# Bank credit data does not yet show a broad downturn

Markdown copy of https://realalpha.app/s/bank-credit-data-does-not-yet-show-a-broad-downturn/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

- Speaker: [Steve Eisman](https://realalpha.app/people/steve-eisman/) (Host)
- Show: The Real Eisman Playbook
- Episode: [Market Cracks Under AI Debt Burden | The Weekly Wrap](https://realalpha.app/v/y45yqiCdfI4/), Oct 16, 2026
- Moment: 12:47 to 17:52 of the episode (5 min), playing on https://realalpha.app/s/bank-credit-data-does-not-yet-show-a-broad-downturn/; the episode from there: https://realalpha.app/v/y45yqiCdfI4/?t=767

## Quote

> “We certainly saw large increases in consumer non-accruals leading up to the GFC. We are **not seeing that now** at all.”
> — [Steve Eisman](https://realalpha.app/people/steve-eisman/) at [14:54](https://realalpha.app/v/y45yqiCdfI4/?t=894) · The Real Eisman Playbook

## Takeaways

1. Non-accruing loans are a key signal of serious credit stress.
2. Large-bank consumer and commercial credit data remain benign.
3. Private credit and AI debt are harder-to-read economic risks.

## About this segment

Steve Eisman examines delinquent-loan data from four large banks and finds no broad deterioration. He argues that private credit and AI-related debt, not banks, may be where the next weakness appears.

From “Market Cracks Under AI Debt Burden | The Weekly Wrap”, Oct 16, 2026. [Watch the full episode](https://realalpha.app/v/y45yqiCdfI4/)

## Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[12:59](https://realalpha.app/v/y45yqiCdfI4/?t=779) Before we discuss valuation, I want to focus on how the large bank results are of particular importance because they provide great insight into the state of the current credit cycle. The four large banks, J.P. Morgan, Citi, Wells, and Bank of America, have large and broad-based lending businesses. I'd note that credit quality is a less important metric for Goldman Sachs and Morgan Stanley. On the consumer side, the big four provide credit cards, auto loans, and several other types of consumer loans. On the commercial side, they make commercial real estate and corporate loans and other types of commercial loans as well. They don't do everything, but they're big enough and broad enough to provide a window, though not a complete window, into the health of the U.S. economy. Now, there is growing fear that credit losses will start to mount for

[13:43](https://realalpha.app/v/y45yqiCdfI4/?t=823) the first time since the great financial crisis. We've had 17 years of amazingly benign credit quality, and investors are wondering if the good times are over. Yes, private credit is exposed to software, and software is getting hit by AI, but are there signs of broad credit problems emerging in the actual data? When the major banks report, they provide reams of data. J.P. Morgan's quarterly earnings release supplement is actually 29 pages long and is filled with data on every page.

[14:13](https://realalpha.app/v/y45yqiCdfI4/?t=853) that is typical of how all the banks report. To see trends and credit quality, the best place to look is to examine consumer and commercial non-accruing loans. Now, a non-accruing loan is a loan that is seriously delinquent, usually 90 plus days. A bank no longer reports interest income from these loans and reserves for future losses. I am putting on the screen consumer and commercial non-accruing loan data for JP Morgan, Wells Fargo, Citi, and Bank of America for the third quarter of last year, the second quarter of this year, and this third quarter for this year.

[14:50](https://realalpha.app/v/y45yqiCdfI4/?t=890) These four banks combined provide a large window into credit quality trends. If, if there was a serious trend of deteriorating credit, we would see increases in non-accruing consumer or commercial loans or both on a year-over-year basis and on a quarterly sequential basis. We certainly saw large increases in consumer non-accruals leading up to the GFC. We are not seeing that now at all. The credit data is benign in both consumer and commercial. Now, for those of you who are audio listeners, I'll just focus on J.P. Morgan and Bank of America. For J.P. Morgan, total non-accruing loans were $9.4 billion in the third quarter of this year. They were down 7% versus last year and flat versus the second quarter. The numbers of Bank of America were also very benign. Total non-accruals were $5.7 billion, up 3% versus last year and down 4% versus the second quarter.

[15:47](https://realalpha.app/v/y45yqiCdfI4/?t=947) Conclusion, the credit trends report. ordered by the large banks are very benign. Now maybe a credit cycle will emerge, but we are not seeing it in the banks. That's an important data point. The banks have the broadest credit exposure so their results are a concurrent indicator. And this brings me to my conclusion about the state of the U.S. economy. Since the GSC, many commentators have tried to predict the next crisis. This is it, they say, and then it isn't. Perhaps the most important story of the past 17 years is how powerful and resilient the U.S. economy has become. Yes, there are definite problems in private credit with respect to overexposure to software, and perhaps these software issues are harbingers of an approaching credit cycle. Yet, the banks have shown once again that credit quality in the U.S. is still okay. Whatever bad happens in the future, I don't think

[16:38](https://realalpha.app/v/y45yqiCdfI4/?t=998) it will be in the banking system. It could emerge in private credit, or it could emerge in AI-related debt like CoreWeave. That's why I'm now much more focused on the fixed income markets, continued reaction to CoreWeave, and whether there is contagion to other parts of AI. Again, in my view, the banks are not leading indicators this time. In the past, I have said, as the banks go, so goes the economy. Not this time. The banks are extremely well capitalized and their credit quality is still benign. The potential issues in AI and in private credit it will determine whether and when there will be a recession. It would be much easier to predict the future if the banks were the leading indicators because the data the banks produce is voluminous and transparent, while private credit is much more opaque, making future economic

[17:30](https://realalpha.app/v/y45yqiCdfI4/?t=1050) weakness much trickier to foresee. Frankly, I'm starting to think that the entire future of the U.S. economy hinges on the success or failure of AI. That's why I talk about it so much.