Featured Analysis: JP Morgan Chase & Co. (JPM) Q2 FY2026 earnings call transcript

Featured Analysis:
Is AI a Successful Story for JPM in Operational Leverage?

 

by Connie Chun
July 16, 2026

LLM AI is prevalent. Companies are searching for success stories with proven playbook to avoid missteps. To date, there is a lot of doubt regarding the ROI of LLMs. Judging on the blockbuster results from JP Morgan Chase recent earnings report, and its view of AI as its most vital transformative technology, such worry may not be true. It sparks my curiosity to analyze JP Morgan Chase & Co. (JPM) Q2 FY2026 earnings call transcript to learn more about their AI ROI.

With less concern about their earnings beat, my interest is to get a qualitative assessment of their near-term upside and downside risk.

I submitted JPM's earnings call transcript to ELAINE for analysis. ELAINE uses Abstractive Symbolic Logic Framework to create a logical semantic structure. Through this structure, it explores what are the concepts that dominate the context and semantics and less on the numbers. As opposed to an analysis that only provides a summary without details, ELAINE offers a qualitative assessment of the near-term upside and downside risk, sentiment readings, arguments that support conclusions and shows the details of each assessment.

ELAINE abstracted the following key focus:

"If we look at year-to-date results, it's been a strong revenue environment, but I think operating leverage on an adjusted basis was negative. You alluded to some expense one-offs potentially"

Looking at the list of abstracted semantic hierarchies, majority of the semantic hierarchies shows a focus on "question" with
(question, leverage, year-to-date) rank the highest on the list.

Follow the hyperlink that connects this entry for more details shows the following excerpt (quoted in italics):

"If we look at year-to-date results, it's been a strong revenue environment, but I think operating leverage on an adjusted basis was negative ..."

The above excerpt is the same as the key focus highlighted by ELAINE. It is a question raised by an analyst regarding JPM's operating leverage with AI. A negative operating leverage on an adjusted basis implies AI is operating at a net loss. In this scenario, revenue increases paradoxically shrink operating income, and vice-versa. Higher AI usage directly results in a higher number of tokens. As productivity increase linearly, token cost goes up exponentially. It is a generally known fact and is affecting AI adoption across industries.

The "read more" hyperlink associated with this excerpt shows the exchange between the analyst and JPM's Chairman and CEO Jamie Dimon (quoted in italics):

James Mitchell, an analyst, asked during the JPM's earnings call:

"Okay, good morning. Jeremy, maybe a follow-up on the expense question and operating leverage question earlier. Understand completely longer term, no bank can generate perpetual operating leverage. If we look at year-to-date results, it's been a strong revenue environment, but I think operating leverage on an adjusted basis was negative. You alluded to some expense one-offs potentially. No question you're investing heavily and should be, so I get all that. Just when I think about the benefits of AI and technology generally, is there a time over the intermediate term where you think expense growth could slow a little and operating leverage kind of becomes more likely in a period of time over the next few years"

Jamie Dimon, Chairman and CEO responded:

"I'm just going to answer that by saying when you have great returns and very good margins, which actually went up this quarter, not down."

"The notion that somehow you can forever increase your operating leverage is a crazy notion. We don't have that. I think it's part of the reason why banks failed, if you go back 20 years ago. We're never going to have that point of view. AI will have its gives and takes. We can't project. I do think you might actually see a slowdown in growth, maybe a slowdown in 2027 or 2028. The teams are looking at all of our opportunities, and we pointed out over and over again when we have an opportunity to spend more money in marketing, with deposit ROI, we're going to do it. We're not going to have false gods. We have to pray that we can't do something really smart."

"I've also pointed out over and continuously that some expenses, if you accounted for them as investments, that they have very good returns, but they're expensive in the short run. AI still remains to be seen because the other thing I think about AI, which is a little bit different than everybody else, is you don't uniquely benefit from AI. The ultimate beneficiary of AI will be our customers. In a competitive capitalist world, we always use AI to do a better job for the customers, and we can't just say, "Oh, it's going to increase our margins and we're going to keep that." If that were true, our margins would be 80% today because of computerization over the last 20 years."

What's the take-away?

Jamie Dimon acknowledged the cost issue:

"AI will have its gives and takes. We can't project. I do think you might actually see a slowdown in growth, maybe a slowdown in 2027 or 2028."

"but they're expensive in the short run. AI still remains to be seen because the other thing I think about AI, which is a little bit different than everybody else, is you don't uniquely benefit from AI"

These remarks support that cost justification for AI is not in productivity but improvement in customer experience. He acknowledged that AI is expensive. LLMs charge per interaction. Token-based costs can easily spiral. This can produce a negative effect on operational leverage. AI is not a fix cost like investment into computer hardware, it is a variable cost that ties to transactions. That being said, AI can potentially become a forever drag to a business's operational leverage. I wonder how many companies can afford to spend billions for a small return on AI buildout?

View ELAINE Analysis:

JP Morgan Chase & Co. (JPM) Q2 FY2026 earnings call transcript

More discussions on use cases:


Analysis of Fed. Chair Powell's Dec 10 2025 Press Conference Transcript Is there an elephant in the room?


Featured Analysis: Nvidia (NVDA) Q3 2026 Earnings Transcript


Featured Analysis: Discover Insights from Home Depot Latest Earnings Transcript