Modular
Inference reimagined, from Kernel to Cloud. One unified stack.
Modular is a unified AI inference and compute platform for serving, optimizing, and deploying generative AI models across heterogeneous hardware. The platform includes Modular Cloud, the MAX serving/framework stack, and Mojo kernels to run models across GPUs, CPUs, NPUs, and ASIC-oriented environments without hardware-specific rewrites.

Recent stories
0 linked stories
No linked stories yet.