AI is reshaping the fundamentals of the U.S. economy.

In the second quarter of 2026, U.S. private enterprises increased their AI-related investment year over year by $300 billion, a growth rate of 25%. The annualized investment scale reached a record high of $1.5 trillion. This signals that AI investment has upgraded from the "arms race" among tech giants to a capital migration at the scale of the broader economy.

The structural features of this investment surge are clear: investment in computers and peripheral equipment is the primary driver, followed by investment in communications equipment, software, and data centers. Over the past two years, AI-related investment has increased by $500 billion, a rise of 50%. In particular, corporate investment in computers and peripheral equipment has grown by more than double, reflecting that AI infrastructure buildout has entered a phase of full-scale acceleration.

From a macro perspective, the multiplier effects of AI investment are being fully unleashed. AI’s direct investment contributes about 25% to 33% of the U.S. GDP growth in the recent period. JPMorgan Chase research further points out that in the first half of 2026, roughly two-thirds of U.S. GDP growth will come from corporate AI investment. This structural shift is far-reaching in meaning. Traditionally, consumption accounts for about two-thirds of U.S. GDP; now, the investment side’s contribution to growth is approaching that of consumption. The U.S. economy is shifting from an “consumption-driven” model to an “AI-investment-driven” one.

The spillover along the AI investment industrial chain is also worth paying attention to. In the capital expenditures of ultra-large cloud providers, about 75% goes to physical infrastructure—data center buildings, power systems, cooling equipment, and network hardware, rather than to chips themselves. This means that AI investment is pulling through traditional Keynesian multiplier channels, boosting real-economy industries such as cement, steel, electricity, and construction. At present, the U.S. has around 4,000 data centers in operation, with another roughly 3,000 under construction or already announced. With nearly double the growth pace, “digital factories” are being built at a rate comparable to factory construction.

At the same time, the AI investment boom also brings structural divergence. Capital expenditure in non-AI sectors is basically stagnant, and the U.S. economy is splitting into two parallel worlds: an “AI economy” and a “non-AI economy.” The former’s spending grows exponentially and employment is strong; the latter’s spending stalls and profit margins come under pressure. This divergence is not cyclical—it is structural.

A deeper shift is that AI investment has become “immune” to monetary policy. Whether the Federal Reserve raises rates or not, ultra-large cloud providers must keep investing in data centers—not because they are betting on AI by borrowing money, but because “if we don’t build data centers, our customers will move to competitors.” A triple set of structural forces locks in the direction of investment: the prisoner’s dilemma of the technology race, the political consensus around reshoring semiconductors, and fiscal stimulus already written into law.

Of course, beneath prosperity there are also hidden concerns. If the AI industry cannot demonstrate sufficient commercial value within a certain period of time, capital markets may no longer tolerate it, leading to a significant slowdown in U.S. economic growth. However, no matter what, AI investment—now the core engine of the United States’ economic growth—cannot be shaken. It not only changes the flow of capital, but also rewrites the logic of how the macroeconomy operates.