Citigroup to Finance $3T AI Infrastructure Buildout
Citigroup is positioning itself to help finance a global AI infrastructure expansion estimated to be a $3 trillion opportunity. The bank's focus on chips, data centers, and cloud platforms signals that AI is now being procured and capitalized like other core utilities such as energy and networks. This shift emphasizes the need for resilience, interoperability, and auditability in AI platforms to meet enterprise procurement standards.
- Citigroup's new "AI Infrastructure Banking" team is a cross-disciplinary group led by senior bankers including Ric Spencer, vice chair of technology, and Ashish Agrawal, co-head of real estate, to break down internal silos. Other leaders include technology financing head Alex Watkins, technology corporate banking head Doug Baird, and M&A banker Ben Mortimer. - An internal memo indicates the financing strategy will involve a mix of bank debt, private credit, infrastructure and real estate financing, and structured investment-grade debt to fund the build-out. - The global AI infrastructure market is projected to reach $758 billion in spending by 2029, with accelerated servers accounting for more than 95% of that total, according to IDC. - The push for this financing comes as individual tech companies are making massive investments, including Amazon's $12 billion for data center campuses in Louisiana and Meta's projected 2026 capital expenditures of up to $135 billion. - The U.S. is the current leader in AI infrastructure, accounting for 76% of total spending in the second quarter of 2025. However, the build-out faces significant headwinds, as planned data centers are projected to require 150 gigawatts of power, a 12% increase over the entire current U.S. grid capacity. - As part of its broader AI strategy, Citigroup's Markets Strategic Investments unit recently made its first investment in Japan, taking a stake in the Tokyo-based startup Sakana AI. - The financing initiative is a response to a global data center construction boom that saw a record $61 billion in investment in 2025, driven by the intensive compute and energy demands of AI workloads.
