Tech News
Microsoft unveils Maia 300 inference chip featuring 2nm chiplets and 288GB HBM4 memory
The new accelerator uses a multichiplet design and HBM4 memory to boost token processing efficiency by 40 percent compared to leading GPUs.
Maia 300
Everything Patch Gazette has reported about Maia 300: 2 stories, newest first. Part of our Tech News coverage.
Tech News
Microsoft unveils Maia 300 chip, targets AI inference dominance by 2027
The tech giant’s third-gen custom silicon aims to reduce reliance on Nvidia, promising higher token output per watt for cloud AI workloads.