JPM Misses, Cuts the Buyback to $1.1B From $6.7B, and Guides 2027 NII ex-Markets to $107B: Management's Own Capital Allocation Now Agrees With Our Hold

Results vs. Consensus

Consensus was an operating estimate, so the comparison that matters is the $5.75 figure excluding the $658M of net investment securities losses, not the $5.57 GAAP figure. Measured either way, this was a miss. Measured correctly, it was a small one, and its composition matters more than its size.

Q3 2026 Scorecard

Year-over-Year

Quarter-over-Quarter

Q2 2026 CET1 and RWA are shown as finalized in the current supplement (14.2% and $2.13T); the July release estimated 14.1% and $2.14T. Q2 ex-items figures exclude the $4.6B Visa gain and $1.0B of equity-investment gains.

Revenue assessment. Ex-securities-loss revenue growth of 9% is a solid quarter for a $200B-a-year franchise. It was simply not the quarter the Street modeled after the second quarter's equities windfall. The composition is healthier than the second quarter's: NII ex-Markets grew 6% year over year on balance growth and a yield that finally stopped compressing, and fee lines tied to clients rather than volatility (asset management fees +15%, card income +25%, lending- and deposit-related fees +10%) did the work. Markets at $10.3B is the more interesting number. It is down 15% from a peak quarter and still up 15% year over year. That is normalization to a higher base, not mean reversion to the old one.

Margin assessment. This is the weakest element of the print and the one we would watch. In the second quarter, a 77% incremental margin on revenue outperformance made the expense raise read as good news. This quarter showed the other side of operating design: when revenue steps down, compensation follows, but the 17% non-compensation growth does not. The FY26 expense guide now implies ~$109.0B, up another $1.5B, and the CFO said the Street's ~$114B for 2027 is "not too far" from his preliminary read. That is ~5% expense growth into a year where NII ex-Markets is guided up ~10%, which is fine. It does assume Markets and IB hold.

EPS assessment. $5.75 against ~$5.90 is a modest miss of moderate quality. The reserve build is the kind a growing wholesale book should produce, and the securities loss is a choice the firm made, not a mark it suffered. But the clean read on earnings power is now ~$5.75 to $6.14 a quarter, not the top of that range. Our FY26 ex-items framework of $23–24 holds, at the lower half: $17.83 through nine months plus a fourth quarter near $5.5.

Segment Performance

Commercial & Investment Bank (CIB): Normalizing to a Higher Base

CIB earned $8.2B on $22.7B of revenue, up 20% and 14% respectively, at an 18% ROE. The year-to-date figures, which include the second quarter's windfall, show the scale of the step-up: nine-month CIB net income of $27.0B is up 32%, Markets revenue of $34.0B is up 23%, and equities alone are up 49% to $15.5B. The third quarter on its own still grew CIB net income 20% and equities 49%. What changed sequentially is that the two cyclical engines both slowed at once. Markets fell 15% to $10.3B, and IB fees fell 8% to $3.0B as debt and equity underwriting both declined. Advisory went the other way: up 21% sequentially and 32% year over year, to $1.2B.

“In the CIB, revenue increased 14%. Markets revenue and IB fees each rose at a mid-teens rate, helped by strong levels of client activity, although activity did moderate somewhat during the quarter. Meanwhile, Payments and Securities Services each posted double-digit revenue increases led by ongoing deposit and fee growth.”
— Jamie Dimon, Chairman and CEO

Two balance-sheet details matter more than the revenue lines. First, CIB loans grew 14% year over year to $675B, with Global Corporate and Investment Banking loans up 27%. This is where the RWA is going and where the reserve build came from. Second, average CIB trading VaR rose to $39M from $25M a year ago (+56%). The revenue is being earned with more risk on, which is reasonable in an active tape and worth remembering when the tape turns.

Assessment: The franchise did what Bull #1 says it should: it held a Markets base above $10B in a quarter where Fixed Income declined, IB slowed in late September, and bank stocks sold off. That is one data point toward the trigger we set in July for turning constructive, and only one. The miss came from Fixed Income, where consensus sat near the second-quarter level. That is the specific form of Bear #5 we flagged: the peak anchoring expectations. The segment is fine. The expectations were not.

Consumer & Community Banking (CCB): The Best Quarter in the Firm

CCB grew net income 13% on 9% revenue growth, helped by a 20% lower provision. Expense still grew faster than revenue, at 12%, on marketing, technology and advisor hiring. Card was the engine. Card Services & Auto revenue rose 15%, average card loans rose 7% to $250B, the card net yield reached 10.67%, and card income in CCB rose 50%. Credit was better again: the Card Services net charge-off rate fell to 2.91% from 3.34% in Q2 and 3.15% a year ago, and 30+ day delinquencies of 1.97% are 17bps below last year. The FY26 Card NCO guide of ~3.2% stands, which implies a seasonal uptick in the fourth quarter.

The deposit franchise is the part we care about most this quarter, given the September bank-stock sell-off on agentic-AI deposit fears. On the evidence so far, it is fine. Average CCB deposits rose 4% year over year to $1.10T, the deposit margin widened to 2.78% from 2.70% in Q2, and account acquisition accelerated.

“We ranked #1 in U.S. retail deposits for the sixth consecutive year, and account acquisition remained strong in the quarter as we added over 650,000 net new checking accounts.”
— Jamie Dimon, Chairman and CEO

Elsewhere, client investment assets passed $1.4T (+15%), debit and credit card sales volume rose 10% to $539.5B, and active mobile customers rose 7% to 64.9M. Home Lending is the one segment line the rate move visibly damaged: production revenue collapsed to $46M from $173M a year ago.

Assessment: This is the cleanest quarter in the firm, and it lands in the business that most worried us in July, when CCB was handed to a markets executive. The numbers do not yet show a leadership problem. They show a deposit franchise that is growing, repricing favorably and acquiring accounts faster, a card book with better credit than guided, and a reserve that held flat while card loans grew 7%. Card allowance coverage drifted down to 6.16% from 6.61% a year ago, which is fine while delinquencies fall and is the line to watch if they turn.

Asset & Wealth Management (AWM): Flows Up, Margin Down

AWM revenue rose 14% to $6.9B and net income 11% to $1.8B, at a 36% pretax margin and a 45% ROE. AUM reached $5.2T (+13%) and client assets $7.9T (+15%). Long-term net inflows were $51B, led by fixed income, and total client asset net inflows were $263B, the largest of the five quarters in the supplement. Markets subtracted $23B from AUM in the quarter, so the growth was all flows.

“Finally, in AWM, revenue grew 14%, and flows were very strong with $263 billion of client asset net inflows.”
— Jamie Dimon, Chairman and CEO

The cost line is the caveat. The pretax margin slipped to 36% from 38% in Q2, flat year over year, as expense rose 4% sequentially against 1% revenue growth and the provision rose to $84M from $13M. Year over year, expense rose 14%, in line with revenue, on revenue-related compensation, distribution fees and continued hiring: Global Private Bank client advisors rose to 4,460, up 8% sequentially and 10% year over year. Average AWM loans grew 20%, while average deposits fell 3% sequentially.

Assessment: Bull #4 continues to confirm on flows, which is the part that compounds. The margin dip is an investment in advisor capacity rather than a pricing problem, and a 45% ROE leaves plenty of room. AWM remains the steadiest growth line in the firm and still too small, at ~14% of revenue, to set the multiple.

Corporate: The Discretionary Loss and the Rate Drag

Corporate lost $455M on $93M of revenue. The $658M of net investment securities losses explains the swing. The release describes them as discretionary, and management did not explain the trade on the call. Underneath, Corporate NII fell $737M year over year to $669M on lower rates earned on the firm's excess cash and securities, and Treasury and CIO net income was $161M against $1.2B a year ago. Corporate deposits more than doubled to $79B, attributed in the supplement to the firm's international consumer initiatives.

Assessment: A discretionary sale of available-for-sale securities at a loss, in a quarter when yields rose sharply, is most often a repositioning into higher-yielding assets. If so, it pays back through NII over time, and part of the higher Q4 and 2027 NII guide may reflect it. That is our inference, not a management statement, and we treat the loss as non-recurring only to the extent such sales do not become a pattern.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Management was calm and unapologetic about the miss, treating it as a normal quarter after an abnormal one, and spent most of the call on capital deployment, rates and agentic AI rather than on the quarter's numbers. The posture on the outlook was more guarded than in July: more caveats on the 2027 NII figure, more talk of uncertainty, an explicit late-September slowdown in banking activity, and a CEO who still sees a higher-than-consensus chance of much higher rates. Succession, the dominant thread in July, did not come up at all.

1. The Buyback Stops: The Capital Hierarchy Made Visible

Gross repurchases fell to $1.06B (3.1M shares at an average of $342.33) from $6.7B in the second quarter and $8.3B a year ago. Net of employee issuance, repurchases were $664M, and the LTM net payout fell to 61% from 73%. In July we wrote that "buy back less as the price goes up" and $6.2B of net repurchases at 3x tangible book were difficult to hold at the same time, and we listed the crowd-out of buybacks by organic deployment as a commitment to watch. It arrived in one quarter, and the release led with it:

“Our capital hierarchy has remained consistent over time, with investing in the franchise and supporting our clients at the top. In line with our hierarchy, we reduced share repurchases this quarter. We believe we can deploy this capital over time at very attractive returns for our shareholders.”
— Jamie Dimon, Chairman and CEO

On the call, Dimon put excess capital at “$30 billion or something,” down from the ~$40B cited in April and July, and pointed to the Security and Resiliency Initiative, a $1.5T ten-year financing commitment whose first year he estimated at over $200B, as one destination. The CFO framed the choice as preserving optionality: given how quickly the firm has been able to deploy organically, “it makes sense to keep dry powder.”

Assessment: This is the correct decision, and it is the most important signal in the print. Capital that earns 17% or more organically should not buy a stock on a ~6–7% earnings yield, and management said so in those terms. But read it from the other side as well. The firm with the best information on JPM's intrinsic value just cut its purchases by more than four-fifths at ~$340. The buyback also supported per-share growth: diluted shares fell 3% over the past year. Without it, EPS growth runs closer to net income growth. Bull #3 is now being defended rather than spent on stock.

2. The Miss, Decomposed: Markets, a Reserve Build for Growth, and a Discretionary Loss

The miss has three identifiable sources, and none of them is credit deterioration. Fixed Income Markets revenue and IB fees came in below consensus. A $539M net reserve build followed loan growth, with $452M in CIB. And the firm chose to realize $658M of securities losses. The CFO was explicit on the reserve:

“Sure. Yeah, the wholesale build is actually just from new activity. So it is not downgrades or change in the Outlook or anything like that.”
— Jeremy Barnum, CFO

Firmwide net charge-offs fell $389M year over year to $2.2B, wholesale nonaccrual loans fell 13%, and the CIB net charge-off rate was 0.14%. The release frames the headline result “excluding discretionary investment securities losses totaling $498 million after tax,” and the word "discretionary" is doing deliberate work.

Assessment: A miss built from Markets normalization, a growth-driven reserve and an elective securities loss is a much better miss than one built from charge-offs or deposit costs. It is still a miss, and it resets the clean quarterly earnings base toward $5.75 from the $6.14 of the second quarter. We would not chase the stock on "the miss was optical."

3. NII Raised Again, and a First Look at 2027

The fourth-quarter guide implies another increase to the full-year numbers and introduces the 2027 central case:

“We expect fourth quarter NII ex-markets to be approximately $25.5 billion and fourth quarter total NII to be about $28 billion. And as we have done for a couple of years, you will note we gave you a central case for 2027 NII ex-markets of approximately $107 billion, which we will refine further next quarter.”
— Jeremy Barnum, CFO

Adding the guide to nine-month actuals gives FY26 NII ex-Markets of ~$97.2B (raised from ~$96.5B) and total NII of ~$106.2B (from ~$105.5B). The second half ex-Markets now runs ~$50.2B, above the ~$49.5B the July guide implied. The fourth-quarter run-rate of $25.5B annualizes to ~$102B, so the 2027 central case embeds about $5B of growth, roughly 10% above FY26. The net yield ex-Markets rose to 3.70% from 3.65%, the first sequential expansion since the fourth quarter of 2025. Dimon added the caveat himself: “I think there is more uncertainty around that number than normal.”

Assessment: Bull #2 strengthened this quarter, and the quality concern we raised in July, a volume story with a falling yield, eased as the yield turned up. The 2027 number is the strongest single argument for the bulls: ~10% NII growth on a $100B base is a large earnings tailwind. We take most of it. We haircut the rate-curve piece, discussed next.

4. The Rate Spike Arrives Through the Balance Sheet

In July, after the CFO conceded that deposit convexity "at some point kicks in," we reframed Bear #3 from a rate-cut risk to a rate-spike risk. The rate spike came. The Fed raised rates in September for the first time since 2023, and long yields moved above 5%. The first-order effect on JPM was not on deposits. It was on capital:

“In terms of the balance sheet, we ended the quarter with a standardized CET1 ratio of 13.9%, down 30 basis points versus the prior quarter, as net income was more than offset by OCI losses, higher RWA and capital distributions.”
— Jeremy Barnum, CFO

The CET1 rollforward shows $7.9B of unrealized losses on investment securities against $15.0B of net income to common. AOCI widened to ($19.8B) from ($7.7B), and tangible book per share fell 0.6% sequentially, the first decline in the five quarters the supplement shows. Meanwhile the firm became less asset sensitive: earnings-at-risk to a +100bps shift fell to $1.4B from $1.8B. The 2027 NII bridge also leans on rates, with about a quarter of the ~$5B growth tied to the hikes now in the forward curve. Asked about a 6% ten-year and further Fed hikes, Dimon reaffirmed his outlier view:

“You know, I have always thought there is a higher chance than other people. I still kind of think that, but that is not a prediction.”
— Jamie Dimon, Chairman and CEO

Assessment: Higher rates are locally positive for JPM's NII and negative for its tangible book, its capital ratio and, with a lag, its credit and deposit costs. This quarter delivered the NII benefit in the guide and the capital cost in the actuals. We move Bear #3 to EMERGING. The deposit franchise has not shown convexity yet, but the CFO does not think the cash sorting is over (see Q&A), and the CEO assigns above-consensus odds to the scenario where it would be.

5. Equities Holds, FICC Breaks: Markets Normalizes to a Higher Base

Markets revenue of $10.3B was up 15% year over year and down 15% sequentially, and the two halves diverged:

“Markets, fixed income was down 4% year on year, lower revenues in commodities, rates, and credit, largely offset by higher client flows in currencies and emerging markets. Equities was up 49% year on year, with continued outperformance across the franchise, notably in prime and derivative.”
— Jeremy Barnum, CFO

Equities at $4.95B was above the $4.3–4.8B quarterly range we carried after the CFO called the second quarter's $6.0B improbable to repeat. Fixed Income at $5.37B posted a year-over-year decline for the first time since early 2024. Asked whether client demand for balance sheet had softened, the CFO accepted an analyst's description of demand for capital as insatiable: “I like your insatiable for capital point. That feels about right.” On the fourth quarter, the CFO's base case is a normal seasonal decline. The CEO was less sure: “Think the fourth quarter will have not the same seasonal decline.”

Assessment: The equities franchise, built on prime financing and derivatives for an expanding hedge fund client base, is holding a much higher base than two years ago. That is Bull #1. Fixed Income is cyclical in the ordinary way, and a rate sell-off hurt rates and credit trading this quarter. Our $9.5–10.0B through-cycle Markets base survives a quarter that tested it, and we raise the equities range to $4.5–5.0B.

6. Investment Banking: The Accordion Closes in Late September

IB fees rose 14% year over year to $3.0B on advisory, and the firm kept the #1 global ranking at 9.2% wallet share. The forward signal was weaker, and the prepared remarks volunteered it:

“While our outlook remains constructive, supported by ongoing client engagement, evolving market conditions have recently slowed activity a bit, so we will have to see how the quarter develops.”
— Jeremy Barnum, CFO

In Q&A, the CFO dated the slowdown to the rate sell-off: “And so we did see a little bit of that slowdown in, you know, the second half of September, basically.” The CEO reached for his usual image of the pipeline as an accordion that “can open and close pretty quick,” and offered one offset: the European pipeline is “actually quite healthy and even more so in IPOs than we would have thought.”

Assessment: In July the CFO conceded some IB pull-forward and called the pipeline "quite robust." This quarter delivered part of the air pocket we listed as a commitment to watch: fees down 8% sequentially and underwriting down sharply, with advisory masking it. The pipeline has not broken, but the fourth quarter starts on a slower footing than the second quarter's guidance language implied.

7. Agentic AI and the Deposit Franchise: Smart Cash Goes Live, Outside Agents Are Blocked

The new item on this call was the one that had moved bank stocks in September: personal AI agents that could move idle consumer cash into higher-yielding products. Management's framing was that this is two old competitive forces combined, yield-seeking flows and fintech experience competition, and that it is not complacent about either. It also reported delivery on the commitment Dimon made in July (“You'll see something this year”):

“there is obviously been a lot of focus on smart cash solution, which I think is a small piece of the overall story. But that is, live now, actually, as of the last couple weeks. Very, very early days, and it remains a kind of targeted pilot type thing.”
— Jeremy Barnum, CFO

The defensive half is that outside agents will not get in. Dimon: “In the meantime, we are not gonna allow just any agent to come, outside agent to come inside the company because it is just too risky for clients and for JPMorgan.” The CFO tied the policy to the firm's long opposition to credential sharing and screen scraping, and placed liability on customers who hand credentials to an agent. On the related fear of a faster, technology-driven deposit run, he was dismissive of the narrative and blunt about the principle: “You are probably not supposed to be hoping that your deposits are stable because the money is difficult to move.”

Assessment: The strategy is coherent: build the agent (Smart Cash), block the competitors' agents, and win on primary-bank relationships rather than rate. The CFO's last line is the honest one. A franchise whose low-cost deposits partly depend on customer inertia has a real exposure to a technology that removes inertia. Blocking access buys time; it does not change the economics if customers and regulators push for open access. The test is measurable: CCB deposit growth and the deposit margin, both of which improved this quarter.

8. AI Economics: Benefits Not Yet Visible, Costs Being Managed

Asked where the AI benefits show up in the earnings release, the CFO conceded they do not yet, sized AI expense at “something like 700 annually or something like that,” and said he is deliberately stepping back from quantifying benefits: “I am really trying to move away from that.” His anecdote of a near one-shot rewrite of legacy finance tools suggested real productivity gains in software engineering. On cost, the firm's approach to token spend is to avoid lock-in:

“So part of the strategy is model agnosticism, harness agnosticism, and ability to run both proprietary models, but also open weight and open source models, some of which we have run on our own.”
— Jeremy Barnum, CFO

Dimon added that model routing will send each query to the best, quickest and cheapest model: “We do not need the most expensive thing for every query you have, and very often people just use the 1 they are comfortable with.”

Assessment: Consistent with July, and still mildly bearish for the multiple. Management will not promise an AI efficiency dividend, has stopped publishing the old benefit estimate, and expects some gains to be offset by token costs. A 2027 expense base near $114B, up ~5%, is the number that would have to fall for AI to become a margin story. Nothing on this call suggests it will.

9. Expenses: ~$109B for 2026, and a ~$114B Street Number for 2027 That the CFO Did Not Fight

The fourth-quarter expense guide implies another increase to the full year:

“On expense, we expect the fourth quarter to be about $27.5 billion with the implied full year increase reflecting higher volume and revenue related expense driven by continued elevated activity levels.”
— Jeremy Barnum, CFO

Nine-month expense of $81.5B plus ~$27.5B gives ~$109.0B for FY26, up from ~$107.5B in July and ~$105B at the start of the year. On 2027, the CFO broke from last year's pattern of correcting the Street:

“That seems to be quite a bit less true this year, so we see the consensus right now for 2027 expenses being around $114 billion and our current preliminary read is not too far from that.”
— Jeremy Barnum, CFO

Assessment: In July the raise came with a 77% incremental margin that justified it. This raise came with a quarter in which revenue fell sequentially and expense did not. Volume-related costs (brokerage, distribution fees) are legitimate, but non-compensation expense up 17% year over year is a structural base. The 2027 setup of ~10% NII ex-Markets growth against ~5% expense growth is still positive operating leverage, provided Markets and IB hold their current base.

10. Credit: Card Better Than Guide, Underwriting Tightened, Rates the Dependency

Consumer credit was good and wholesale credit was stable, with the reserve build tied to growth. The more forward-looking comments were about underwriting discipline and the rate path. Asked where the firm is pulling back, Dimon pointed at loosening terms:

“I am not gonna give you very specific stuff, but we are tightening up a little bit credit. I think some of the credit terms have got a little bit loose.”
— Jamie Dimon, Chairman and CEO

On the consumer, the CFO described current health as strong because of the labor market and drew the conditional line plainly: if the central bank needs an aggressive extended rate cycle, “that is gonna involve higher unemployment, which will produce weaker consumer credit results.” On middle-market clients, he described caution rather than distress: “You know, no question that higher rates, I think input costs, inflation, the diesel narrative, like, these things are tough.”

Assessment: Bear #4 stays CONTAINED on the data: card NCOs at 2.91%, delinquencies below last year, nonaccrual loans down 7%, and CIB charge-offs well under half of last year's. The dependency has sharpened, though. In July the risk was a labor market that weakens by itself. Now there is a policy channel too: a hiking Fed and $106 Brent bear directly on the middle-market and consumer borrowers the firm is growing into.

11. Top of the Cycle: "Every Single Business," AI Concentration, and Hedge Fund Leverage

Asked whether anything could pull ROTCE toward the 17% through-cycle target from 21%, management again described the environment as unusually broad-based in its favor:

“And we do not want to be pointlessly negative, and I will not sort of bore you with the speech. But if you just look across the entire franchise, every single business is performing exceptionally well.”
— Jamie Dimon, Chairman and CEO

The CFO added that historically a high-rate environment produced more internal diversification, with some businesses doing better than others: “So the conditions that we have had recently were essentially every part of the franchise is, you know, beating records or not typical by historical standards.” On the AI build-out, he said exposures are "still quite modest" relative to total wholesale exposure, but warned that the theme now runs through construction lending, working-capital revolvers and the equity markets, so the firm stresses the whole enterprise for a sentiment shift. On hedge fund leverage, highlighted by the IMF, management called it a known risk, and Dimon noted a couple of recent, modest fund unwinds that the Street absorbed easily.

Assessment: When every business is at a record at the same time, earnings are most correlated to a single factor, here an active, high-priced, risk-on market. That is the core of Bear #6. The AI-exposure answer was careful and correct, and the absence of a dollar figure is the gap. A bank at 3.0x tangible book has priced the "every single business" environment; it has not priced the correlated unwind.

Guidance & Outlook

The guidance pattern this quarter was favorable on NII, unfavorable on expense and unchanged on credit. The 2027 framework is the substantive addition. The CFO walked through the bridge from the fourth-quarter run-rate:

“That $5 billion, probably about a quarter of that, I would say, is related to the yield curve. So we remain asset sensitive and probably a little bit more asset sensitive than the EAR shows ... So, you know, about a quarter of that is the hikes that are currently in the forward curve. I think there is 1 in December and then 2 through the first half of the year. And the rest of it, I would actually describe pretty straightforwardly as just a continuation of the balance sheet growth and mix trends that we have seen this year.”
— Jeremy Barnum, CFO

Implied fourth-quarter math. The guide implies Q4 NII ex-Markets up ~3% sequentially and total NII up ~3%, with Markets NII near $2.5B. Expense of ~$27.5B is ~1% above the third quarter. If Markets follows the CFO's normal seasonal decline toward ~$9.0B and IB stays near the third-quarter level, fourth-quarter EPS lands near $5.5, giving FY26 ex-items EPS of ~$23.3.

Street position. The third-quarter consensus was set after the second quarter and leaned on its Markets level. Post-print, the NII guide pushes FY26 and 2027 NII estimates up, while lower Markets and IB run-rates and higher expense push the other way. The net effect is roughly neutral on FY26 EPS and positive on 2027 NII. Coverage after the print was mixed in tone, with at least one target raised and most commentary focused on the buyback.

Guidance style. Unchanged and credible. The firm guides to frameworks and refreshes them every quarter, in both directions. Expense guidance has risen repeatedly in 2026 and NII guidance in each of the last two quarters, and both moves track the revenue environment rather than sandbagging. The 2027 NII central case comes with the usual explicit caveat, plus an unusual one: Dimon said uncertainty around it is higher than normal.

July's Commitments, Graded

Analyst Q&A Highlights

Q&A ran on three threads: capital deployment after the buyback cut, the 2027 NII bridge, and agentic AI from several angles (deposit disintermediation, AI's own economics, and lending concentration to the AI build-out). Credit and rates got shorter but pointed questions. Unlike July, nobody asked about succession.

Why the buyback fell to roughly a seventh of its usual pace

The first capital question relayed investor inbound directly: why a ~$1B buyback against a typical ~$7B quarter, in the run-up to final Basel III endgame and G-SIB rules expected around year-end. The answer ruled out the defensive explanations by name and made the positive case for organic redeployment, with an explicit comparison to the stock's earnings yield.

Q: "We were fielding a lot of inbound questions from investors in terms of the $1 billion buyback versus the normal cadence of around $7 billion. And I know that there is no normal cadence, I know that buyback is the last priority."
— Erika Najarian, UBS

A: "So the positive news is the main reason the buyback is down is that ... nothing to do with AOCI or guessing about the environment, even Basel III. We believe now, which is different than we told you in the past, that we can redeploy that capital at very good returns over time. And redeploying that capital at very good terms over time is far better than buying stock at a 6% or whatever the earnings yield would be if you bought it today."
— Jamie Dimon, Chairman and CEO

Assessment: "Different than we told you in the past" is the admission. Two years of buybacks at rising multiples are being replaced by a deployment thesis. Disclaiming AOCI is fair as to motive, but the AOCI hit is a fact about this quarter's capital: $7.9B came out of CET1. The comparison to a ~6% earnings yield is management telling you how it values its own stock.

The bridge to ~$107B of 2027 NII ex-Markets

The question asked for the assumptions behind the new 2027 central case: balance growth, deposit pricing and rates. The answer was unusually clean: about a quarter from the forward curve, the rest from continuing this year's balance and mix trends, including card loan growth at a comparable pace. The CEO then added his own caveat on top of the CFO's.

Q: "I was hoping Jeremy, you could unpack the new outlook for 2027 NII, the central case that you gave, including some of the assumptions around the balance sheet growth, deposit pricing and the impact of higher rates?"
— John McDonald, Truist

A: "And the rest of it, I would actually describe pretty straightforwardly as just a continuation of the balance sheet growth and mix trends that we have seen this year. So both continued-- you know, modest consumer deposit growth, continued wholesale deposit growth, probably the wholesale a bit weaker than what was very strong this year, but we do not actually know. And then card loans, which obviously can be a big driver of NII, continued at comparable levels."
— Jeremy Barnum, CFO

Assessment: The non-rate three-quarters of the bridge is the credible part. It extends trends that the third quarter just confirmed. The rate quarter depends on three hikes the market may or may not deliver. We underwrite ~$106B, which keeps the growth story intact and leaves out half the curve.

Whether personal AI agents threaten the deposit franchise

The opening question of the call went straight to the theme that sold bank stocks in September. The answer reframed agentic competition as the intersection of two long-standing pressures, yield-seeking flows and fintech experience competition, and restated the strategy of competing for primary-bank relationships rather than every rate-sensitive dollar.

Q: "1, when we think about the disruption risk to agentic AI, I know you have talked about in the past, Maybe just frame that given your visibility into the technology, how you think about that, and also the challenge of implementing agentic AI inside a bank given the emergence of all the rogue agents, and does that make it harder for a bank?"
— Ebrahim Poonawala, Bank of America

A: "So if you think about what is the nature of the competition that comes from these agents, it is really the intersection of 2 things that we have been talking about for a long time. 1 is yield seeking flows, So obviously, we deal today and have dealt with all the various rate cycles. A lot of competition from yield-seeking flows. And our strategy remains focusing on primary bank share, not chasing every dollar of yield seeking money."
— Jeremy Barnum, CFO

Assessment: "Not new" is accurate about the forces and less accurate about their speed. An agent that sweeps idle balances automatically removes the friction that made the old forces slow. The firm's answer is a credible two-part defense, its own agent and closed access for outside ones. The evidence this quarter favors the defense. The argument will be decided by deposit data over several quarters, not by this exchange.

Where the AI benefits show up in the numbers

The question asked to see the AI payoff in the earnings release, and to update a previously cited figure of roughly $2B of spend and $2B of benefit. The answer corrected the spend figure downward, declined to re-quantify benefits, and conceded the core point without hedging.

Q: "Can you say the degree that AI helps out JPMorgan? ... But, you know, I squint and I look at the earnings release. I am trying to find the benefits, and I am not sure if I can."
— Mike Mayo, Wells Fargo Securities

A: "So on that point, and in terms of the visibility of the benefits, I think you are right to point out that the benefits are not observable in our current quarterly earnings. ... it takes a long time to really do things that are visible at an institution of our size. ... And, obviously, some of those benefits are gonna be offset with token cost, which is a whole other conversation."
— Jeremy Barnum, CFO

Assessment: The candor is the point. Management is resetting expectations away from a quantified AI dividend at the same time as it lets the Street's 2027 expense number stand near $114B. Investors pricing an AI-driven efficiency step-change into JPM are doing so without management's support.

Whether higher rates will restart the consumer cash sort

The question asked whether consumer deposit expectations had changed in a higher-rate, more competitive market. The answer kept the base case intact and then volunteered a personal view against the consensus that cash sorting is over, which is the most useful risk disclosure on the call.

Q: "And I am just wondering, any change in your expectations for either growth or how you are adjusting to the current marketplace and competition for deposits?"
— Ken Usdin, Autonomous Research

A: "So in the normal course, you will obviously have higher rate paid from the high beta products from CDs and premium deposits. ... And it is very much a function of which side of the argument you land on in terms of some people believe, well, all the cash sorting that gonna happen has already happened, and therefore, you know, these things are relatively insensitive to slightly higher rates. I do not personally believe that."
— Jeremy Barnum, CFO

Assessment: In July the same CFO said the deposit convexity his models expected had not arrived but must at some point. With the Fed now hiking, he expects "little bit of additional pressure," higher weighted-average rate paid, and some migration into CDs. That pressure is not in the 2027 NII central case in any visible way. It is the main reason we haircut it.

Concentration risk in financing the AI build-out

The question asked whether the hyperscaler capital-expenditure race can be financed without breaching concentration limits across bank loan books and bond and insurance portfolios. The answer separated single-name exposure, which is managed, from thematic exposure, which is harder, and stopped short of claiming immunity.

Q: "So I guess my question is this amazing hyperscaler race to expand compute power, and everybody wants to finance it. The money in capital's there, but from your vantage point, can we finance it without breaking some concentration limit issues?"
— Glenn Schorr, Evercore ISI

A: "And so you really need to be running stresses that test your resilience to a meaningful change in overall thinking across the entire enterprise. ... You know, currently, our exposures are still quite modest in the context of our overall wholesale exposure, but that is not the same as saying that we would be immune to a major change in sentiment in the AI space. I think I would be surprised if anyone, frankly, was immune to that."
— Jeremy Barnum, CFO

Assessment: This is the right answer and the right worry. Thematic exposure through revolvers, prime brokerage, equity underwriting and wealth client portfolios is larger than any construction-loan line. "Quite modest" without a number leaves investors to trust the stress process.

How far 21% ROTCE sits above the through-cycle level

The question asked whether anything in the macro or competitive environment could pull returns toward the 17% through-cycle target. The answer did not defend 21% as sustainable. It restated 17% as the through-cycle goal and named the cyclical inputs flattering the current return.

Q: "So JPMorgan's, ROTCE target through the cycle is 17% versus 21% today. Is there anything you are seeing either in the macro environment or competitive environment that could accelerate a path to 17%, or is that not anywhere on the horizon?"
— David Chiaverini, Jefferies

A: "I always talked about the 17% being through the cycle. Things are pretty good now with high volume, high asset prices, and fairly low credit losses and all that."
— Jamie Dimon, Chairman and CEO

Assessment: High volume, high asset prices and low credit losses are the three inputs most likely to mean-revert together. The stock at 3.0x tangible book capitalizes roughly a 20% sustainable ROTCE on our framework, and the CEO keeps describing 17% as the through-cycle number. That gap is our Bear #6, stated by management.

What They're NOT Saying

1. No explanation of the discretionary securities losses. The release calls the $658M of losses discretionary, and the call never explained what was sold, what the proceeds were reinvested in, or the earn-back period. If it was a repositioning into higher yields, it partly funds the higher NII guide, and investors should know how much. If more such sales are coming, the "significant item" framing becomes a recurring cost of a rate strategy.
2. No bridge from ~$40B to ~$30B of excess capital. Excess capital fell by about a quarter in one quarter, with the $7.9B AOCI hit and $51B of RWA growth as the obvious drivers. The CEO said the buyback cut had "nothing to do with AOCI," but the shrinking excess was never reconciled. With final Basel III endgame and G-SIB rules expected around year-end, the denominator may move again.
3. Nothing on succession. The first call after the co-president appointments to pass without a single succession question. The two-into-one mechanism we flagged in July remains unaddressed, and its absence from the agenda is not the same as resolution.
4. No dollar figure on AI-related credit exposure. The CFO called exposure to the AI build-out "quite modest" relative to total wholesale exposure, while describing a theme diffused through construction loans, revolvers and equity markets. A firm that stresses the whole enterprise for an AI sentiment shift can size the result, and chose not to.
5. No data on deposit migration or agent blocking. Management expects some migration into CDs and higher rate paid, and is blocking outside agents. It gave no figure on the share of deposits in high-beta products, the pace of migration since the September hike, or Smart Cash adoption. These are the leading indicators for Bear #3, and none was disclosed.
6. No reconciliation of the two fourth-quarter Markets views. The CFO's base case is a normal seasonal decline; the CEO said the decline may not be the usual size. Neither offered a number, and the Street is left to pick a side for a line that was the main source of this quarter's miss.

Market Reaction

• Pre-print setup: Close of $334.30 on October 12. The stock entered the print +3.7% YTD against the S&P 500's +14.0%, +8.5% over twelve months and −6.2% over thirty days. The 52-week closing range was $282.84 to $365.18. The high was set on August 12, so the stock had given back 8.5% from it, in a September bank sell-off tied to rising rates and agentic-AI deposit fears. Most large peers fell further over the same thirty days.
• Print-day session (before the open, October 13): Premarket trading was weak on the headline miss, but the regular session opened at $339.97, a 1.7% gap up, and reached $344.36 (+3.0%). The stock then faded through the day to close at $331.43, down 0.9% (−$2.87), just above the session low of $331.17. The high-to-low range was 3.9% of the prior close.
• Volume: 13.8M shares versus a 7.6M 30-day average (1.8x), heavier than July's 1.4x.
• Benchmark and peers: The S&P 500 was flat and a bank index ETF fell 0.7%. Among peers reporting the same morning, Citigroup fell 5.6% and Wells Fargo 3.1%, while Goldman Sachs slipped 0.2%. The next day, Bank of America fell 4.8% and Morgan Stanley 2.7%.
• Following sessions: On October 14, JPM opened at $327.17 and recovered to close at $335.28 (+1.2%) while the S&P 500 fell 1.1%. By October 16 it closed at $339.72, up 1.6% for the print week against the S&P 500's 1.9% decline.

The session shape was the mirror image of July's. In July the stock gapped down on the expense raise and the one-time gains, then rallied to a 52-week closing high as the market worked through the expense walk. In October it recovered from a weak premarket as the NII raise was absorbed, opened higher, and then sold off through the day as same-day peers traded lower. We read the fade as mostly sector-driven: Citigroup fell 5.6% on a beat, and the Dow gave up an early gain. The JPM-specific content that could have weighed (the buyback cut, the late-September IB slowdown, a higher expense guide) was all public before the open.

The more informative signal is relative. Over the print week JPM outperformed the S&P 500 by about 3.5 points and outperformed every large peer whose shares we track. When the sector is selling off on rates and deposit fears, the market is still paying for the balance-sheet quality, deposit franchise and earnings diversity the third quarter displayed. That is the franchise premium working as it should.

What the tape does not change is the valuation arithmetic. Since our July downgrade at $342.89, the stock has rallied 6.5% to a new high, given it all back, and sits at $339.72, down 0.9% against the S&P 500's +1.5%. That is roughly the in-line outcome a Hold implies. At 3.02x tangible book, the stock trades where it did in July on a tangible book that just fell, and a forward EPS framework that moved to the lower half of its range.

Street Perspective

Debate: Is the buyback cut a vote of confidence or a capital constraint?

Bull view: The bull case being made on the Street is that this is capital allocation at its best: a bank with 17%-plus organic deployment opportunities in lending, the security and resiliency initiative and client financing should not buy its own stock at 3x tangible book. Keeping dry powder ahead of final capital rules adds optionality, and the earnings from deployed capital will exceed the EPS lost from a lower buyback within a few years.

Bear view: The bear camp contends that the cut is at least partly forced. Excess capital fell about $10B in a quarter, $7.9B of unrealized losses hit CET1, and the ratio has fallen five quarters in a row. Buybacks have contributed roughly 3 points a year to EPS growth through share count, and that tailwind is now mostly gone, just as 2027 earnings growth slows.

Our take: The bulls are right on economics and the bears are right on per-share math. Both point to Hold. The deployment logic is sound, and we would rather the firm earn 17% on a loan than 6–7% on its stock. But the firm's own pricing signal (cut purchases by more than four-fifths at ~$340) supports our view that the shares are fairly valued, not cheap.

Debate: Do personal AI agents threaten JPM's low-cost deposits?

Bull view: Some sell-side desks argue the agentic-AI deposit scare is a rerun of every fintech scare before it. JPM has the #1 retail deposit share for six years, added more than 650,000 net new checking accounts, grew consumer deposits 4% and widened the deposit margin in the quarter the fear peaked. It can block credential-sharing agents and has its own Smart Cash tool in pilot.

Bear view: The bear camp contends that low-cost deposits partly rest on customer inertia, and agents are built to remove inertia. Blocking outside agents is a temporary defense that customers and regulators may not tolerate, and a sweep to 3–5% yields across the industry's checking balances would cost the largest banks a large share of their earnings. Higher policy rates make the incentive to sweep larger.

Our take: Real, slow, and measurable. The bear case is directionally right about the mechanism and premature about its speed: the third-quarter data show no damage. We fold it into Bear #3 as an accelerant of deposit repricing in a hiking cycle, and we will grade it on CCB deposit growth, the deposit margin and the share of high-beta deposits each quarter.

Debate: Has capital-markets normalization begun?

Bull view: A growing consensus view is that the third quarter merely came off an exceptional second quarter. Markets revenue was up 15% year over year, equities up 49% on durable prime and derivatives franchises, IB fees up 14% with advisory up 32%, and the CEO expects a lighter-than-usual seasonal decline in the fourth quarter. The base has reset higher.

Bear view: The bear camp contends the turn is visible: Fixed Income declined year over year for the first time since early 2024, IB fees fell 8% sequentially, underwriting fell sharply, and management described a slowdown after the rate sell-off. Fixed Income came in roughly $0.55–0.6B below consensus, and the fourth quarter starts slower.

Our take: Both are describing the same quarter. Markets at $10.3B in a quarter with a rate shock and a sector sell-off is the best evidence yet that the base is structurally higher, which is our Bull #1. The decline from $12.1B is the normalization that management warned about in July. Our $9.5–10.0B base survives; Street numbers that still lean on the first half do not.

Debate: Are higher rates a net positive for JPM?

Bull view: The bull case is mechanical. JPM is asset sensitive, NII guidance rose again, the net yield ex-Markets turned up, and the 2027 central case embeds about $1.25B from forward-curve hikes. Securities losses taken now become higher NII later.

Bear view: The bear camp contends that the costs arrive later and in more places: $7.9B of AOCI losses already hit capital and tangible book, consumer credit weakens if real rates stay high long enough to raise unemployment, middle-market borrowers are strained by input costs and energy, and deposit migration accelerates. The CEO himself sees a higher-than-consensus chance of a 6% ten-year.

Our take: Positive for NII now, negative for capital now, and negative for credit and deposits later. The 2027 guide books the first. The third quarter booked the second, and the third is what Bears #3 and #4 are for. We haircut the rate component of the 2027 bridge and treat higher-for-longer as a modest net negative for the multiple, despite the NII.

Model Update

Valuation impact: Fair value stays at $320–380, midpoint ~$350. Two changes offset. The tangible book base fell 0.6% sequentially on AOCI, and the earnings framework moved to the lower half of its range. Against that, a much smaller buyback at 3x tangible book retains more capital per share, so our 12-month-forward TBVPS estimate holds near $125. On that base, 2.4x (cycle normalizes, ROTCE toward 17–18%, rate costs arrive) gives $300; 2.9x (base) gives ~$363; and 3.2x (Markets holds above $10B, 2027 NII ex-Markets lands at $107B, ROTCE 21%+) gives $400. Cross-checked on earnings, 15x the ~$24 midpoint of our FY27 range gives $360.

At the $339.72 close on October 16, the stock trades at 3.02x tangible book and ~14.6x our FY26 ex-items EPS of ~$23.3. The ~$350 midpoint implies roughly +3% price appreciation plus a 1.9% dividend yield, about +5% total return. That is in line with or slightly below an S&P 500 expected return in the high single digits. It is a Hold.

Thesis Scorecard Post-Earnings

Overall: Thesis unchanged in direction, with the balance of evidence slightly negative for near-term earnings momentum and positive for 2027 NII. Bull #2 strengthened and Bull #1 passed a harder test than the second quarter set. Bull #3 changed character, from a cushion being spent on buybacks to one being protected for deployment. Against that, the rate spike we reframed Bear #3 around has started to show up in capital and tangible book. The earnings base settled at the lower end of our range, and management's own buyback decision lines up with our valuation call.

Action: Maintaining Hold; fair value $320–380 (midpoint ~$350), unchanged. We downgraded in July because the thesis was proven and priced. Three months later the thesis is intact, the price is about where it was, and the firm has stopped buying its own stock at this level. Existing holders own a franchise with a growing deposit base, a 21% ex-items ROTCE and ~10% guided NII growth into 2027, and have no reason to sell. We would not add here. We would turn constructive on a pullback toward the low $300s (roughly 2.7x tangible book), on Markets holding a $10B-plus base through the fourth and first quarters, or on a named successor with a date. We would turn negative if CET1 breaks below 13.8% without a matching return story, if CCB deposit growth or the deposit margin turns down as rates rise, or if AOCI losses keep tangible book flat for another quarter.