HeadsUpAI

Modular Open-Sources GPU Kernels for Multivendor Consumer Hardware

Mojo

· Updated

Modular, the company behind the Mojo programming language and MAX AI acceleration platform, is open-sourcing its AI models and GPU kernels. The kernels will run on multivendor consumer hardware, enabling the same code to work across GPU vendors without platform-specific rewrites.

GPU kernel work has typically required vendor-specific rewrites, making hardware portability expensive. Modular is opening that layer and explicitly welcoming community contributions that outperform its own kernels — a bet on open development over proprietary optimization.

Developers building AI workloads can now run Modular's open-source GPU kernels on NVIDIA, AMD, Intel, ARM, or Apple Silicon. Contributing kernel improvements back to the stack is also on the table — the company says it welcomes work that beats its own.

View the full update on modular.com
Chris Lattner
Chris Lattner
@clattner_llvm
X

@Zyyon_ Please don’t tell anyone: we aren’t just open sourcing all the models. We are doing the unspeakable: open sourcing all the gpu kernels too. Making them run on multivendor consumer hardware, and opening the door to folks who can beat our work. Plz keep it quiet, ok? 😉

28retweets
View on X

Share this update