The new U.S. earnings season begins—can the AI-driven bull market deliver on profits?

The setup: prices up, confirmation pending

Valuation is a forward exercise, but it runs on reported accruals. The S&P 500's trailing multiple has expanded on the assumption that AI-related capital expenditure will convert into operating leverage — higher margins, lower unit costs, durable topline. The bull thesis is not wrong in theory; it is simply unverified. Q2 reports from the largest hyperscalers and semiconductor platforms are the first credible test. If operating income fails to outrun the depreciation and stock-based compensation embedded in those capex buildouts, the multiple does the math itself. The market's hesitation captured in the "earnings boom meets market caution" framing is not pessimism. It is the discount function repricing risk in real time.

What to watch in the filings

Bitcoin World's preview highlights the two sectors where the signal will arrive first: bank profits and consumer health. These matter more than the megacap tech prints for one reason — they test the real economy underneath the AI narrative. Net interest margin trends, commercial loan loss provisions, and credit card delinquency rates will reveal whether the consumer that funded corporate earnings in 2024 and 2025 is still solvent. On the bank side, the forensic question is simple: are provisioning ratios moving with charge-offs, or are reserves being released to flatter earnings? Released reserves inflate reported income without improving cash conversion. That is a one-quarter benefit, not a structural one.

For related context, see QuantRate Launches Free AI Trading Platform for Stocks, Crypto, and Forex.

The structural test for AI capex

The harder analytical work sits in the technology segment. AI infrastructure spending shows up across three accounts: capitalized internal-use software, property and equipment, and acquired intangibles. Each carries a different useful life and a different impairment profile. The risk for investors is not that capex is too high. It is that the amortization periods are too long, smoothing what is actually a shorter commercial window. When Q2 filings land, the disclosures worth isolating are: the useful-life assumptions on newly capitalized AI assets, any segment-level impairment charges, and the gap between reported operating income and operating cash flow. A widening gap is the clearest evidence that earnings quality is deteriorating — accruals rising faster than cash, the classic precursor to a downward revision.

The earnings season ahead is not a referendum on AI. It is an audit of how that story has been capitalized.