Alibaba unveils Qwen-branded AI smart glasses at MWC Barcelona with pre-orders starting March 2, challenging Meta's dominance in a wearable AI market that tripled last year.
Groq's LPU chips deliver inference speeds that make GPUs look slow - 1,200+ tokens per second on Llama 4. We benchmark latency, throughput, model availability, and pricing against the GPU-based competition.