Expanding U.S. artificial intelligence infrastructure will require trillions of dollars in funding, and the money could come from ordinary Americans’ deposits and pension assets, a projection said. As AI competition moves beyond software development into investment in physical infrastructure such as data centres, power grids and semiconductors, the role of capital markets is also growing, it said.

On May 25, blockchain outlet Cryptopolitan reported that BlackRock CEO Larry Fink (래리 핑크) stressed in an annual letter to shareholders that maintaining U.S. AI leadership requires massive capital investment. He said the United States must keep investing in research and development, data centres, power grids, semiconductors, telecommunications cables and talent to stay in front in AI. “To support innovation of this scale, a strong capital market is essential,” he said.

Attention has focused in particular on the scope of potential funding sources. Fink said AI infrastructure investment could draw not only on institutional money but also on bank deposits and pension funds. That would mean a structure could form in which U.S. households’ long-term savings underpin the foundations of the AI industry.

He also said the pace of investment in the United States is still not sufficient. Fink said at the Milken Institute Global Conference in May that “we are not moving fast enough.” On claims in parts of the market that AI is overheating, he pushed back, saying, “There is no AI bubble. It is the opposite.”

BlackRock already holds large stakes in key AI-related technology companies, including Apple, Microsoft and Nvidia. At the same time, it is expanding investment in physical infrastructure. BlackRock bought infrastructure investment firm Global Infrastructure Partners in 2024 for $12.5 billion, increasing the share of energy and infrastructure assets.

BlackRock and Global Infrastructure Partners then jointly pursued an AI data centre investment project in March 2025. Microsoft, Nvidia, xAI and Abu Dhabi sovereign wealth fund MGX also took part. Data centres are key facilities for operating large AI models and require massive investment that includes not only servers but also power equipment, cooling systems, fibre-optic networks and backup power units.

Microsoft CEO Satya Nadella (사티아 나델라) said at the time that AI infrastructure would become a key factor shaping future industry and regional economic growth, and stressed the need to expand joint investment.

Major Wall Street financial firms are also backing increased AI infrastructure investment. JPMorgan Chase CEO Jamie Dimon (제이미 다이먼) said at a recent event in New York that as much as $1 trillion could be invested in the future in data centres, semiconductors and network equipment. He said that even if technology investment does not generate linear returns in the short term, it is likely to create economic value in the end.

Dimon said, “Technology generally ends up paying for itself, but the process is not linear.” He added that while it is difficult to predict in advance which companies will win or lose, the influence of AI technology itself is large enough.

Against this backdrop, U.S. AI competition is spreading beyond simple model development into a race to secure physical infrastructure such as power, semiconductors, telecommunications networks and data centres. As the possible funding sources are discussed as expanding beyond Wall Street institutions to include ordinary people’s deposits and pensions, AI is emerging as a factor that could reshape not only the technology sector but also financial markets and the structure of long-term capital management.