Accelerator Demo Day  (May 16, 2019), which    

-- reduces AI inferencing latency by up to 100 times,

 -- reduces energy consumption by 6 times,

 -- increases throughput by up to 100 times, and

 -- works for many workload types

Why Low Latency Important for AI Processing?

3D View of AI Processors

Best Latency for Many AI Workloads

DinoPlusAI Trex performance estimated based on RTL simulation