JAKARTA - The explosion of artificial intelligence has caused the electricity demand for data centers to skyrocket. In China, a number of nuclear fusion companies are now trying to cut the commercialization target for the technology from decades to only five to 10 years.
Yicai Global, quoted on Friday, July 31, reported that NovaFusion Energy Technology or NovaFusionX and a number of similar companies are targeting the large electricity needs that arise from the rapid development of AI.
"The explosive growth of AI has driven the electricity demand to rise exponentially, forcing us to shorten the commercialization window of nuclear fusion to only five to 10 years from the previous 30 to 50 years," said NovaFusionX founder Guo Houyang, in a forum on July 28.
"If we can't solve the electricity problem in five to 10 years, AI computing power will definitely hit the electricity limit," he said.
Citing Goldman Sachs, Guo said the global data center electricity demand is likely to reach around 1,130 terawatt-hours by 2030. This figure is roughly equivalent to Japan's annual electricity consumption.
According to Guo, AI data centers require a stable, carbon-free electricity supply, can be placed in many locations, and remain competitive in terms of cost. He assessed that conventional power grids and renewable energy cannot meet all of these requirements at once.
Nuclear fusion is the process of combining light atomic nuclei to generate energy. The technology has long been considered a future source of clean energy, but the path to commercial plants is still arduous.
Guo said there were three factors that made the commercialization target accelerated. First, the surge in AI computing creates a huge need for stable and carbon-free electricity 24 hours a day.
Second, the technology of American companies such as Helion Energy has been shown to have scientific and engineering feasibility for the field-reversed configuration or FRC line.
FRC is an approach to contain plasma using a certain configuration of magnetic fields. Plasma is a very hot gas containing charged particles which is the key in the fusion process.
Third, the Chinese hard-tech startup ecosystem is considered mature because it brings together talent, industry, policy, and capital.
However, a number of experts in the same forum reminded that the distance between laboratory devices and commercial products is still far away. The biggest challenge is engineering and manufacturing so that the technology can be mass-produced and work reliably.
NovaFusionX was founded in April last year. The company is targeting the first plasma discharge by the end of this year and wants to become the first Chinese private company to exceed 100 million degrees Celsius next year.
Guo said NovaFusionX is targeting net energy gain from deuterium-tritium fusion with Q greater than 1 by 2029.
Q greater than 1 means the energy produced is greater than the energy used to trigger the fusion process.
The company also targets the construction of the first small and dispersed fusion demonstration plant in China in the early 2030s. NovaFusionX is talking to a number of prospective customers for electricity purchases, but has not signed any agreements.
Unlike tokamaks, which use a dominant approach using a large ring-shaped magnetic field to hold the plasma, NovaFusionX chooses a FRC path combined with a small modular reactor design. This approach is known as FRC-SMR.
According to Guo, a small modular unit costs about 100 million US dollars to build. The cost of electricity could potentially fall below 7 cents yuan or about 1 US cent per kilowatt-hour.
However, he reminded that the figure was still a projection and had not been proven through commercial operations.
Another player, Dongsheng Fusion, is sticking with the tokamak path. But it's opting for a harder-to-find and more environmentally friendly deuterium-helium-3 fuel.
Dongsheng Fusion General Manager Nie Lin said the choice was made so as not to repeat the investments already made by state-backed companies. About 90 percent of Dongsheng's basic technology is said to overlap with national programs.
According to data from the Fusion Industry Association quoted by Yicai, 56 fusion companies in the world have raised a cumulative fund of 14.2 billion US dollars. As much as 4.5 billion US dollars was obtained in the past year, up 69 percent from the previous year.
For the first time, private capital also surpassed government research funding as the main source.
Shanghai-based Hongfu Investment has set up a fund of more than 200 million yuan, or about 28 million US dollars, for the controlled fusion sector. Hongfu Investment Director Luo Yusheng said that the development of AI has heated up the fusion sector, from plasma control to simulation.
Luo assessed that suppliers of components and materials such as magnets, power electronics, and superconducting materials could generate revenue faster than reactor developers.
However, he gave a harsh note. Reactor developers do have the opportunity to save trillions of yuan, but no one can be sure who will succeed and how long it will take.
Helion Energy in the US is a concern for Chinese companies. Helion has signed a 50-megawatt power purchase agreement with Microsoft, with a target for delivery in 2028. However, the company has also not achieved commercial electricity generation.
Guo said the US excels in the validation of prototypes from zero to one. China, according to him, has strengths in complete supply chains, large-scale AI application scenarios, high-end manufacturing, and policy support for the 15th Five-Year Plan.
"The United States does have a big advantage in the engineering iteration from zero to one, but China still has a huge potential in the one-to-many phase in large-scale industrial production," said Guo.
The manufacturing actors still remind us that the fusion path is not easy. Ni Jun from Shanghai Jiao Tong University and Contemporary Amperex Technology said that energy breakthroughs must solve scientific, technical, and engineering problems before becoming a product that can be produced.
The chairman of Shanghai Firm Lithium New Energy Technology, Zhang Xi, said many companies claimed to have almost entered the market, but most of them never really got there.
Before fusion produces commercial electricity, Wison New Energies Product Technology Center Deputy General Manager Deng Shujun assessed that a more realistic transition is renewable energy combined with energy storage and natural gas.
Deng also said many clean energy projects were still stuck at the conceptual design stage.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
