WWDC26: Optimize custom machine learning operations with Metal tensors | Apple
Apple's TensorOps library now natively accelerates quantized ML kernels on Apple Silicon, leveraging the M5 chip's new neural accelerator.
“The neural accelerator is a new hardware block in M5, located directly in each shader core.”
At WWDC26, Apple detailed how developers can write optimized custom ML kernels using the Metal TensorOps library, which now natively supports quantized data types (down to 2-bit) and exploits the M5 chip's new per-shader-core neural accelerator. It matters as a developer-facing on-device inference play, but it is a niche engineering session rather than a broad AI industry signal.