Wall Street Banks: AI Capex Super Cycle Drives M&A and Financing
Taylor Wilson
Goldman Sachs, Citi, Bank of America and JPMorgan all disclosed major AI-linked deals this earnings season — AI infrastructure spending is now delivering a concentrated surge in underwriting fees and lending income across Wall Street.
What exactly is Wall Street earning from AI?
Tech companies are pouring money into data centers and compute capacity. Every dollar spent requires bond issuances, equity offerings, or loans — and banks collect fees and interest on each one.
Goldman CEO David Solomon said on the earnings call that "AI infrastructure buildout is still in its early stages." This means → he sees years of spending ahead and a multi-year revenue runway for the bank.
He described the industry as being "in the middle of an AI capex super-cycle," with demand rising across all financing instruments. In plain terms = stocks, bonds, or loans — if it touches AI, it sells.
Which landmark deals have already landed?
Goldman won the lead underwriter role for SpaceX's record $86 billion IPO and will co-manage Anthropic's listing alongside Morgan Stanley. Rival OpenAI has also filed for a U.S. IPO.
Citi served as joint global coordinator for SK hynix's $26.5 billion ADR offering, earning more than $70 million from that single deal.
Bank of America extended a $520 million credit facility to OpenAI — the bank's first-ever loan to the AI company. This means → even traditional commercial banks are now opening direct credit lines for AI firms.
What does Bank of America's own data reveal?
Internal BofA figures show the bank has raised nearly $500 billion for AI-related companies since the start of 2025.
That sum accounts for 60% of its combined investment-grade debt, leveraged finance and equity-capital-markets activity. This means → six out of every ten dollars BofA intermediates are AI-linked — AI is no longer a "growth vertical" but the core book of business.
How does AI money reach the "plumbers"?
JPMorgan CFO Jeremy Barnum said the bank is financing data centers and seeing loan demand from companies only indirectly tied to AI.
His example: "Data centers drive demand for plumbers and electricians." In plain terms = building server farms requires pipes and wiring, so traditional trades are borrowing to expand too.
He cautioned, however, that it is hard to pinpoint whether such demand stems directly from AI. This reflects a real spillover effect whose boundaries remain blurry.
Reuters reported that Meta is working with Morgan Stanley and JPMorgan to arrange roughly $13 billion in financing for a data center in El Paso, Texas.
Can the super-cycle last?
Argus Research financial-services director Stephen Biggar said the "AI-driven capex super-cycle benefits equity issuance, M&A activity and debt financing."
Yet since July, tech stocks — chip stocks in particular — have come under pressure, with investors questioning high valuations and whether AI capex momentum can hold.
This means → Wall Street's current wave of AI mega-deals rests on one premise: tech companies keep spending aggressively. If they pull back, the middlemen's fees cool in lockstep. That makes spending durability the key variable for this cycle.
Content is for reference only, not financial advice.