The rapid growth of generative AI is expected to significantly increase electricity consumption, especially in data centers. According to Bloomberg Intelligence (BI), by 2030, data centers could use ten times more energy than they do today. This surge in energy demand is driven by the need to support AI workloads, which require a lot of computing power.
Solar and Battery Solutions to Meet Energy Needs
As AI becomes more integrated into businesses, the need for solar power and battery storage solutions is growing. Companies worldwide are racing to incorporate AI into their systems, leading to a greater strain on energy resources. Nvidia, a leading tech company, is creating hardware solutions to meet this rising demand. For example, their upcoming Blackwell GPU is expected to cause a significant increase in energy use.
In fact, when Nvidia releases the Blackwell GPU in 2025, BI predicts that it could raise total U.S. power consumption by 1% within its first year. Furthermore, it could lead to a 50% increase in energy usage by data centers.
Renewable Energy as a Solution
As the energy infrastructure is already under pressure, BI suggests that data centers may turn to renewable energy sources like solar and wind power to meet their needs. By 2030, data centers could consume up to 17% of the U.S.’s total electricity. This growing demand offers opportunities for renewable energy companies.
BI predicts that companies like First Solar, Enphase, and Sunnova could see their revenues approach $40 billion by 2026. These companies are well-positioned to benefit from the rising demand for alternative energy solutions.
Role of Natural Gas and Decline of Fossil Fuels
Despite the push for renewables, BI also believes that the natural gas industry will benefit from the rising energy needs of data centers. By 2030, data centers could increase natural gas demand by up to 10 billion cubic feet (bcf) per day. At the same time, the need for coal and other fossil fuels is expected to decline as renewables take center stage.
Nuclear Energy May Arrive Too Late
While nuclear energy could be a clean energy solution, BI suggests that nuclear power may not be able to support the AI revolution in time. Building new nuclear reactors takes at least 10 years, which may be too late to meet the immediate energy needs driven by AI growth.
Conclusion
In summary, the growing demand for AI is set to significantly increase energy consumption, particularly in data centers. To meet this demand, the focus will likely shift towards renewable energy sources like solar and battery storage. At the same time, natural gas may still play a crucial role, while nuclear energy might not be ready to contribute in time for the AI boom. As AI continues to grow, companies involved in energy and technology stand to benefit from the rising demand.
