A breakthrough by engineers at BitEnergy AI could drastically reduce the energy demands of AI applications.
The team has developed a method called Linear-Complexity Multiplication, which replaces energy-intensive floating-point multiplication (FPM) with efficient integer addition.
This innovative approach reportedly slashes electricity use by an impressive 95% without compromising performance.
Currently, AI systems like ChatGPT consume vast amounts of power—enough daily energy to supply 18,000 U.S. homes.
With predictions of AI energy consumption matching Bitcoin mining, this discovery could bring much-needed sustainability.
While specialized hardware is required for …
