Nations accelerate AI expansion amid rising energy strain and investment risks

Monday 22 December 2025 - 09:20

By: Dakir Madiha

Nations accelerate AI expansion amid rising energy strain and investment risks

Japan has unveiled a five-year plan to invest around 1 trillion yen (approximately $6.34 billion) in a new artificial intelligence venture, positioning itself among global powers seeking control over the technological and economic foundations of AI infrastructure. The initiative, backed by SoftBank and Preferred Networks, aims to develop a domestic foundation model by 2026 and strengthen the nation’s independence in advanced computing as competition intensifies with China and the United States.

The strategic effort reflects Tokyo’s growing concerns over national security dependence on foreign-developed AI systems. It comes as China pushes for technological self-sufficiency, with semiconductor companies such as Biren Technology and Baidu’s Kunlunxin preparing major stock listings to secure capital and challenge international chip leaders. In parallel, SoftBank is accelerating plans to secure $22.5 billion in funding for AI-related ventures, marking one of the largest private investments in the field to date.

Mounting power demands pressure grids

The rapid construction of data centers is beginning to test global energy capacity. Industry forecasts suggest total electricity consumption by data centers could nearly triple from 415 terawatt-hours in 2024 to about 945 terawatt-hours by 2030, surpassing the combined energy use of Germany and France. In the United States, projections indicate that data centers could consume up to 12% of national power supply within three years.

Concerns over rising energy bills have drawn the attention of lawmakers. Three U.S. senators recently launched an inquiry into whether large technology companies are driving up domestic electricity prices. Meanwhile, utilities are investing heavily in grid modernization, with estimated costs of up to $1.4 trillion over the next 25 years expenses likely to reach end-users.

The surge in demand has boosted the outlook for energy technology firms. Jefferies upgraded its rating for GE Vernova, citing strong performance in gas turbines and grid services as AI-driven electricity needs grow rapidly. The company expects to maintain a substantial generation capacity backlog extending into 2029.

Economic doubts and sustainability constraints

Analysts are voicing concerns over potential market distortion. With numerous new entrants ranging from private investors to cryptocurrency miners repurposing infrastructure financial risks associated with overbuilding are rising. The influx of speculative capital has pushed U.S. data center financing to nearly $180 billion this year, adding volatility to an already overheated market.

Skeptics warn that rapid hardware obsolescence could undermine profits, as state-of-the-art chips may need replacement every few years. If annual capital expenditures of $400 billion depreciate faster than projected, the resulting costs could outpace revenue across the industry. Environmental pressures are also mounting, with global water usage for chip cooling expected to reach 1,200 billion liters by 2030.

Despite the hurdles, momentum remains strong. Nvidia is expected to boost production of its H200 chips following expanded export approvals, while U.S.-based AI infrastructure projects, including the massive Stargate initiative, continue to attract enormous funding and political backing.