Watch Desk posted an update
Sparxity, formerly known as SpiNNcloud, says it has closed a Series A, bringing the funding secured by its team to $90 million. The company is developing computing systems for sparse AI workloads, which aim to avoid spending effort on inactive parts of a model.
Why it mattersSparxity says it has completed three chip tapeouts, produced more than 70,000 chips and serves over 30 customers. It is targeting up to 44 times faster inference on ultra-sparse models, a company target rather than an independently established result. That makes this more than a funding headline: it is a sizeable bet on whether specialised hardware can make some AI workloads faster and more energy-efficient. The AI Insider reports that the round was led by ICM HPQC Fund, with GlobalFoundries among the participants.
Discuss: Will specialised sparse-AI chips prove a better route to efficiency than making general-purpose accelerators bigger?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.