JPMorgan Asset Management Executive Says Foreign Capital Is Funding North America's AI Compute Buildout

Foreign investors are becoming a bigger part of the financing stack behind North America's AI compute expansion, with JPMorgan Asset Management executive Charles Wu saying that banks and institutional investors from Asia and the Gulf are helping fund a sector whose capital expenditure is close to $1 trillion. The remarks put a financial frame around the AI race: the constraint is no longer only model capability, but who can pay for the data centers, chips, power and networks needed to run them. Wu, JPMorgan Asset Management's head of Asia Pacific alternatives institutional client strategy, told attendees in remarks at the SuperReturn conference in Singapore that foreign banks, pension funds, sovereign wealth funds and insurers are playing a major role in North American AI computing growth. The pull is obvious: J.P. Morgan Asset Management estimates hyperscaler capital spending surpassed $400 billion in 2025 and is on track to approach $800 billion in 2026, roughly 10 times 2019 levels, while capex is set to absorb 90% to 100% of operating cash flow in 2026. It also says cloud revenue across major platforms grew an average 44% year over year in the first quarter of 2026 and that more than 70% of data center capacity under construction is already leased. Physical capacity remains tight. CBRE's H1 2026 North America data center report says primary-market supply reached 10,903 megawatts, under-construction capacity hit 7,481.1 megawatts, vacancy fell to 1.4%, and 80.4% of under-construction capacity was preleased. That scarcity is why global capital is moving into debt financings, development equity, joint ventures and operating platforms rather than only public tech stocks. High-profile projects show the same pattern. The Stargate Project said it intends to invest $500 billion over four years in U.S. AI infrastructure for OpenAI, with $100 billion deploying immediately and initial equity funders including SoftBank, OpenAI, Oracle and MGX; later updates put Stargate at nearly 7 gigawatts of planned capacity and more than $400 billion of investment over the next three years. Separately, Brookfield launched a $100 billion global AI infrastructure program with NVIDIA and Kuwait Investment Authority, anchored by a fund targeting $10 billion of equity commitments, while NVIDIA announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to mobilize more than $500 billion of third-party capital over time. The reception is split. Supporters see foreign and institutional money as a way to relieve compute scarcity for developers and enterprises, finance long-lived power and real-estate assets, and give pensions and insurers exposure to contracted AI infrastructure instead of only volatile software equities. Skeptics argue that the boom is moving risk into less transparent places. A Brookings Papers on Economic Activity draft estimates AI infrastructure investment in data center buildings, power systems, networking, chips and other equipment will total $10.3 trillion from 2025 to 2032, equal to an average 3.63% of U.S. GDP per year, and warns that risks are migrating into off-balance-sheet joint ventures, private credit, securitizations, special-purpose vehicles, lease commitments and loan guarantees. Power and local consent are the other hard limits. Lawrence Berkeley National Laboratory says U.S. data centers consumed about 4.4% of total U.S. electricity in 2023 and could consume 6.7% to 12% by 2028. Gallup found seven in 10 Americans oppose AI data centers in their local area, including 48% who are strongly opposed. The upshot is that Wu's comments do not mean global cash has solved the AI buildout. They mean the buildout is becoming a global infrastructure trade whose returns depend on end-user AI demand, electricity delivery, local approvals, chip replacement cycles and whether investors can see the leverage accumulating around the boom.