One Europe’s leading climate-tech VC firms has called on a “fundamental rethink” of traditional computing methods in light of the escalating energy consumption by AI applications.
Berlin-based World Fund warns a simple switch to renewable energy will not be sufficient to decarbonise AI computing.
Daria Saharova Said, “We must rethink how we do computing, from the chips and materials we use to the software we run,” at the future green computing conference in Munich today.
World Fund, Intel’s deeptech accelerator Ignite and Dealroom teamed up at the event to unveil a report that outlines a range of emerging technologies – from chips made in outer space to brain-like processors – to curb AI’s huge appetite for energy.
The report maps the green computing eco-system using data from Dealroom. It identifies 65 green computing startups, 54 of them European, who have raised $900mn collectively. Over half of the companies in this list were founded within the last five years. Twelve of them were born just 12 months ago. The report highlights three key technologies which have the greatest potential to decarbonise AI.
The report highlights three key technologies that have the most potential to decarbonise AI. These materials could reduce AI’s energy usage by improving the efficiency and thermal performance of computing hardware.
A Welsh startup Space Forge is one of the leading innovators within this field. The company uses the microgravity and extreme temperatures in space to produce semiconductors it claims are up to five times more pure than those made on Earth.
“We have pushed the efficiency and performance of silicon chips to its limit,” said Joshua Western. CEO and cofounder of Space Forge.
New computing paradigms such as quantum, optical, and neuromorphic computing are also promising. Quantum computers promise to solve complex computations much faster than traditional machines, potentially reducing overall energy consumption and computational time.
Ines de Vega,VP of Innovation at IQM, Europe’s most-funded quantum computing start-up, said: “Classical computers have become too large, too expensive and use too much water and energy.” Quantum computing can find new solutions for climate change, but also drastically reduce energy consumption in computing itself. Black Semiconductor, a German company, has demonstrated that it could dramatically increase processing speeds. The company’s photonics chips could transmit signals up to 1,000 times faster than electronic chips. Anastasiia
Nosova, former chip engineer for German semiconductor giant Infineon as well as host of Anastasi In Tech, argued that photonic chipscould be 100 times energy efficient than regular semiconductors. She said that photonic chips are “one of the most important developments” in computing today. Deep Render, a London-based startup, is one of those working on this. The company uses deep learning to compress files and retain quality beyond what had previously been possible. This reduces the amount of data that must be transmitted or stored and, therefore, the computing power required. These technologies are still in their early stages of development, but they have great potential. According to OpenAIthe energy required to train AI models doubles every three to four month. Saharova said that in order for these computing solutions to scale up in Europe, “we need a lot venture capital as well as government backing.” She believes that Europe must allocate EUR1 Trillion to bring climate tech and green computing to the “level” it needs to reach.
At TNW Conferenceon June 19-20, Amsterdam, the future of AI is a hot topic. Tickets for the eventare now available Use the code TNWXMEDIA2025 to get 30% off.
