Citigroup creates a dedicated team for AI infrastructure investments

When one of the world's largest banks restructures its organizational hierarchy for a specific technology, it's no longer just a corporate announcement—it's a marker of the era. Citigroup announced the creation of a dedicated group that will exclusively handle deals in the AI infrastructure sector. The team includes key figures from investment and corporate banking divisions, which speaks to the seriousness of the initiative.

To understand the scale of this decision, one must look at the context. Over the past two years, global spending on AI infrastructure—data centers, GPU clusters, energy capacity, network equipment—has grown exponentially. According to various estimates, in 2025 alone, the largest technology companies invested more than $200 billion in this sphere, and by 2027, cumulative investments could exceed a trillion. Such capital volumes inevitably attract bankers. But Citigroup went further than simple interest—the bank institutionalized this direction, carving it out into a separate business vertical.

Structurally, the new group combines the expertise of two key divisions. Investment bankers will manage deals involving mergers and acquisitions, IPOs, and debt placements for AI-sector companies. Corporate bankers will handle lending and project financing—data center construction, equipment procurement, energy infrastructure deployment. This cross-functional approach allows Citigroup to offer clients a complete spectrum of financial services under one roof, which is critical in a sector where deals often require both project financing and M&A consultations simultaneously.

Citigroup is not the first to take such a step. Goldman Sachs began scaling teams working with AI companies back in 2024, and Morgan Stanley actively consults on data center deals. However, the formalization of a separate group at Citi is a sign that AI infrastructure has ceased to be a niche subsector of tech banking and has become an independent direction comparable in significance to energy or real estate. Essentially, Wall Street is beginning to treat computational capacity for AI the same way it has for decades treated oil infrastructure—as a foundational layer of the economy requiring specialized financial services.

Behind this decision lies pragmatic calculation as well. Commissions from AI-sector deals are growing rapidly. Building one major data center costs billions of dollars, and there are now dozens of such projects worldwide. Companies like CoreWeave, Lambda, and Crusoe Energy are attracting massive funding rounds. Hyperscalers—Microsoft, Google, Amazon—are entering long-term contracts for energy supply and capacity leasing, each requiring complex structuring. For banks, this is a gold mine, but to exploit it, you need people who simultaneously understand both financial instruments and the technological specifics of AI infrastructure.

There is also a deeper signal. The formation of such teams at the largest banks means the financial system is seriously betting on the long-term nature of the AI boom. Banks don't create new divisions for short-term trends—it's expensive and organizationally complex. If Citigroup is willing to restructure, it means the bank's internal analysts see sustained multi-year demand. This, in turn, could become a self-fulfilling prophecy: the presence of specialized banking teams makes it easier to attract capital for AI projects, which accelerates infrastructure development, which generates even more deals.

For Russian observers, this news is interesting in the context of the global race for AI infrastructure. While Western banks are building financial pipelines for trillion-dollar investments in computational capacity, the question of who and how will finance similar infrastructure in other regions remains open. One thing is clear: artificial intelligence has definitively moved beyond the technology sector and become a system-forming factor for global finance. When banks at the level of Citigroup reshape their organizational structure around a specific technology, arguing with that is already meaningless—only adaptation remains.