JPMorgan (JPM.US) Q3 Earnings Call: Bank Moats Are Solid Enough to Withstand "AI Stealing Deposits," Hedge Fund Leverage Risks Are Manageable, Prolonged High Rates Will Drag on the Economy

JPMorgan (JPM.US) Q3 2026 results beat expectations again, and the company held an earnings call. CFO Jeremy Barnum reported in his opening remarks that the company generated net income of $15.3 billion, or $5.57 per share, with a return on tangible common equity (ROTCE) of 20%; revenue was $50.9 billion, up 8% year over year, or up 9% excluding a $658 million net loss on investment securities, driven primarily by growth in markets revenue, deposit balances and asset management fees. Expenses were $27.3 billion, up 12% year over year, mainly due to higher volume- and revenue-related expenses as well as front-office hiring and wage inflation; credit costs were $2.7 billion, including $2.2 billion of net charge-offs and $539 million of net reserve build; the standardized CET1 ratio at period-end was 13.9%, down 30 basis points quarter over quarter. By segment, Consumer & Community Banking net income was $5.7 billion, revenue up 9% year over year, debit and credit card spend up 10% year over year, deposits up 4%, and client investment assets surpassed $1.4 trillion; Corporate & Investment Bank net income was $8.2 billion, revenue up 14% year over year, with investment banking fees up 14%, fixed income trading down 4% year over year and equity trading up 49%; Asset & Wealth Management net income was $1.8 billion, with a 36% pretax margin, $5.2 trillion in AUM and $7.9 trillion in client assets. On the outlook, the company guided to fourth-quarter net interest income ex-markets of about $25.5 billion, total net interest income of about $28 billion, and expenses of about $27.5 billion, and provided a baseline scenario for 2027 net interest income ex-markets of about $107 billion, with the credit card net charge-off rate maintained at about 3.2%. The Q&A session centered on two main threads: the AI disruption and the rate path. On the risk of agentic AI disrupting deposits, Barnum stressed that yield-seeking flows and fintech competition are not new problems, and the company's strategy remains to defend its share of primary banking relationships; Chairman and CEO Jamie Dimon said he would not allow any external agents into the company's internal systems, and did not want thousands of "rogue agents." On the 2027 NII guidance, Barnum broke it down as roughly $5 billion of growth over the annualized fourth-quarter level, about a quarter of which comes from rate hikes reflected in the yield curve (one expected in December and two in the first half), with the rest from balance sheet growth and the continuation of structural trends; the 2027 expense consensus is about $114 billion, and the company's preliminary view is close to that. On capital allocation, Dimon explained that the buyback decline from about $7 billion to $1 billion was mainly due to better organic investment opportunities, and had nothing to do with AOCI, speculation about the environment, or even Basel III; the company has about $30 billion of excess capital, and the first year of the 10-year, $1.5 trillion Security and Resilience Initiative deployment has already exceeded $200 billion. Management also responded to questions on tightening credit terms, consumer resilience, deposit beta, hedge fund leverage and leveraged finance, noninterest-bearing deposit growth, liquidity rule reform, and AI-related concentration risk; on middle-market clients, the company said it has not seen the so-called insatiable capital demand, with clients generally "worried about others while doing fine themselves," though high rates and rising input costs have introduced some caution. The market is most focused on three threads. First, the AI threat to banks' low-cost deposits: bank stocks reacted after several personal agent products were launched, and the company acknowledged that competitive pressure is real, but stressed that decades of branch networks, value proposition and customer trust form a moat, while making clear it will block shared-credential fund transfers and screen scraping. Second, rates and the credit cycle: Dimon said whether the economy can withstand a 10-year Treasury yield rising to 6% and several Fed rate hikes is a "prudent scenario" that must be prepared for, but made clear it is not a forecast; the company is also tightening credit standards and warned that prolonged high rates will eventually push up unemployment and weigh on consumer credit performance. Third, regulation and capital returns: the market is watching the Basel III endgame and G-SIB surcharge finalization by year-end, possible adjustments to liquidity rules (discount window, HQLA, LCR), and how the company will deploy about $30 billion of excess capital after slowing buybacks. In addition, management gave no clear guidance for fourth-quarter investment banking and markets: the investment banking fee outlook remains constructive, but activity slowed slightly in the second half of September; markets revenue is not guided in principle, with only a reminder that the third to fourth quarter typically sees a seasonal decline, and that recent seasonal patterns have not been typical. Presentation Session Host Good morning, ladies and gentlemen. Welcome to JPMorgan's Third Quarter 2026 Earnings Call. This call is being recorded. [Host instructions] We will now proceed to the presentation. Please note that the forward-looking statements and non-GAAP financial measures discussed on this call can be found in JPMorgan's earnings press release and investor presentation, which are available on the Investor Relations website. Please hold. I will now turn the call over to JPMorgan Chairman and CEO Jamie Dimon, and CFO Jeremy Barnum. Mr. Barnum, please begin. Jeremy Barnum, Executive Vice President and Chief Financial Officer Thank you very much, and good morning, everyone. The company reported net income of $15.3 billion, or $5.57 per share, with a return on tangible common equity (ROTCE) of 20%. These results include a $658 million net loss on investment securities at the corporate level. Revenue was $50.9 billion, up 8% year over year. Excluding the net loss on investment securities, revenue was up 9% year over year, driven primarily by markets revenue, higher deposit balances, and increased asset management fees in Asset & Wealth Management and Consumer & Community Banking. Expenses were $27.3 billion, up 12% year over year, mainly due to higher volume- and revenue-related expenses, as well as front-office hiring and wage inflation. Credit costs were $2.7 billion, including $2.2 billion of net charge-offs and $539 million of net reserve build. On the balance sheet, the standardized CET1 ratio at quarter-end was 13.9%, down 30 basis points from the prior quarter, as net income was offset by OCI losses, higher risk-weighted assets (RWA), and capital distributions. You will notice that we updated the categorization of the RWA movement table to make it more informative. Standardized RWA increased by about $52 billion this quarter, primarily from loan growth in Banking, Markets, and Asset & Wealth Management, as well as the impact of higher energy prices and increased market volatility. Now let's look at the segments. Consumer & Community Banking (CCB) reported net income of $5.7 billion. Revenue was $21.2 billion, up 9% year over year, driven primarily by higher balances and yields in credit card, asset management fees, deposit balances, and auto lease income. A few points worth highlighting. Debit and credit card spend was up 10% year over year, with growth continuing across consumer categories, including gas and travel, supported by consumer demand. Deposits were up 4% year over year, with stronger growth this quarter, as we continued to expand our customer base in Consumer Banking and Business Banking; according to FDIC data, we maintained our number one position in retail deposit share. Client investment assets surpassed $1.4 trillion, up 15% year over year, driven by market performance and strong inflows. Next, Corporate & Investment Bank (CIB) reported net income of $8.2 billion. Revenue was $22.7 billion, up 14% year over year, with strong performance across all businesses. Investment banking fees were up 14% year over year, reflecting strong performance across products, particularly advisory. While our outlook remains constructive amid continued client engagement, recent changes in the market environment have slightly slowed activity, so we will see how the rest of the quarter unfolds. In Markets, fixed income was down 4% year over year, with lower revenue in commodities, rates and credit largely offset by higher client flows in currencies and emerging markets. Equity was up 49% year over year, with continued outperformance across businesses, particularly prime brokerage and derivatives. Turning to Asset & Wealth Management (AWM). AWM reported net income of $1.8 billion, with a 36% pretax margin. Revenue was $6.9 billion, up 14% year over year, driven primarily by higher management fees on higher average market levels, strong net inflows, and higher loan balances. Long-term net inflows were $51 billion, driven primarily by continued strength in fixed income. Assets under management (AUM) were $5.2 trillion, up 13% year over year; client assets were $7.9 trillion, up 15% year over year, driven by continued net inflows and higher market levels. Finally, Corporate reported a net loss of $455 million, which includes the previously mentioned net loss on investment securities, with revenue of $93 million. Now to the outlook. I will update the full-year guidance by providing fourth-quarter guidance, with the corresponding implied full-year figures shown on the page. We expect fourth-quarter net interest income ex-markets of about $25.5 billion, and fourth-quarter total net interest income of about $28 billion. As in prior years, we provided a baseline scenario for 2027 net interest income ex-markets of about $107 billion, which we will refine further next quarter. As a reminder, this baseline scenario is based on a set of assumptions that are clearly highly uncertain even under the best of circumstances. On expenses, we expect about $27.5 billion in the fourth quarter, with the corresponding full-year increase reflecting higher volumes and revenue-related expenses from persistently elevated activity levels. Finally, we continue to expect the credit card net charge-off rate to be about 3.2%. That concludes our opening remarks. We would now be happy to take your questions, so please go ahead. Q&A Session Host [Host instructions] Our first question comes from Ebrahim Poonawala with Bank of America Corp. Ebrahim Poonawala, Bank of America Corp Securities, Research Division I wanted to ask a question there has been a lot of discussion over the past few months around everything AI. If there are two areas you could specifically address: first, on the disruption risk from agentic AI, I know you have talked about this before, but could you share your views based on your understanding of the technology? Second, against the backdrop of various "rogue agents," what challenges do banks face in deploying agentic AI internally? Does this make it harder for banks to actually deploy agentic AI in responding to competitive challenges? Jeremy Barnum, Executive Vice President and Chief Financial Officer Let me start, and do you want to add? Okay. On the disruption point, there are a few things to say. The market is clearly very focused on the Smart Cash solution, and I think it is just one small piece of the overall story. But this product has actually been live for the past few weeks, still at a very, very early stage, essentially still in the nature of a targeted pilot. Still, it is interesting to learn what customers really care about and to observe how such a product sells in real time. So there will be more to come. It is early days, and so far it has been interesting. I think the bigger picture is that several personal agent products were launched this quarter, and they did generate a lot of attention in the market, and we saw bank stocks react a bit after those launches. From our perspective, on the one hand, this is not new but that does not mean we will be complacent. If you think about the nature of the competition these agents bring, it is really the intersection of two things we have long discussed. One is yield-seeking flows. Obviously, we have faced a lot of competition from yield-seeking flows today and through various rate cycles in the past. Our strategy has always been to focus on share of primary banking relationships, rather than chasing every dollar of yield-seeking money. That does not mean the competitive pressure is not real, just that it is not new. The other part is competition from fintech companies that keep innovating and offering excellent customer experiences. We have discussed this for a long time too, and we know we must keep pace with them and invest aggressively enough to ensure we deliver the experience people want through all channels, including digital. In that sense, so-called agentic personal agents and the like are just the intersection of these two things, not new. As I said, that does not mean we are complacent. We spend a lot of time on it, and how this space develops and what we ourselves launch in the future are worth watching. But if you truly step back and look at our competitive position, we feel quite good about the hand we hold, largely thanks to decades of investment in our branch network and value proposition, and especially the trust our customers place in us. Jamie Dimon, Chairman and Chief Executive Officer I would just add one thing: internally, we use agents to do beneficial things, and we will continue to do so. We will also use them to serve customers, but our focus is always on serving customers well. So it is not that we are against agents or for agents or anything like that. They are tools, and you use them to serve customers. There is also a defensive dimension, which is that you must ensure your customers are not harmed by them, and that is indeed a concern. You can think of it as a kind of AI agent concern, and it is amplified because you do not want rogue agents, and there could be thousands of them. So we always stand on the side of the customer. People who use these tools should understand the risks. There is a transfer of liability here, and they should understand that liability has shifted. If it is misused, they should know it is being misused. So we must protect customers in the right way, and we are developing tools for that. At the same time, we will not allow any external agents to freely enter the company, because that is too risky for customers and for JPMorgan. Ebrahim Poonawala, Bank of America Corp Securities, Research Division Understood, that is very helpful. Separately, in assessing the risk that rates may rise further than what we saw in the third quarter, what is your latest view on the resilience of the economy can it withstand a 10-year Treasury yield reaching 6% and several more Fed rate hikes? As we discount that, do you think the risk of a market event or economic stress is significant? Jamie Dimon, Chairman and Chief Executive Officer I think we must we always prepare for a range of different events. That is not the same as predicting them, because we want to be a strong company that can help clients, communities and economies around the world in good times and bad. So these things can certainly happen, and we never rule them out. They can happen for good reasons, such as strong capital demand, a big productivity boost, and everything is fine; or for bad reasons, such as sticky global inflation; and it may not be because of the U.S. itself, but because of what happens overseas. So we are not forecasting. I have always thought the probability is higher than others think, and I still roughly think so now, but that is not a forecast just treat it as prudent preparation. Host Our next question comes from John McDonald with Truist. John McDonald, Truist Securities, Inc., Research Division Jeremy, I was hoping you could unpack the new 2027 net interest income outlook you gave, the baseline scenario, including some of the assumptions around balance sheet growth, deposit pricing and the impact of higher rates. Jeremy Barnum, Executive Vice President and Chief Financial Officer Okay, John, sure. If you start from our fourth-quarter net interest income ex-markets guidance of $25.5 billion and annualize it, you get roughly $102 billion. So we are essentially guiding from that level to $107 billion, or about $5 billion of growth. Of that $5 billion, about a quarter is related to the yield curve. We remain asset-sensitive, and perhaps slightly more asset-sensitive than the EAR (quantitative interest rate risk measure) suggests, because consumer deposit beta is lower than modeled. So about a quarter of it comes from the rate hikes reflected in the current forward curve. I recall one in December, and then two in the first half. The rest, I would describe very plainly as a continuation of the balance sheet growth and structural trends we saw this year. That is, consumer deposits continuing to grow modestly, wholesale deposits continuing to grow wholesale deposits may be a bit weaker than this year's very strong performance, but we are not really sure. Then credit card loans, which can obviously continue to be an important driver of net interest income at comparable levels. So really, it is a fairly straightforward story. I gave the caveats, and I already stated those caveats in my opening remarks. Jamie Dimon, Chairman and Chief Executive Officer Let me add one thing. I think there is more uncertainty around this number than normal, because so much is happening in the world. And obviously, it also relates to what we ourselves do, and we can adjust at any time. John McDonald, Truist Securities, Inc., Research Division Okay, that is helpful. Jeremy, turning to expenses, do you have a preliminary view on next year's investment spending plans, other expense drivers, and how they compare with the current Wall Street forecast for next year's expenses? Jeremy Barnum, Executive Vice President and Chief Financial Officer Okay. We try to make a bit of a distinction between net interest income and expenses: although Jamie is right that we can indeed change net interest income by adjusting the balance sheet mix, overall, absent intervention, net interest income at any point in time is more like a property of the balance sheet and is not very sensitive to the budget cycle. So at this point in the year, we do not feel there is a good reason to pretend we have no view on 2027. Expenses are a bit different they remain sensitive to the budget cycle, and there are many discretionary decisions to make. While we have some preliminary views, this is not fully settled yet. John, you certainly remember that last year we pointed out that our preliminary view on expenses showed the consensus was a bit off. This year seems clearly different. Currently, we see 2027 expense consensus at about $114 billion, and our current preliminary view is not far from that number. So there is arguably no news on this front. Now it is just a matter of how the budget cycle ultimately concludes. As Jamie noted, I think the uncertainty is not only in the macro environment, but also in how fast the investment landscape is changing, and the impact of AI and so on. So there are many variables. But for now, consensus does not look far off. Host Our next question comes from Erika Najarian with UBS Group AG. L. Erika Penala, UBS Group AG Investment Banking, Research Division My first question is on capital management. We have received a lot of investor questions about the difference between the $1 billion buyback and the roughly $7 billion regular pace. I know there is no such thing as a regular pace, and I know buybacks are the last priority. But could you walk us through how we should think about your capital priorities, especially given that we hear Basel III endgame and the G-SIB rules will be finalized by year-end. Jamie Dimon, Chairman and Chief Executive Officer Okay. I think we have told everyone that we have roughly $30 billion of excess capital that is just a rough estimate of the various endgame rules, and we do not know the final outcome but that is a plentiful amount. The capital hierarchy has not changed: after paying dividends (which we hope will remain stable), we love investing in organic growth. For our shareholders, that is the highest and best use of capital. Obviously, buybacks come after that, especially when the stock price is rising. So the decline in buybacks is mainly good news, and has nothing to do with AOCI, speculation about the environment, or even Basel III. We now believe, and this is different from what we told you in the past, that we can redeploy this capital over the long term into places with very good returns. Redeploying capital over the long term into high-return areas is far better than buying back our own stock at an earnings yield of around 6% today. The logic is simple, and this is good good for shareholders. One example, and I am sure there are many, is the Security and Resilience Initiative (SRI), which we told everyone would involve $1.5 trillion over 10 years. I believe we have not yet disclosed the specific number, but the first year may already exceed $200 billion, and we will disclose that and do some related work. But I suspect the actual scale will be much larger. The world needs capital, the world needs advice, the world is becoming more complex, and global assets under management are growing. So we think this is good news over the long term. We cannot tell you what the next six months mean, but that is why we are doing this. We want to allocate this capital well for shareholders. Jeremy Barnum, Executive Vice President and Chief Financial Officer I would just add one thing, Erika. If you look at the trajectory of the CET1 ratio we started at about 15.7%, and now it is down to 13.9%, and during that period we did a lot of internal deployment. There was a lot of organic RWA growth, and some of the investments Jamie mentioned. During that period, we also did a batch of buybacks. But now, as Jamie said, relative to the excess capital we hold and the environment, our pace of organic deployment makes retaining dry powder clearly reasonable. Jamie Dimon, Chairman and Chief Executive Officer I want to praise the investment banking team a bit. They have hired a lot of very talented people for SRI (Security and Resilience Initiative), and we are fortunate to have Todd Combs here investing some of the capital. They have come up with many ideas, and now it is no longer just meetings and discussions. They have helped fund some very interesting projects, and we have more ideas. They have also done an excellent job on the policy work here, such as how to accelerate U.S. security and resilience in areas like shipbuilding, and we had a major announcement in Philadelphia. So I am proud of them. We have quite a few retired people who volunteered to come back, less for anything else than out of patriotism. It has been a remarkable effort for the company. As we noted, this is global, and the global response has also been outstanding. I have a document right here showing more than $10 billion or $20 billion of financing in just one country. So there will be a lot more. I would group the American Dream Initiative, the Innovation Economy Initiative, and the Commercial Banking initiative globally into the same category. These things are working. We believe we are doing the right thing both doing good business and earning good returns. L. Erika Penala, UBS Group AG Investment Banking, Research Division A follow-up on that. It sounds very positive for future activity levels, and broadly speaking. Jamie, I know you have participated in some media events over the past few weeks and talked about the pipeline. Against that backdrop, especially given market concerns that the global bond selloff has had a chilling effect on activity levels, could you talk about what the pipeline means for activity levels, both on Wall Street and on Main Street? Jamie Dimon, Chairman and Chief Executive Officer Okay, before I answer, one point: bond market volatility and that fear do not change how we operate. We serve clients in good times and bad. We understand I call some economic winds weather, and that is not exactly what I mean here. The pipeline is fine. But as I have pointed out many times, the pipeline is like an accordion it can open and close very quickly. In the U.S., as Jeremy said, things have arguably been "pushed out," but we do not really know, and we will have to see. I recently went to Europe, and the pipeline there is actually quite healthy, and IPOs are even better than we expected. So we do not know but serving clients well and doing the work well is what matters most. The pipeline, we always forecast the pipeline, using that probability you mentioned... Jeremy Barnum, Executive Vice President and Chief Financial Officer Probability-adjusted. Jamie Dimon, Chairman and Chief Executive Officer And one thing we can be sure of: it will definitely be wrong. Jeremy Barnum, Executive Vice President and Chief Financial Officer It is worth emphasizing that when Jamie and I say "we do not know," I think we mean: of course we do not know exactly what fourth-quarter investment banking fees will be. But I think we do know that the overall environment is solid and supportive, with a lot of business that wants to get done and needs to get done. As you noted, Erika, given this rate selloff, it is not surprising that the market would take a slight breather and pause to make sure buyers and sellers can agree on valuations. So basically in the second half of September, we did see a bit of that slowdown, and now it depends on how the fourth quarter unfolds sorry, as we enter the fourth quarter. But the big picture, I think, is quite clear. Host Our next question comes from Chris McGratty with KBW. Christopher McGratty, Keefe, Bruyette, & Woods, Inc., Research Division Following up on capital allocation. Could you talk about where you are less willing to allocate capital? It seems there are many opportunities to deploy. Where are you pulling back? Where are you seeing deal structures or peers doing things you may not be willing to do? Jamie Dimon, Chairman and Chief Executive Officer I will not get too specific, but we are tightening slightly on credit. I think some credit terms have become a bit loose. You have heard us and others talk about this before: a bit too much PIK (payment-in-kind), a bit too weak on covenants, a bit too much priming of new debt ahead of existing debt in the capital structure, a bit too aggressive. We just feel there is no need to chase that kind of thing now. Christopher McGratty, Keefe, Bruyette, & Woods, Inc., Research Division Okay, great. Second question, there is a lot of focus on consumer health. What I want to ask is about the durability of consumers in withstanding these rates. I recall you mentioned debit and credit card spend up 10% year over year. If rates remain where they are a year from now, what is your expectation for consumer resilience? Jeremy Barnum, Executive Vice President and Chief Financial Officer Obviously, if you keep real rates very, very high for a long time, at some level, that is precisely intended to cool economic activity to fight inflation. It is hard to imagine that not affecting consumers. So far, based on current data, that has indeed been the case over the past few years: consumers have been quite resilient in the face of a series of headwinds headwinds that in other circumstances could have been challenging. But I think conditioning the consumer credit health outlook on higher rates is a fairly nuanced question. Fundamentally, though, the picture is simple: consumers are very healthy right now, mainly because the labor market is healthy. But central banks need to cool the economy, and that involves an aggressive and prolonged hiking cycle; almost by definition, that means higher unemployment, and therefore weaker consumer credit performance. Host Our next question comes from Ken Usdin with Autonomous Research. Kenneth Usdin, Bernstein Autonomous LLP On the consumer side, it is good to see Consumer & Community Banking deposits up 4% year over year. What I wanted to ask is, apart from agent-related factors, what are you seeing in terms of yield-seeking behavior? Deposit margins have been rising over the past few quarters, which is also good. What I want to ask is whether your expectations for deposit growth, or how you are responding to the current market and deposit competition, have changed? Jeremy Barnum, Executive Vice President and Chief Financial Officer Okay, that is an important question and a good one. I think the simple answer is: our core expectations have not changed significantly. Over the past few quarters, as yield-seeking flows faded, we saw consumer deposit growth recover, and that remains our basic expectation. That is the big picture. I do think that since that outlook was formed, we have seen rates move somewhat higher, which does raise the question: what impact will that have on consumer deposits? Normally, higher-beta products such as CDs and premium deposits will obviously pay higher rates. All else equal, you might expect some internal migration, with funds moving from checking and savings accounts into CDs or similar products, leading to a higher weighted average rate paid. You might also see a bit of extra outflow pressure, making deposit growth slightly weaker. But that is all speculation. It depends a lot on which side of the debate you are on: some argue that the cash sorting that is going to happen has already happened, so these deposits are relatively insensitive to small rate increases. I personally do not think so. I think it is more of a continuum than black and white. So I would expect a bit of extra pressure. But ultimately it is a matter of degree, and no one really knows. The good news is that the core business is performing exceptionally well, largely helped by net new checking account openings I think about 650,000 this quarter. So performance is really, really strong, which is a testament to our value proposition, the customer engagement we are seeing through our marketing efforts, and the strength of the brand. So from a long-term perspective, that is what matters most; the short-term rate impact will just be what it is. Kenneth Usdin, Bernstein Autonomous LLP Great. A follow-up on markets. Obviously, we are seeing clear divergence: equities remain strong; while, as you mentioned earlier, there are tougher year-over-year comparisons and the impact of higher rates on fixed income, currencies and commodities (FICC). Has there been any change in client demand for balance sheet usage, that is, capital consumption? Does it look as "insatiable" as it does in equities? Please talk about the state of activity in markets and your willingness to provide balance sheet but more importantly, what is happening on the demand side. Jeremy Barnum, Executive Vice President and Chief Financial Officer Okay. Overall, the simple answer is no material change. Yes, I like your use of "insatiable" exactly. I very much agree with your description of insatiable capital demand; that phrase feels quite apt. Frankly, I have not heard anything to the contrary recently. But with respect to Jamie's point about tightening credit, that is more of a general reminder to make sure we comb through everything carefully. At a moment when AI is the overwhelming theme, it brings a lot of innovation and appears in many different areas. Capital is being mobilized through various different mechanisms. We just need to track it very, very closely to make sure nothing is missed. Host Our next question comes from Mike Mayo with Wells Fargo & Company. Michael May, Wells Fargo & Company Securities, Research Division Could you talk about how much AI has helped JPMorgan? I see you rank highly in a recent survey on AI and banking. But when I squint at the financials, I try to find the benefits and I am not sure I can. I think if these numbers are outdated, you can update them. You spend roughly $2 billion on AI, and the benefits are also roughly $2 billion? If that data is outdated, please update it. Also, please address the risk that AI agents could pull deposits away from JPMorgan. On that point, could you talk about your value proposition, customer trust, and how you protect customer privacy and how much customers value that? Jeremy Barnum, Executive Vice President and Chief Financial Officer Okay, Mike, let me sort out a few points. First, on AI spending, you are right, I think we did say that number in some media setting. I think it is not entirely accurate, and we corrected it later. That AI spending number at the time was clearly much lower I think around $700 million a year and I think that data is now a bit outdated too. But more importantly, there are a lot of accounting effects in this: if you convert the time everyone spends thinking about this into dollars, you can probably come up with a higher number. So at this point, it is no longer something that can be singled out separately. It is now fairly widely distributed across the organization, and it is obviously a very, very high priority effort. On the benefits question, you are right, we have quantified it in the past. I actually want to move away from that now. I think that past quantification was accurate in its own way, but it included a lot of what you might call "pre-generative AI" machine learning-type things, which are especially important in areas like consumer fraud prevention. That is relevant, but I think when you and others ask about AI now, you are really asking about the impact of generative AI and agentic AI on productivity and cost lines in multiple areas, especially software engineering, but also elsewhere. Maybe you are also asking about the exciting new revenue opportunities at the margin. So I think that is the more relevant focus now. And I am very reluctant to spend a lot of effort quantifying it, because the result is often wasting a lot of people's time inside the organization when we actually need them focused on doing the work. So in terms of observable benefits, you are right, these benefits are not visible in our current quarter's results. I think that is probably not surprising, because no matter how important, dramatic or exciting the theme is, it takes a long time to produce visible results in an institution of our size. Personally, I can say that even in my own technology team supporting finance, I am starting to see increasingly exciting use cases and types of benefits. We experimented with some minor legacy tools that needed rewriting, trying to see if we could rewrite them in one go in about two or three weeks the results were not perfect, but you start to see some very exciting opportunities for acceleration and productivity changes. How long it takes for that to show up remains to be seen. And obviously, some of those benefits will be offset by token costs, which is another topic. But we are excited, and there is a lot going on. I think your other question was about agents and trust, and I addressed part of that earlier. But overall, as Jamie noted, the true north here is protecting customers and of course protecting the company, but more focused on the fact that agents enable customers to move funds in their accounts. I think at present, some services that provide this capability do so through shared credentials; we have long opposed that, and also opposed screen scraping and similar methods. It is not safe. So as long as we are able to block it, we will block it, to protect customers and ourselves. It is important for customers to understand: if they give their credentials to an agent and the agent does something they did not want, the consequences are on them. That is not ideal. We are working to find the right approach. It is part of the whole puzzle. From a broader perspective, trust is obviously an asset, and we have spent decades building it. As I said earlier, we feel good about the hand we hold on this. But we must be clear: being trusted is not an excuse for complacency, nor a reason to provide a mediocre customer experience. So we really have to do both well. Michael Mayo, Wells Fargo & Company Securities, Research Division Just a quick follow-up on token costs. Do you have a scale advantage on token costs versus smaller banks? And are you also shifting more toward much cheaper open-source model AI systems? Jeremy Barnum, Executive Vice President and Chief Financial Officer A bit of everything. I would say our strategy is very focused on minimizing lock-in, maximizing optionality, and maintaining cost management capability. So part of the strategy is: model-agnostic, harness-agnostic, and able to run both proprietary models and open-weight and open-source models, some of them in our own data centers. So it is very much a "have it all" strategy, aimed at getting the best available capabilities at the best price and controlling our own destiny. Jamie Dimon, Chairman and Chief Executive Officer And there are tools emerging that do model routing, sending requests to the best, fastest, cheapest one you do not need to use the most expensive thing for every query, and often people just use the one they are most used to. These will gradually come into place and keep costs under control. Host Our next question comes from Gerard Cassidy with Royal Bank of Canada. Gerard Cassidy, Royal Bank of Canada Capital Markets, Research Division Could you talk about the outlook for deposit beta? Obviously rates are starting to rise, and fed funds futures show more hikes over the next 3 to 6 months. And given the growth in your consumer deposit base since the last tightening cycle in 2022 and 2023, how do you think this time may play out? And obviously, we are starting from a much higher level than in 2022. Does that mean your deposit beta this time may not be as high? Jeremy Barnum, Executive Vice President and Chief Financial Officer Gerard, that is indeed a very good question. I think it is important for everyone to think about carefully, because this particular rate cycle the recent one is arguably even more unprecedented than the last one. We do not have much historical data on how people react when rate hikes begin from a starting point like this one. So I would say this is an unusual decade-plus. I think one thing that is very important for everyone, especially those training models on historical data, is that this historical dataset is simply not rich enough. So thinking outside the box is very important. That said, in that context, I think for us the beta story is actually fairly simple: you can almost think of our business as a barbell with two ends one end is a set of very high-beta deposits, such as most wholesale deposits, as well as CDs, premium deposits and consumer deposits; the other end is a large pile of very low-beta deposits. The question is probably not so much the pricing of any