As artificial intelligence expands, its soaring electricity use is putting unprecedented pressure on power grids, driving up costs and threatening to slow Europe’s energy transition. But a new report shows that only light-based computing can keep AI’s ever-expanding power demands at bay without triggering a surge in emissions.
The study, published today by Photonics21 and compiled by market intelligence agency TEMATYS, finds that as electricity use from AI rises sharply, and conventional silicon chips struggle to keep pace with the twin pressures of ever-larger models and expanding data centres, Europe risks undermining its clean-energy goals and increasing reliance on fossil fuels.
Entitled “AI Desperately Needs Photonics,” the report concludes that photonics, which uses light rather than electricity to move and process information, offers the only viable pathway to scale up AI computing while dramatically reducing energy consumption. The study reveals that integrating photonic technologies into data-centre and chip architectures can help AI growth remain compatible with Europe’s climate commitments.
A Growing AI Problem
AI is scaling faster than the world’s power grids can handle. Each new generation of larger, more complex models demands exponentially greater computing power, sending data-centre energy use soaring.
The International Energy Agency puts global consumption at roughly 415 TWh in 2024, and TEMATYS warns it could more than double by 2030 as AI workloads dominate digital infrastructure.
Sébastien Bigo, Nokia Bell Labs Fellow and Photonics21 Work Group Leader for Digital Infrastructure, said, “Photonics can provide the infrastructure that will determine whether AI becomes cleaner and more competitive or simply costlier and dirtier. Europe has the research base to lead; what it lacks is coordinated investment and industrial scale.”
The report shows that bringing photons much closer to the compute (whether through co-packaged optics or other emerging photonic integration approaches) and using light to handle increasing parts of data movement or specific acceleration tasks can considerably improve the carbon and cost profile of AI. While photonics is not a replacement for CPUs or GPUs, it can become a critical complement that eases the pressure on them. If Europe wants sovereignty in future AI hardware and to meet its climate commitments, the report stresses that Europe’s policy and investment decisions in the coming years will be decisive.
Photonics is Already Essential
The report finds that photonics is already indispensable to the digital world. Fibre optics underpins today’s internet and data-centre infrastructure, and the next step of integrating light directly into chip architectures through co-packaged optics is already underway.
These advances, however, will not be enough on their own. Even with industry claims of 3.5 times greater power efficiency, TEMATYS concludes that these gains cannot fully offset the surge in energy demand driven by AI’s rapid expansion.
The study also shows that optical computing is no longer theoretical. Laboratory breakthroughs and start-up prototypes, including MIT’s 2024 demonstration of an integrated photonic chip performing neural-network computations entirely with light, reveal a clear technical path towards fully photonic processors.
Finally, the report warns that Europe has the expertise but not yet the scale. World-class research and promising start-ups exist across the continent, but without faster investment, large-scale manufacturing, and workforce development, Europe risks surrendering this strategic market to competitors abroad.
The report urges European institutions, national governments and private investors to treat photonics as strategic infrastructure for AI. Recommended actions include: dedicated funding for pilot manufacturing, incentives to scale photonics start-ups, inclusion of photonics in Chips/AI/Green tech funding streams, and skills initiatives to build a manufacturing workforce. Without these steps, Europe risks ceding critical parts of the AI hardware value chain to overseas competitors while facing higher energy bills and emissions.