Wall Street banks ride AI super cycle to record quarters as financing demands reshape markets

Wall Street Reaps Record Gains from AI Infrastructure Boom

American megabanks delivered stunning second-quarter results that signal AI financing has become a dominant force reshaping Wall Street. Goldman Sachs posted record quarterly revenue of $20.3 billion, up 39%, while JPMorgan Chase saw revenue rise 27% to $58 billion2

2

. The surge reflects what Goldman CEO David Solomon calls an AI super cycle, where "demands on financing into every single financing instrument, in every region of the world and across every single industry" are creating unprecedented opportunities3

3

.

Goldman Sachs booked $3.4 billion in investment banking fees during the quarter, a record, with equity underwriting jumping 130% to $985 million and debt underwriting rising 75% to $1.03 billion3

3

. JPMorgan reported $3.3 billion in investment banking fees, up 30% and its highest since 2021, while Morgan Stanley climbed 58% to $2.44 billion3

3

. Bank of America revealed it has helped raise nearly $500 billion for AI-related companies since 2025, accounting for 60% of such fundraising across investment-grade debt, leveraged finance and equity capital markets5

5

.

Morgan Stanley Pioneers Novel AI Debt Deals Structure

Morgan Stanley has emerged as the chief architect of AI debt deals, devising hybrid financial instruments that funnel tens of billions into data center financing4

4

. The bank's leveraged finance co-head William Graham designed a template deal for TeraWulf Inc., a Maryland-based company that pivoted from bitcoin mining to AI data centers. TeraWulf raised $3.2 billion through privately placed bonds at just 7.75% yield, despite being junk-rated at BB1

1

4

4

.

The structure combines elements of traditional bond deals with project finance protections, creating what industry insiders call construction bonds "wrapped" by Big Tech backstop agreements4

4

. Google guaranteed roughly half of the $3.7 billion in lease payments that startup Fluidstack pledged to TeraWulf over 10 years, while taking warrants for nearly a tenth of the company1

1

. TeraWulf CFO Patrick Fleury explained they "were able to borrow against the strength of Google's balance sheet," avoiding the slow stage-by-stage reviews of traditional project finance4

4

.

Since the TeraWulf deal, Morgan Stanley has sold more than $40 billion of construction bond products and is expanding the structure across Asia and Europe4

4

. Graham predicted that "AI infrastructure bonds are the majority of the annual supply of new-money non-investment-grade debt," calling it the fastest-growing segment and the first new market segment in 20 years4

4

.

Big Tech Balance Sheets Become AI Capital Markets Backbone

Hyperscalers including Google, Meta, Amazon and Microsoft have deployed their pristine balance sheets as emergency backstops, enabling unproven AI companies to access capital formation at dramatically lower rates1

1

4

4

. When a hyperscaler guarantees data center leases, financing costs roughly halve, according to Morgan Stanley's Mo Assomull, who noted the choice between "mid to high single digits or double that"4

4

.

Meta raised $27.3 billion in a private placement with Blue Owl and Pimco for a single Louisiana campus, part of more than $40 billion placed in that market since November3

3

. The social media giant is also working with Morgan Stanley and JPMorgan Chase on a roughly $13 billion financing package for a data center in El Paso, Texas5

5

. These arrangements benefit both parties: tech giants advance their AI infrastructure without massive upfront cash, while data center developers access institutional capital previously unavailable to them.

Google's involvement in TeraWulf extends beyond financial guarantees. The company's chip segment manufactures TPU chips that power AI computations, meaning TeraWulf's success directly benefits Google's hardware business1

1

. Major clients downstream include Anthropic, the frontier AI lab that Goldman Sachs and Morgan Stanley are preparing to take public5

5

.

AI-Related Capital Expenditure Forecast Spans Years

Morgan Stanley CEO Ted Pick quantified the AI infrastructure boom's trajectory, estimating the industry is "around 10%-15% of the way through the investment cycle"3

3

. His firm's research forecasts data center financing capital expenditure of roughly $850 billion this year, $1.3 trillion in 2027, and possibly $1.5 trillion in 20283

3

. Goldman's Solomon told analysts the bank is preparing for a three-to-five year investment cycle still in early stages2

2

.

JPMorgan CFO Jeremy Barnum described AI as "everywhere in financial markets," pointing to "booming environments with a ton of activity, big IPOs, big index rebalancing, a lot of activity in Asia"2

2

. Investment banks have profited from major deals including SK Hynix's $26.5 billion ADR offering and SpaceX's record $86 billion IPO5

5

. Citigroup CEO Jane Fraser noted that "wherever there's a bottleneck in that whole energy power compute memory ecosystem, we're seeing a lot of activity"3

3

.

Emerging Risks in Power Supply and Revenue Projections

Despite the enthusiasm, caution signals are emerging. JPMorgan's Barnum revealed the bank "passed on some deals," specifically citing concerns about power supply risk and tenant stability in data center financing3

3

. "We saw some deals come through where we were just like, 'Yeah, we're not doing that,'" he said, though no bank disclosed dollar figures for AI infrastructure exposure3

3

.

The financial engineering carries inherent risks if AI revenue generation disappoints. Analysts project TeraWulf's revenue will grow from about $300 million this year to 10 times that by 20291

1

. Hit those numbers and Google's backstop remains academic. Miss them badly, and while Google's $126 billion cash pile will barely be dented, its reputation will take a hit1

1

. Pick acknowledged the uncertainty: "It's really early, and I'm not sure we altogether know because of the known unknown element of this. One has to have humility in all of this"3

3

. JPMorgan's Jamie Dimon was even more direct: "It's getting close to as good as it gets. We just don't know how long it's going to last"3

3

.