A faster way to estimate AI power consumption
Reports and Proceedings
Updates every hour. Last Updated: 12-Jun-2026 08:15 ET (12-Jun-2026 12:15 GMT/UTC)
The EnergAIzer technique can predict how much power a certain AI workload will consume when run on a particular processor. This method could help data center operators and algorithm developers improve the sustainability of AI workloads.
After a surprising discovery that overcomes a longstanding problem in fiber optics, MIT researchers demonstrated a biomedical imaging technique that is faster and more precise than other methods, which could help scientists and clinicians study new brain therapies.