Vaire, a U.K. startup, has announced that initial tests of its new computer chip technology, reversible computing, show it could cut the electricity needed for many computations, including those used in AI, by half. The company’s CEO, Rodolfo Rosini, and CTO, Hannah Earley, are hopeful about the impact of their chip on the efficiency of AI workloads. The startup’s tests have not yet been independently verified, but Vaire’s chip test kits are being sent to potential customers and academic labs for more evaluation.
If the results are confirmed, this breakthrough technology could be widely adopted commercially, changing how AI workloads are handled and significantly reducing the energy they need. Reversible computing could lead to chips that recycle nearly all the energy they use, giving off very little heat. This technology tackles one of the AI industry’s biggest challenges: the rapidly growing electricity demand of data centers filled with graphics processing units (GPUs), special computer chips commonly used to train and run AI models.
The International Energy Agency predicts electricity demand from data centers will more than double by 2030 to 945 terawatt-hours, about 3% of global energy demand and equal to Japan’s energy use. AI workloads will drive most of this growth, with energy needs in data centers expected to quadruple in the next five years. Many AI industry insiders think these projections are low.
It is hard to estimate the carbon footprint of AI systems, as it mostly depends on the power sources for any given data center.
Energy-efficient AI chip technology
While many AI data centers in the U.S. and Europe are being built using renewable energy like solar and wind power or nuclear power, which emit little carbon dioxide, other regions still depend heavily on gas or coal-fired power, which increases CO2 emissions.
There is also worry that tech companies are buying so much green power for their data centers that other potential customers might have to use fossil fuels. The cooling needs of AI data centers are another challenge. GPUs use much more power than other computer chips, like central processing units (CPUs), and thus generate more heat, requiring more energy and water for cooling.
Also, many AI data centers are being built with low groundwater supplies. To address these issues, Rosini co-founded Vaire in 2021 to use reversible computing to reduce the energy demands of increasingly large AI models. Vaire worked in stealth for three years, with $500,000 in seed funding from London-based 7percent Ventures and a group of angel investors, perfecting the design of its reversible computing chip.
Last year, the company raised more seed funding from 7 percent Ventures, Lifeline Ventures, SeedCamp, and Jude Gomila, co-founder of mobile ad company Heyzap. Vaire has raised a total of $10 million and has about 20 employees across offices in London, Cambridge (England), Sunnyvale (California), and Portland (Oregon). Reversible computing’s ability to greatly reduce the energy use of AI workloads could help lessen the environmental impact of the AI boom and ensure a more sustainable future for the growing industry.
Rashan is a seasoned technology journalist and visionary leader serving as the Editor-in-Chief of DevX.com, a leading online publication focused on software development, programming languages, and emerging technologies. With his deep expertise in the tech industry and her passion for empowering developers, Rashan has transformed DevX.com into a vibrant hub of knowledge and innovation. Reach out to Rashan at [email protected]
