I wrote in this newspaper two years ago that America was still the economic envy of the world and standing at the edge of a boom. We were further ahead in AI than anyone. The technology, along with the infrastructure build-out it requires, seemed likely to produce tremendous growth. That growth was the answer to America’s economic Achilles’ heel, its federal debt, which has now reached a record $40 trillion. Expanding the economy by more than 3% a year is ultimately how countries meet the promises they have made to their citizens without simply borrowing more to pay for them. AI makes that kind of growth more achievable than it has been in decades.

All this remains true. But just as America needs to invest more than ever, capital is becoming scarcer. Governments across the developed world are running huge deficits. That borrowing is now colliding with tremendous demand from the private sector, much of it from technology companies. Add the trillions needed for power, housing, industrial capacity and other infrastructure, and you get the biggest competition for capital I have seen in 50 years in finance.

Some people look at the sums going into AI and see a bubble. As fiduciaries, we have to distinguish durable investments from speculation. Yes, some companies may fail. The railroads produced both fortunes and bankruptcies. So did the internet. It isn’t difficult to imagine a future with a handful of companies worth $10 trillion each.

I see a different risk. When all that investment demand collides with a limited supply of capital, borrowing becomes more expensive for everyone, from families buying homes to businesses looking to invest. The companies with the deepest pockets have an obvious advantage. Those are the hyperscalers. The goal should be to finance their growth without starving the rest of the economy of capital.

We can’t finance this investment cycle simply by moving the same dollars from one asset to another. Our capital markets will have to tap new pools of investable capital. After the financial crisis, banks had less capacity to lend. The capital markets filled much of the gap. Private credit grew. Companies, entrepreneurs and small businesses found new sources of financing. That evolution financed the recovery and helped keep the U.S. an attractive place to build a company.

To keep it that way, we need to do three things: First, scale the way this investment cycle is financed. Much of AI infrastructure financing today is bespoke, involving a relatively small number of companies and lenders. The investment required over the next several years is too large to finance that way alone.

I began my career as a mortgage trader when the mortgage-backed securities market was still young. It took decades to build the data, standards and market structure that allowed mortgages to become one of the world’s largest asset classes. We don’t have decades this time. We need to turn new kinds of infrastructure into investments that far more pools of capital can finance. That is what securitization has done before: Mortgages became securities, loans found new markets, and assets once held on bank balance sheets became investable at enormous scale. Computing infrastructure may be next.

Second, recognize that data centers are only one part of this build-out. Before the technology can make workers and businesses more productive, we have to finance and build the power and other infrastructure it requires. China’s enormous expansion of power generation is a reminder that leadership in technology will depend on more than who builds the best models. Abundant and affordable power will be a competitive advantage on par with technology itself.

That will take faster, more predictable permitting and a concerted effort to train electricians, welders and technicians. These are family-sustaining jobs that can become careers, most of which don’t require a four-year degree. These skills will remain valuable long after today’s construction projects are finished.

AI’s enormous potential comes with legitimate concerns about safety and its effect on people. Realizing its benefits will require the public and private sectors to manage those risks responsibly.

AI promises abundance while its build-out is creating scarcity. We already have trillions of dollars chasing limited supplies of power, chips, equipment and skilled workers. Building more capacity is how we keep those shortages from driving prices higher. It also determines how broadly the technology can spread, keeping it affordable for startups and small businesses.

Third and most important, America needs more investors. Helping more people become long-term investors would give them a greater stake in the country’s growth while adding another source of capital to finance it. For much of the past half-century, capital has grown faster than wages, rewarding people who owned financial assets more than those who relied on their paychecks. Roughly 4 in 10 Americans own no stock at all. This time should be different.

We may be entering one of the great investment cycles in American history. I am bullish on the technology and on America’s ability to finance this build-out while keeping capital available to households and businesses. If we unlock more investment through new methods of financing, make permitting much faster, and build the infrastructure we need, the economy can grow quickly enough to keep our debt manageable.

This boom can produce something larger than a handful of extraordinarily successful companies. The true measure of success will be how many Americans benefit from the growth. Mr. Fink is chairman and CEO of BlackRock.