The AI Infrastructure Financing Boom

Tapping New Sources of Capital and Technical Expertise

By Sridhar Nagarajan 
Head of Digital Infrastructure, IDFG, Standard Chartered

Digital infrastructure is booming, with AI data centers being planned and developed at an unprecedented pace and scale. Given the hundreds of billions of dollars of required investment, AI data centers resemble infrastructure, rather than commercial real estate projects such as the traditional data centers of the past. Or, as Jensen Huang, founder and CEO of Nvidia describes them: “productive, investment infrastructure: AI factories.”

As we look ahead, the size of the commitments to finance AI infrastructure could potentially exceed the capabilities of current capital sources, which include banks, public bond offerings, and private capital. This points to a need to broaden the range of global capital sources beyond the traditional lender base. For Standard Chartered, providing access to multiple sources of capital, potentially including sovereign-backed financing institutions, could significantly expand the pool available for AI infrastructure.

As this article will discuss, our differentiation at Standard Chartered is anchored in an overarching theme: the skill sets required in AI finance encompass the ability to analyze power risk, data center risk and permitting risk. For me personally, my background in infrastructure, across traditional asset classes such as oil and gas, power and core infrastructure, enables me to approach AI data centers as utility-like infrastructure, requiring a comprehensive risk allocation framework for securing financing, obtaining permits and other approvals, and working with, not against, the local community. In short, at Standard Chartered, our bankers and investment advisors possess the diverse backgrounds needed to assess risks that are becoming increasingly correlated — and who are backed by the cross-border efficiency of a bank with deep relationships in 55 markets.

As a “superconnector” bank, with a presence in 55 markets, Standard Chartered can help channel capital from global sources as well as technical know-how to support AI infrastructure development.

The Scale of AI Financing

To understand the AI infrastructure landscape, we start with the hundreds of billions of dollars committed this year to date, already eclipsing the total for all of 2025, with the numbers still climbing.

For example, in August, Nvidia announced partnerships with several investment firms to access more than USD500 billion of third-party capital to build out AI infrastructure over time. Nvidia said it is working with the investment firms to create dedicated pools of capital for Nvidia customers “to independently underwrite AI infrastructure.”      

This latest deal further expands on previous projections of AI capital commitments. According to Bloomberg NEF, analyst projections of expenditures by 14 of the largest publicly owned data center operators were estimated at USD800 billion as of February 2026, a 60% increase from about USD500 billion in August 2025. Other estimates show hyperscalers Amazon, Alphabet, Microsoft and Meta deploying a total of USD750 billion in 2026, an increase of more than 80% from 2025. As of the first quarter of 2026, a total of 831 data centers were under construction globally, with 311 sites in the Americas and the rest spread across Europe, the Middle East and Africa (EMEA) and the Asia Pacific (APAC) region, according to Bloomberg NEF. New IT capacity (a measure of power drawn by server equipment) amounts to 15.9 gigawatts currently under construction in the US alone, and accounting for the majority of the 23.1 gigawatts of capacity being developed globally.

To finance these projects, banks have been the dominant players. Recently, the focus began to shift toward more public bond issuances to complement the hyperscaler operating cash flows and the bank financings. However, there are indications that the dynamics in AI infrastructure financing are changing. For example, banks and private capital sources have made massive capital commitments for AI infrastructure, which could saturate their portfolios and make them reticent to provide additional financing in this space. Another indication is that the spreads on recent corporate bond issues by hyperscalers have widened compared to issuances earlier this year and last year.

In our view, this is the optimal time to explore alternative sources of capital, including sovereign-backed financing. Given that memory, chips and power-related components are sourced internationally, from US, European and Asian suppliers, it seems logical that capital sources from multiple countries could be explored to help finance AI infrastructure. For example, to further diversify capital and bridge a potential funding gap, export credit agencies in the US, Asia and Europe could be a source of infrastructure financing.

At Standard Chartered, we understand these connections and, with our deep relationships in local markets, are actively pursuing how best to tap all relevant pools of capital. Given our strengths in global capital flows and expertise in infrastructure development — including building governmental, municipal and community relationships to support large-scale projects — Standard Chartered is uniquely positioned to help advance the AI agenda.

With our global reach across Asia and EMEA, as well as a growing presence in the US, we can help ensure hyperscalers and their partners access the repeatable, scalable and non-dilutive capital sources they require for future development.