Sections 3-14 are for paid subscribers. These sections include commentary on First Brands and Tricolor implications, as well as updates on PayPal, Uber, Flutter, Robinhood, SoFi, Amazon etc.

Welcome back to earnings season. Our Taiwan Semi earnings review was published during the week. Next week’s content schedule includes Tesla and Netflix reviews, as well as IBM and Intel coverage. My current portfolio & performance vs. the S&P 500 can be found here.

Important Credit Lingo:

Results:

JPM beat revenue estimates by 1.5%. Missed 2.48% net interest margin (NIM) estimates by 3 basis points (bps; 1 basis point = 0.01%). It also beat $4.81 GAAP EPS estimate by $0.26 or 5.4%.

“

While there have been signs of softening, particularly in jobs, the economy generally remained resilient. However, there continues to be a heightened degree of uncertainty

.” – Jamie Dimon

Credit:

“While we are closely watching the potentially softening labor market, our credit metrics, including early stage delinquencies, remain stable and slightly better than expected.” – CEO Jamie Dimon

“Wholesale charge-offs were slightly elevated as a result of a couple of instances of apparent fraud in certain secured lending facilities. Otherwise, in both wholesale and consumer, credit performance remains in line with our expectations.” – CEO Jamie Dimon

The current facts on the consumer side are that the consumer is resilient, spending is strong, and delinquency rates are actually coming in below expectations… we now expect the 2025 card net charge-off rate to be approximately 3.3% because of favorable delinquency trends driven by the continued resilience of the consumer.” – CFO Jeremy Barnum

“As we sit here right now and we sort of update the macro environment, a few things are true. One is that the personal savings rate is a little bit lower than expected. Consumer spending remained robust while income was a bit lower. That's all else equal decreasing balances per account in CCB.” – CFO Jeremy Barnum

“From the perspective of our credit franchise, this moment of revived animal spirits is driving demand. We're seeing very healthy deal flow. We're seeing acquisition finance come back.” – CFO CFO Jeremy Barnum

It’s great to see NCO meaningfully tick down on a sequential basis alongside card service NCO. Improving card service NCO guidance for the year was equally encouraging. Provisions continue to rise, which is something to keep an eye on, but generally speaking, credit metrics look at equity resilience on a Q/Q and Y/Y basis. Encouraging.

Outlook & Valuation:

The firm slightly raised annual net interest income and net interest income ex-markets guidance for the year. Improved card service net charge-off rate expectations from 3.6% to 3.3% for the year. JPM trades for 15x forward EPS. EPS is expected to slightly fall this year, and compound at a 6% clip over the following two years.

Bank of America beat revenue estimates by 2.3%. It also beat $0.95 GAAP EPS estimates by $0.11 and beat 1.98% net interest margin (NIM) estimates by 3 basis points (bps; 1 basis point = 0.01%).

Credit:

The data below looks good. 30+ day card delinquency rates and non-performing loan (NPL) rates both improved meaningfully on a Y/Y basis, while NPL matched its best result in 2 years. Credit card NCO reached its best level in over a year. Both consumer and commercial NCO data looks resilient and improving, while overall provisions are falling, which points to reasonably strong expectations for future repayment. For a giant consumer bank, this bodes quite well for durable consumer credit health.

“Asset quality remains sound with improvements in several key indicators… Focusing on total net charge-offs again and looking forward, in the near term, we would not expect much change in total net charge-offs given the steady consumer delinquency trends, stability of C&I and reductions in CRE exposures.” – CFO Alastair Borthwick

“In addition to improving consumer losses, note the reductions in both reservable criticized and NPL metrics for the commercial portfolios… We had a modest reserve release associated with improved outlooks for both credit card and commercial real estate.” – CFO Alastair Borthwick

Guidance & Valuation:

Bank of America raised its Q4 net interest income guidance from $15.6B to closer to $15.7B. This would represent about 8% Y/Y growth for Q4. For 2026, they expect to replace about $10B-$15B in assets with higher-yielding products, which should allow them to maintain 2026 growth that’s similar to 2025. Specifically, they guided to roughly 5%-7% growth for next year. BAC trades for 12x forward EPS.

Results:

American Express beat revenue estimates by 2.1%. This was its fastest rate of revenue growth in 7 quarters, with the acceleration being in excess of easier Y/Y comp help. Pre-tax income beat estimates by 6.3%, its 35.9% return on equity (ROE) beat 35.5% estimates by 40 bps (basis points; 1 basis point = 0.01%). Finally, it beat $4 GAAP EPS estimates by $0.12. Strong showing.

Guidance & Valuation:

Thanks to the strong Q3 performance, it raised annual revenue growth guidance from 9% Y/Y to 9.5% Y/Y. That compares favorably to 8.5% Y/Y growth estimates. It also raised $15.25 EPS guidance to $15.35 (about 15% Y/Y growth), which met estimates. Estimates for both 2025 and 2026 are up modestly year-to-date

Credit Data & Commentary:

Credit metrics look as robust, healthy and stable as you’d expect from an ultra-prime originator like this.

“

Credit performance remains excellent with both U.S. consumer and small business delinquency rates still below 2019 levels

.” – CFO Christophe Le Caillec

“

Q3 delinquency and write-off rates were low with delinquency rates flat to last quarter, while write-off rates declined. This performance is supported by our focus on premium products, which tend to attract high-income, highly creditworthy customers

.” – CFO Christophe Le Caillec

More Big Bank Commentary from the Week:

"

Credit performance remained strong and continued to improve. Commercial net loan charge-offs were stable from the second quarter with lower losses in our commercial & industrial loan portfolio, largely offset by higher commercial real estate losses... Consumers continue to be resilient as income growth has generally kept pace with increases in inflation and debt levels

." – Wells Fargo CFO Mike Santomassimo

"

And it's worth noting that across our U.S. cards portfolios, delinquency and NCO rate continue to perform in line with our expectations

." -- Citi CFO Mark Mason

"

The macro environment reflects more global resilience than many anticipated. The U.S. continues to be a pacesetter, driven by consistent consumer spending as well as tech investments.... While growth is cooling somewhat, and we're keeping an eye on the labor market, America's economic engine is indeed still humming

." -- Citi CEO Jane Fraser

Quite the interesting integration announcement this week for Lemonade’s auto business. Going forward, it will be much easier for Tesla owners to use Lemonade’s pay-per-mile, telematics-based car insurance offering. No longer will they need to install a dedicated piece of hardware to gather data. Now, through a tight integration with Tesla’s API, that data will effortlessly stream behind the scenes right to Lemonade's real-time risk measurement.

Legacy players like State Farm and Allstate also have integrations with this API, but their auto products require a separate hardware installation. They both also have a much harder time perpetually and cohesively using this context-rich information. Progressive’s integration requires no dedicated hardware and they do have continuous monitoring/data collection like Lemonade. That’s the closest alternative, but Lemonade is confident in having superior data depth and breadth vs. the field. They say it all the time in earnings calls.

How can one expect a company to know a driver as well as Lemonade if they’re not collecting as much data and not as capable when it comes to making sense of it? They can’t. That’s why it’s able to uncover who is overpaying for their plans due to improper incumbent understanding of risk and a need for these people to subsidize those underpaying for plans. That’s why Lemonade takes granularity to another level. And that’s why Lemonade’s underwriting trends remain so promising despite its overall access to data being tiny compared to the big boys. A superior ability to leverage all of this valuable insight is the key.

And that key didn’t manifest by accident. It’s in their DNA. Lemonade Car was built with scalable data ingestion and telematics at the center. Their data pipeline orchestrates and delivers all needed driver context to associated teams through cohesive, interoperable, light-weight APIs. Unstructured driver data needs tagging and organizing to be used in the right ways, with Lemonade’s plumbing making that wonderfully automated and speedy. Simply put, this tech stack wasn’t built over decades of stitching together disparate systems that struggle to communicate or leverage each other. The tech and AI-native model was purpose-built for wonderful malleability, seamlessly constant software updates, bottleneck-free data consumption and integrations like this one. That means Lemonade Car can almost effortlessly integrate with a wide array of different sensors without nearly as much heavy lifting. They just move faster and more precisely.

“Our telematics pipeline architecture was built in advance so we could connect new data sources with only minor adjustments.” – Lemonade Sr. Director of Engineering Daniel Korn

This also means Lemonade will be positioned to holistically use the highly valuable data Tesla collects on these drivers to offer (it thinks) even steeper discounts vs. the field. 

Others are allowed to build what Lemonade just built using the same Tesla API. The integration certainly isn’t only available to this specific company. That’s not why this is exciting. It’s exciting because it’s another signal of Lemonade’s ability to build slick integrations like this while others haven’t done nearly as good of a job. Incumbents can’t easily bend their applications to connect to other modern systems like Tesla’s. They can, and do, get part of the way there… but not all the way. Nobody else matches the onboarding ease, lower hardware needs, extensive data collection and lower Lemonade pricing. Others can match some of that equation… but not every piece. That should be a compelling offering for Tesla owners.

“In the end, a move like this is only possible thanks to great people and real technological infrastructure.” – Lemonade Senior Director of Engineering Daniel Korn

A few other notes:

The company continues to beef up its board. They added Meta’s VP of AI products today. Added PayPal’s Chief Marketing Officer earlier in the month.

There has been a ton of noise surrounding a few bankruptcies for two auto companies and the potential implications on the private credit market. First Brands is an auto parts supplier that grossly over-levered its balance sheet for acquisitions and somehow saw billions of dollars “vanish” from its books. It collapsed. Alternatively, Tricolor is a subprime auto lender that got too aggressive with originations and couldn’t weather any degree of material downturn in that market. Some believe they had a heavy focus on originating loans for undocumented immigrants, and so stricter immigration rules from this administration have hurt that company a lot. Furthermore, Tricolor allegedly used the same loan pools from its balance sheet to secure funding for several different loans. If the allegations are true, and they sure look to be, that’s fraud.

In both cases, horrible decision-making, extreme risk-taking and illegal activity played much larger roles in both company failures than anything to do with macro. This reminds me a bit of the regional banking crisis, where a few especially aggressive players (Silicon Valley Bank) loaded their balance sheets into ill-advised bonds that led to large unrealized losses as rates were hiked. Nearly all banks and creditors were entirely fine then and I think that will be the case here.

And for more confidence in these issues not spreading widely, the Fed is set to cut rates and end quantitative tightening. That means this modest dose of stress on capital markets comes with a Fed supporting overall liquidity and credit spreads… rather than fighting against those things (like during the Regional Banking Crisis). That makes the likelihood of this spreading even lower. When we look across the creditors that we discussed in section one (and others like Ally who also play in auto) the metrics look very good. The commentary and corporate spreads (FRED data pictured below; lower levels = better liquidity & risk appetite) are equally strong. Even CCC-rated spreads still look reasonably good. So while this could lead to healthier capital markets and private credit discipline with especially risky auto loans, I do not anticipate anything close to the liquidity freezing we saw come in 2022.

“On private credit overall, I haven't heard anything to suggest that the private deals are performing differently from the public deals.” – JP Morgan CFO Jeremy Barnum This Week

Galileo was added to the AWS partner network. I think this is material. We know how many potential clients like to do their building within AWS... and now they can do that Galileo-related building in that environment as well. We also know how powerful of a selling partner AWS has been for CrowdStrike and so many other enterprise software titans. This will make selecting Galileo's software on the AWS marketplace (where it's already listed) more compelling with less associated vendor friction.

We now know who Anthony Noto was referring to when he previously told investors Galileo had signed 2 other large travel companies besides Wyndham. These are Southwest Airlines and United Airlines. Two high-quality, long-standing companies trusting their financial product evolutions to Galileo. This provides more proof of concept and reputation building that should lead to more Galileo deals. We already know the pipeline is swelling (per the team). This is already part of guidance.

In terms of the First Brands and Tricolor impact on SoFi’s capital market access, I don’t expect it to be material. For review, SoFi focuses on the largest, highest-quality institutions looking for multi-year relationships. It doesn’t do much in subprime and is overly conservative when it comes to leverage and using its balance sheet. Because of strict ultra-prime focus and strong underwriting, its credit health has been durable across cycles and has built trust with private credit giants like Blue Owl. For reasons discussed in section 2, I do not think this will aggressively impact liquidity or growth appetite for private creditors. I think this will make them a lot more picky, but that’s great news for SoFi. This isn’t your average Fintech… this is the 750+ FICO, $150K+ borrower income fintech.

It looks like the omni-channel push is going well in Germany. According to a popular German magazine (DER AKTIONÄR), they're up to 5 million customers opting into contactless digital wallet usage in physical stores since April's launch. For those people, their PayPal app is now their brick and mortar wallet.

This is an important sign of progress and an encouraging indication that eventual expansion will be fruitful (even if Germany is arguably its best market). PayPal has plans to debut this direct app payment offering across the globe (hopefully the USA soon if regulation cooperates). That will mean more users can access in-person tap-to-pay convenience without going through other competing wallets like Apple or Google.

Providing this seamless transaction method is vital for PayPal’s compelling product suite. I realize that many other players have this already... but the reality is that PayPal didn't. It needs to match alternatives like Apple Pay in this area just like SoFi needed to match competing brokerages by filling their own product gaps. In both cases, it's not that these launches will become product differentiators. They simply remove two large product gaps. I am hoping PayPal Wallet tap-to-pay quickly launches across the rest of the planet, as I think it will be great for incremental market share and volume.

Goldman downgraded the company to sell. They don’t believe in the guidance management has offered, as they now see low single-digit transaction margin growth next year. They think lapping Braintree profit tailwinds, credit growth, tariffs & general branded competition will make it hard for the firm to meet their guidance for next year.

I disagree. Every sign we’ve seen pertaining to this firm’s turnaround  has been strong. Venmo acceleration, branded stabilizing, Braintree price hikes going better than expected, budding ads business, rising value-added service momentum, and a long list of expanding partnerships with industry leaders. Leadership has repeatedly told us how important it is for them to meet their promises after years of the old team burning Wall Street trust. And Alex Chriss has done exactly that since taking over. I think Goldman is wrong. They think I’m wrong. Time will tell.

Waymo is debuting in London with Moove as the fleet manager and is testing a DoorDash delivery partnership in Arizona. Waymo leadership has talked about tinkering with many, many different business models while they figure out which service providers offer the best customer experience and unit economics. Waymo would be silly not to do this. There is every reason to believe Uber will be that partner down the road. Uber's best-in-class scale is already clearly driving sky-high utilization rates for these cars on a consistent basis. If there's a future competitive landscape with many players (looking highly probable), I don't see anyone else matching this strength. And if I'm right, Uber will be able to pay Waymo and everyone else more money for fleet access than Lyft or others can. While we want every deployment to be with only Uber, the ride-sharing giant does own a 10%+ stake in Moove and has two board seats. Considering more partnerships are all but inevitable, it’s encouraging to see a chunk of these including companies that Uber already owns pieces of.

Uber is offering payment to drivers for completing AI data labeling assignments. These assignments include simple things like uploading pictures of surrounding areas. This will help Uber keep building a world-class base of training data for AV fleets to utilize. All of those algorithms are only as good as the physical and simulated data they’re trained on. Uber has a lot of value to provide thanks to its years as the market leader. With the help of recently acquired data labeling company Segments.ai, they’re more capable in tagging and unleashing the value of this data. I think this is an exciting utility complement to the fleet utilization maximization edge Uber will provide. It’s a great focus area in my opinion.

Guggenheim initiated Uber with a buy rating and a $140 price target based on the strength of the core business and their optimism surrounding it staying strong. They see autonomy as a net positive for the company and see the delivery business as “underestimated.” 

BetMGM is executing. Strong demand prompted them to raise revenue and EBITDA forecasts for the year by 2% and 33%, respectively. Confidence to again raise these targets stems directly from underlying momentum. It is really great to hear this as prediction market debates heat up. This is not nearly as strong of a player as FanDuel or DraftKings, and they're doing quite well. If they can hold their own, I'm even more confident in the other two doing so as well. And this week’s New York State data again supports that notion. Both saw volume rise by 13% Y/Y (hold rate was also a strong 12%). Game timing can influence this, but the growth has been consistently strong since football season started. Healthy growth in one of the most mature states as prediction market risks supposedly ramp.

Building on last week’s Quick Suite launch, Amazon announces Bedrock AgentCore this week. In essence, this is Amazon’s fully-managed environment for enterprise agent creation. This building can be done with just a few lines of code. AgentCore plugs into the same core Amazon capabilities to fully manage agentic compute and cover mission-critical security. It connects memory-equipped agents to relevant data, context and agentic integrations to ensure these autonomous assets are capable, optimally productive and can freely race throughout the digital ether. From there, AgentCore covers vital identity integrations, provides the required agent memory and all other tools a developer needs in one place. With this, they can focus on building next-gen software. Like everything else within Bedrock, AgentCore’s foundation is built with model and developer kit flexibility in mind. Under a singular interface, users can tap into virtually whatever model or 3rd-party tool they need. And for more flexibility, AgentCore tools can be purchased separately or bundled together… whatever the customer wants. Specifically, AgentCore features the following products:

The software development kit already has 1M+ downloads and customers like Cohere Health are cutting patient review times by 30%-40% and “raising clinical determination accuracy” by 30%. Other early customers include Ericsson, Experian, Sony etc. More tools that make Amazon a more holistic and capable GenAI and agentic AI cloud partner for all of its customers. I expect all of these launches to gain meaningful traction and elongate this still massive cloud growth runway.

Some Robinhood users took to social media last weekend to voice their displeasure with the firm. This came in response to broad-based crypto trading outages on Friday and into the weekend. Many were having their transactions blocked as volatility sharply ramped. Certainly not a positive… but social media “boycotts” are always loud… always intimidating… and usually noisy. There are absolutely exceptions, but far more examples of these events turning out to be irrelevant. 

Netflix deals with a “boycott” seemingly every year. Disney Plus has had its own “boycotts” in recent quarters that led to no material impact on user growth. For every social media account you see claiming they’re leaving the service, there are many, many more who either don’t know these outages occurred, don’t plan on leaving or both. Believe me… as a SoFi bull… I’d love all of their customers to leave for my holding because of this. I just don’t think that’s going to happen. The evidence we’re seeing on X is entirely anecdotal and that app is not emblematic of the overall population. I think this is noise. The company is rapidly ramping prediction markets, accelerating credit card issuance and gearing up for a banking launch. Those things are more important than a few hundred angry people venting on social media.

Datadog is again exploring buying GitLab for $60/share. This would be a large purchase for DataDog (13% of its market cap) and would help it move "further left" (closer to initial code creation) in the Development, Security and Operations (DevSecOps) category. DDOG's core observability niche is firmly entrenched in the operations side, while GTLB is well-versed in the development side as they work more closely with software engineers in the source code writing phase of DevSecOps. Both have budding security portfolios spanning the entire software package lifecycle, and GTLB's talents here would surely be a welcomed addition as well. The combination would make DDOG a much more end-to-end DevSecOps platform and a more powerful vendor consolidator.

Meta and Blue Owl closed a $30B private credit data center deal. Meta will own 20% of the 5-gigawatt project with Blue Own owning the rest. This keeps a lot of debt off of Meta’s balance sheet and makes them a tenant rather than a full owner. That will shift more of their cost structure from CapEx to OpEx and help with incurred depreciation expense growth. Expect a lot more of these deals in the future. Meta also announced a new $1.5B data center in Texas and a partnership with ARM to help improve its ad ranking and content recommendations.

Citizens JMP reiterated an outperform rating on Duolingo. That’s not super notable, but the reasoning is. They’re seeing the social data that they track show signs of improvement as DUOL moves away from the AI-first backlash and normalizes social media content. Great to hear. Social Blade is showing a meaningful YouTube improvement but only modest improvements for other platforms. Good to hear that their own alternative data is telling an even more positive story.

RBC channel checks are pointing to an outperforming Shopify quarter as that company continues to execute at a high level. They think volume will beat and slower hiring will drive margin outperformance. All subscriptions are performing well and market share is rising. 

Oracle guided to $225B in FY 2030 revenue and $21 in EPS and told investors that its remaining performance obligations (RPO) crossed $500B. I don’t have access to 2030 estimate data, but this led to 2029 revenue estimates jumping 8% and EPS estimates rising by nearly 3% – implying that the revenue will likely be margin dilutive.

Starbucks CEO Brian Niccol told investors that the China business is worth “more than $10B.” Previously, the expected price tag ceiling was $10B. He also said the protein add-on program and the overall Green Apron Initiative are both going well.

Alphabet announced a new $9B South Carolina data center investment and a $15B data center deal in India. Waymo is also launching in London in 2026. Gemini 3 is also coming in December.

JP Morgan published a note that explains the Sea Limited and Mercado Libre mid-week selloffs. The analyst was seemingly able to get in touch with the investors that SE spoke to this week. Their takeaway aligns with the assumptions I made in the Discord a few days ago. SE is leaning aggressively into growth in Latin America. That could lead to a more competitive environment and some margin pressure in 2026, which obviously has implications for MELI. Some think that could lead to operating leverage halting next year, but JPM called that too negative and expects margin expansion to continue. While this could continue to create short-term noise, I think it's the right decision for SE. And for MELI? It's just more of the same. This is what happens when you play in arguably the most appealing e-commerce growth markets in the world. There will be competition. There always has been competition. MELI has had to overcome it for years and years and years. I expect this to be more of the same. Their fulfillment/e-comm value proposition (for consumers and merchants) is fantastic, their economies of scale are unmatched, their cross-selling engine is humming, their leadership team remains elite and their brand is both ubiquitous and loved in that region. They'll keep taking market share and keep enjoying strong structural tailwinds. Any extreme noise related to items like this is noise to be taken advantage of –in my opinion. There was nothing on credit health concerns in LatAm, which is why Nu didn't react.

We’re not getting much data with the government shutdown continuing on. The Philly Fed Manufacturing Index for October was -12.8 vs. 8.6 expected and 23.2 last month. Furthermore, Powell said this week that passive quantitative tightening (QT) will end in the coming months. As a reminder, passive QT is fed balance sheet reduction via not replacing maturing bonds with comparable assets. This drains liquidity from the overall U.S. economy and that drain is ending.