Sparxity raises $90M Series A for sparse-native AI inference chips
Dresden-based Sparxity raised a $90 million Series A for silicon that computes only active model weights, which it says can speed inference up to 44x.
Dresden-based Sparxity has raised a $90 million Series A for silicon that computes only the model weights active at a given time. The company says this design can speed up AI inference by up to 44 times.
The core idea is sparsity. The company targets inference, the stage where a deployed model answers queries.
The chips are already deployed at the Jülich Supercomputing Centre, a German research facility. Sparxity also has a partnership with Deutsche Telekom's T-Systems.
Investors include ICM HPQC Fund, Lockheed Martin Ventures, GlobalFoundries, and EIC Fund.
What changed
Sparxity raised a $90 million Series A for sparse-native inference chips that compute only the active model weights.
Type: Business · Applies to: Sparxity
Why it matters
Investment in inference-focused hardware suggests that more alternatives to general-purpose GPUs may emerge for running AI workloads.
Original source
We have no link to the original source for this story yet.
Where we saw it
The AI Report
AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it. How it works