JPMorgan Chase (JPM), Bank of America (BAC), Citigroup (Citi), Wells Fargo (WFC), and Goldman Sachs (GS), the five major U.S. banks, disclosed their second-quarter earnings on July 14. Goldman Sachs and JPMorgan Chase exceeded market expectations, with Goldman Sachs' stock price hitting an intraday record high, while bank stocks showed divergent performances.
This set of earnings reports was magnified because a new pricing issue has emerged: Has the AI investment frenzy begun to fuel Wall Street? In the past, investors looking at bank stocks mainly focused on net interest margins, credit cycles, and capital returns. Now the market is starting to ask whether banks can become the "toll booths" for AI capital expenditures.
This toll booth is not mysterious. Technology giants and corporations invest in AI by building data centers, purchasing chips, financing, going public, and engaging in mergers and acquisitions. Hedge funds engage in AI stock trading and leverage, and banks earn underwriting fees, loan arrangement fees, trading spreads, and financing interest from these activities. AI will not be directly recorded on the bank's balance sheet but will enter the income statement.
The differences are also evident. Goldman Sachs CEO David Solomon emphasized that the trading pipeline, client network, and business flywheel still have momentum, and AI development is still in its early stages. JPMorgan Chase CEO Jamie Dimon is more cautious, and the market is also concerned that companies will reassess costs, prices, and investment returns.
Wall Street Takes a Cut of AI Capital Flow
Ordinary investors may easily understand the AI dividend as chip, cloud vendor, and data center rental income, but banks make money from another layer. As long as AI-related assets become more expensive, more active, and require more financing, banks can take a cut from trades and capital flows.
Stock trading revenue is the most direct example. It is not the bank's own bet on the rise and fall of stocks, but rather providing trade execution, hedging, financing, and liquidity for institutional clients. The greater the volatility of AI concept stocks, the more frequent the fund rebalancing, and the busier the bank's trading desk.
Goldman Sachs' second-quarter stock trading revenue reached $7.42 billion, setting a record and becoming the market's reassessment trigger for investment bank stocks. For companies like Goldman Sachs with a higher proportion of capital markets business, trading activity will quickly reflect in profit elasticity.
Another link is AI capital expenditure. Tech giants such as Microsoft, Google, Meta, etc., building AI infrastructure will drive data centers, electricity, chips, and private asset financing. Reuters reported that Wall Street banks are seeing the AI supercycle as one of the sources of future trading and financing activities.
Banks are not the most prominent part of the AI industry chain but are easily overlooked second-order beneficiaries. The more capital-intensive AI development is, the more it needs financial intermediaries. The more AI assets are traded frequently, the more beneficial it is to the trading department.
Goldman Sachs and JPMorgan Chase Provide Two Examples
The second quarter was most illustrative of Goldman Sachs' resilience. The company's net revenue was $20.34 billion, net profit was $6.63 billion, and EPS was $20.98. In the official financial report, Solomon mentioned "One Goldman Sachs" and the business flywheel, emphasizing that customer activity and trading channels still have momentum.
The market rewards not just the quarterly profit itself but Goldman Sachs' position in the AI narrative. If the continuous AI development over the next few years continues to bring IPOs, M&A, equity financing, data center loans, and trading volume, Goldman Sachs' profit elasticity will be more prominent than that of traditional retail banks.
JPMorgan Chase provides a more comprehensive sample. The company reported revenue of $57.3 billion in the second quarter, managed revenue of $58 billion. Reported net income was $21.2 billion, with net profit of $16.9 billion after excluding significant items. It has both capital markets business and consumer finance and corporate lending.
Restraint is needed here in attribution. The strong performance of large banks in the second quarter is not solely due to AI. Investment bank business recovery, active stock market trading, changes in the interest rate environment, and hedging demand due to macro uncertainty are all boosting revenue. AI's more accurate role is to increase a high-elasticity source of demand during the trading and investment banking cycle recovery.
Morgan Stanley will also be under observation. It disclosed its financial report on July 15, later than the other five major banks, but its business structure is closer to Goldman Sachs. If the market continues to trade in capital market activities brought by AI, it will be grouped with Goldman Sachs for valuation comparison.
Solomon's Flywheel and Dimon's Brake
Solomon's narrative is clear: a global client network, strategic transaction pipelines, and "Integrated Goldman" are forming a flywheel, with AI construction still in its early stages, and room for capital markets activity. As long as the AI investment cycle continues, client financing and trading demand will keep flowing back to Wall Street.
This logic is crucial for bank stock valuation. Traditional bank valuations are easily influenced by net interest margin, allowances, and regulatory capital constraints. If investment banking and trading revenues are seen as having structural increments, the market may assign a higher weight to Goldman Sachs, Morgan Stanley, and JPMorgan's capital markets business.
Dimon's restraint forms another boundary. He does not deny the value of AI, and JPMorgan itself is investing in AI and discussing efficiency improvements. But market concerns about corporate AI spending are rising: budgets will not expand infinitely, and clients will start asking how much revenue growth or cost savings each dollar invested can bring.
This is very real for banks. AI capital expenditures can bring trading and financing heat, but it does not necessarily follow a linear path. If companies slow down data center construction, postpone IPO financing, or see rising costs in private asset financing, both bank-related expenses and trading revenues could cool down.
Loan Quality and Trading Revenue Determine Revaluation
Whether bank stocks can truly use AI as a new valuation anchor will depend on whether AI-related financing can continue to enter investment banking revenue, whether data center loan quality can remain stable, and whether stock trading revenue will mean-revert after volatility.
Data center financing is particularly worth watching. It is now the fuel for AI infrastructure expansion and a source of costs for banks and private credit. However, if leases, utilization rates, electricity costs, or financing costs deviate, such assets may shift from revenue increments to risk exposure.
The traditional credit cycle has not disappeared. High borrowing costs, energy prices, and geopolitical risks will lag in credit card, auto loan, and corporate credit quality. Goldman is more likely to enjoy trading heat, while banks more reliant on traditional lending, like Citigroup and Wells Fargo, will be differentiated and priced by the market.
The signal from the second-quarter earnings reports is that the AI dividend has spread from tech stocks to financial intermediaries but is still in the "incremental amplifier" phase. What can support revaluation is not just an AI slogan but the continuous realization of investment banking expenses, trading revenue, data center loan quality, and corporate AI investment returns.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
