The second quarter of 2026 marked a turning point for the US financial industry. The country's largest banks recorded unprecedented growth in investment banking revenue, driven by a global boom in infrastructure construction for artificial intelligence. While traditional markets fluctuated, capital flooded into projects to build data centers, power grids, and computing power necessary for neural networks.
Goldman Sachs emerged as the leader in this marathon. The bank reported record investment banking fees of $3.4 billion. This is a 55% increase compared to the previous year and the best result in the financial institution's history. Bank CEO David Solomon did not hide his optimism, calling the current period an 'AI capital expenditure supercycle.' According to him, demand for financing is growing everywhere, as companies worldwide are seeking funds to create the physical foundation for artificial intelligence development.
Explosive Growth in Securities Operations
An analysis of Goldman Sachs' internal revenue structure shows that the main driver is securities operations. Revenue from equity offerings (IPOs and secondary offerings) jumped by 130%, reaching $985 million. Even more impressive was the growth in the debt instruments sector: revenue increased by 75% to a record $1.03 billion.
The scale of interest in AI projects is confirmed by specific deals. Eight major banks participated in SoftBank's $40 billion loan deal related to OpenAI. This indicates that financing AI giants has become a priority for the entire banking system.
Growth is observed not only at Goldman Sachs. JPMorgan recorded $3.3 billion in fees, which is 30% higher than last year's figures and a maximum since 2021. Morgan Stanley showed a 58% increase, raising revenue to $2.44 billion, while Bank of America grew by 50%, to $2.14 billion.
Forecasts: From $850 Billion to $1.5 Trillion
Bank executives assess the current stage as the beginning of a massive investment cycle. Morgan Stanley CEO Ted Pick predicts colossal investment volumes in data center construction. According to the bank's research, spending on this segment could reach about $850 billion as early as 2026. In 2027, this figure will rise to $1.3 trillion, and by 2028, it will reach $1.5 trillion.
However, Pick warned that the cycle is only in its early stages, and its further development remains uncertain. Bank executives emphasize that the growth in activity is linked not only to AI developers but also to the entire ecosystem around them: from memory and chip suppliers to companies providing power supply and construction.
Citigroup CEO Jane Fraser noted that AI has become the central topic of negotiations with clients. Demand for financing is growing in all segments of the value chain — from high technology to energy.
Risks and Caution: The Energy Problem
Despite the euphoria, banks are beginning to show caution regarding individual projects. JPMorgan CFO Jeremy Barnum admitted that the bank declined some data center deals. The main reasons were questions regarding the availability of electricity and uncertainty about future tenants of such facilities.
None of the largest banks have yet disclosed the exact volume of lending allocated specifically for AI infrastructure, but the overall picture is clear: the total net income of the top five US banks in the second quarter was about $49 billion. This is 39% higher than a year ago. At the same time, the main growth was provided by capital placement and securities issuance operations, while the traditional advisory business grew significantly slower.
David Solomon issued an important warning: the current cycle will not last forever. 'The market will eventually face a reset and a decline in activity, after which a new stage of growth is possible,' he stated. The main challenge for banks remains the economics of data centers and the ability of the energy infrastructure to support the further expansion of AI computing power.
