In case you missed it, here’s what I published recently:

So much more is coming soon. Earnings mania starts next week. I’ll send out detailed earnings reviews on Meta, Amazon, PayPal, Microsoft, AMD, Uber, SoFi, Alphabet and Apple. I’ll also send coverage of Visa, Mastercard, Intel, Coinbase, Robinhood and Chipotle earnings.

Coupang (CPNG) will also be the topic of my next deep dive. Tons of other coverage to get to while I make time for that.

Enphase exists in an extremely cyclical solar energy cycle. Higher rates recently made its installations more expensive for consumers and enterprises. Rate cuts take time to flow through to its sector (especially on the consumer side) to support demand. Bulls will say that easy monetary policy will be like wildfire for this business and that times are about to get much more fun. Bears point to company-specific production issues, tightening California solar energy re-selling regulation, and hefty Chinese competition/Europe struggles as the causes of this firm’s recent decline. While there do seem to be micro-level headwinds, I find it extremely probable that part of this weakness is macro-related. We will see which factors have the strongest impact in the coming quarters.

“While every country in Europe has its nuances, the overall business environment in the region is challenging. Power prices have declined from the early 2023 highs, the economic growth is slow, and the consumer confidence is limited.”

CEO Badri Kothandaraman

ENPH trades for 24x forward earnings. EPS is expected to fall by 51% this year and grow by 82% next year (2-year CAGR of -10%).

IBM reiterated its annual $12 billion+ FCF guidance. This compares to $12.2 billion estimates. It also guided to stable growth rates in Q4 vs. Q3, which essentially met estimates.

IBM trades for 21x forward earnings. EPS is expected to compound at about a 5% clip over the coming years.

LRCX trades for 22x forward earnings. Earnings are expected to grow by 17% this year and by 19% next year.

Deckers owns Hoka, Teva, Ugg and other shoe brands.

Deckers trades for 21x forward earnings. EPS is expected to compound at a 13% clip for the next two years.

Carahsoft provides information technology (IT) solutions for the public sector. It resells products created by companies such as SAP, Microsoft, ServiceNow and also CrowdStrike. Last month, the FBI raided its Virginia headquarters due to rumored contract price fixing and false claims. This week, Bloomberg published an article about a $32 million purchase Carahsoft made in 2023 from CrowdStrike for its Identity Threat Protection to deliver to the IRS. The IRS never bought the software, but CrowdStrike has been receiving regular payments for the “non-cancellable order,” with the last payment due in a few weeks.

This merely builds on the negative sentiment surrounding post-outage CrowdStrike, but I think this is noise to hopefully take advantage of if Mr. Market cooperates. Why? CrowdStrike vetted this contract “extensively,” with company representatives calling all of this “inaccurate speculation” and “misleading.” It was not creating a fictitious purchase on its own; it was providing software (that, again, it was paid for) to a federal client through a legitimate distributor. It is not CrowdStrike’s fault that this distributor is being investigated for other issues. 

Ask yourself: Do you really think this leadership team would risk their entire company to try to trick investors with a hollow contract to beat a single quarter’s earnings estimates? Do you think they’d risk their sterling reputation, world-class business model and unmatched financials for this? I think the answer is glaringly obvious “no.” Just go look at how admirably this team handled the historically bad July 19th blunder. This contract represents less than 0.4% of current ARR considering payments to CrowdStrike were spread over two years. 

I find it nearly impossible that these funds will be clawed back by Carashoft, especially as their own company told Bloomberg that it “stands by the transaction” this week. I look forward to CrowdStrike leadership dispelling all of these concerns whenever it publicly speaks next. That is what I fully expect… An anonymous source saying they wish the company would have waited another week to announce the deal does not change my thoughts on this company in the slightest.

Also consider this: ServiceNow’s relationship with Carahsoft is more material than CrowdStrike. And? On ServiceNow’s last call, it called this drama a complete, immaterial nothing burger and told concerned sell-siders to relax. 

Those are my thoughts on CrowdStrike, the company. CrowdStrike the stock has raced back to pre-outage multiples and is back to “extremely expensive” territory. And? It’s there with materially more execution risk looking ahead. I am selfishly rooting for noise like this to drag down the stock in the near-term so that I can buy more shares.

Spotify is building a supply side platform (SSP) for its advertising business called the Spotify Ad Exchange (SAX). The product easily plugs agencies and buyers into its universe of impressions for open bidding. Notably, The Trade Desk was named as the first demand side platform (DSP) for this project. The two started testing last week. As part of the arrangement, Spotify will also use TTD’s OpenPath. OpenPath blazes a more direct trail between publishers and advertisers and allows capable sell side players to directly practice their own yield management – rather than using another vendor like Magnite to do so. Spotify will also adopt unified ID 2.0, which is TTD’s open internet identifier. With this, TTD can tell buyers exactly which eyeballs they want to target, what they’re worth and where to find them. Powerful stuff… and Spotify agrees.

For the audio giant, this matters a lot. Most of its advertising business has been built from direct sales, rather than real-time, programmatic auctions. These auctions are by far the best tool to ensure optimal price discovery for the seller (Spotify). They’re also fully equipped with all of TTD’s targeting algorithms and data to ensure buyers get maximum return on ad spend (ROAS). This added precision and context used in each decision creates a compelling win-win. Buyers will pay more for impressions because they lead to more revenue and profit. And? Spotify is happy to collect higher premiums.

For The Trade Desk, Spotify is an ideal partner to grow its programmatic audio business. This partnership is starting with podcast impressions, but will soon expand to audio. The podcast industry continues to rapidly grow and is well behind video streaming in proportion of ad impressions that are fully biddable. The Trade Desk can help it rapidly catch up… can make all stakeholders more successful… and can turn audio into a third powerful growth vector to join connected TV and retail media.

I love it when best-in-breed, complementary companies partner.

Lending Club is a loan marketplace that pairs qualified borrowers with lenders and investors and also stores some of this credit on its own balance sheet. The borrower here is not ultra-prime like at SoFi, but instead more subprime. Lending Club is a great hint for what to expect from SoFi in terms of origination appetite and outlook, fair value markings and capital market loan sales. Considering lending can be the biggest driver of SoFi’s financial outperformance or disappointment, this matters a lot. LC is also a decent gauge for credit health, although SoFi’s should be more resilient amid fragile macro thanks to its more affluent borrower.

The Lending Club quarter was mostly strong for reasons that are encouraging for SoFi as we head into its own earnings report. That does not necessarily mean the stock will continue to rocket higher and higher – it could easily cool off. It simply means that I’m quite optimistic about the near future data coming.

Overall results were much better than expectations. This growth engine is more cyclical than SoFi’s but the quarter did deliver clear evidence of next year’s 17% revenue growth estimate being attainable. Outperforming origination volumes and loan pricing (more later) led to $202 million in quarterly revenue or a 6% beat vs. expectations. Specifically, origination volume rose by 26% Y/Y to reach $1.91 billion. And as planned, LC delivered 56% Y/Y growth in held for investment loans as it expanded its balance sheet alongside an improving macro backdrop. The outperformance on the top line facilitated a near doubling of GAAP EPS estimates, with $0.13 topping expectations by $0.06.

Interestingly, Lending Club’s origination guide of $1.85 billion missed estimates by about 4%. All other results and commentary surrounding capital markets (more later) were enough to offset this negative.

Lending Club’s capital ratios continue to offer meaningful cushions for accelerating origination volume and stashing more balance sheet risk. Its common equity tier (CET) 1 ratio of 15.9% and tier 1 leverage ratio of 11.3% are both well in excess of regulatory minimums. And? Its originated credit “continues to outperform.”

Notably, credit loss provisions fell by about $17 million Y/Y to $47.5 million. They did tick higher Q/Q, but that was due to the growth in held-for-investment loans (loans it keeps on the balance sheet). And for evidence of getting through peak credit losses, net-charge off ratio of 5.4% improved sharply Q/Q from 6.2%. More balance sheet items to love:

The sources of quarterly outperformance were ideal indicators for SoFi. Outperforming credit trends pair perfectly with a 100 bps inter-quarter 2-year yield decline, like a bottle of red and a filet, to set the table for the success worked through above. Lower yields mean lower cost of capital, lower loan hurdle rates and more liquid capital markets.

With the table set, Lending Club has enjoyed a “return of bank buyers” and brand new capital market buyer demand too. Specifically, it placed $475 million in loan sales during the quarter with two parties who Lending Club expects to buy another $500 million+ over the coming year. General commentary about risk appetite was more positive compared to recent quarters. Even more encouragingly, “pricing on these sales will support improvement of average selling price across all loan sales,” meaning gain on sale margin dynamics for these transactions were better than average for LC. Favorable pricing and outperforming credit metrics led to it making up its held-for-sale loan portfolio by $9 million. Ironically, many of the SoFi bears, who love Lending Club, openly criticize SoFi for doing the exact same thing. Quite the double standard, but I digress.

For years I’ve been pounding the table that SoFi the company is a gem. Its combination of margin accretive growth, successful product expansion, rising customer engagement, stellar leadership, rocketing margins and surgical balance sheet management are finally being rewarded. Rate cuts have certainly helped brighten sentiment, capital market demand and risk appetite, but cuts only help companies that position themselves to take advantage of them. SoFi had to effectively underwrite through a nasty cycle to keep some capital market liquidity prevalent and its balance sheet healthy… it had to bob and weave to create a large capital ratio cushion… it had to find growth while its large student business was shuttered… it had to rapidly inflect its margin profile to be taken seriously… it had to weather the storm for its tech business the best it could. It can now enjoy the economic elixir that rate cuts represent because it did all of these things. “Luck is what happens when preparation meets opportunity.” That opportunity is now around the corner.

This is all encouraging. I’d be lying to you if I said I wasn't excited to see this investment case finally begin to bear fruit (long way to go). At the same time, SoFi the stock has been on an absolute tear since the cuts were announced. Profit estimates haven’t budged much, so the multiple has expanded. It’s worth noting that it’s expanding from absurdly cheap to just cheap, but it’s still getting more expensive.

It is not healthy for stocks to go up in a straight line — even if they’re simply returning to prices from previous years with much stronger fundamentals like SoFi is. Digesting explosive moves is what’s healthy. While I am very optimistic in SoFi the company delivering great results and even better guidance next week, I also think those numbers could easily be shrugged off as SoFi chops around for the time being. It will probably take numbers that lead to brisk upward profit revisions to keep this train running for now; that’s entirely possible if it leans into lending like most expect.

I have no interest in selling any shares at all. I am still considering a put option hedge heading into earnings, but haven’t made any decisions (you’ll know if I do anything – I’m eyeing $11.50/share). It could be time for this stock to temporarily exhale.

While Microsoft got the mega-cap tech GenAI investment ball rolling with a large OpenAI stake, others have closely followed in its footsteps. Amazon (and Alphabet) placed very large bets in Anthropic, which is a direct OpenAI competitor. Notably, Anthropic’s founders came from OpenAI… and even more notably… Anthropic CEO Dario Amodei was asked by OpenAI’s board (while Sam Altman was temporarily ousted) to merge the two firms and run the combined entity. He declined and Sam was quickly brought back. While I’m speculating, between Microsoft’s challenges on Copilot product quality and OpenAI’s revolving door of an executive team, I think Amazon and Google are elated to be invested in Anthropic instead. Note that the UK is investigating the legality of Alphabet’s investment in Anthropic (probably because Gemini is direct competition) but not Amazon’s involvement.

This week, Anthropic made a series of model enhancement announcements. First, it launched its upgraded Claude 3.5 Sonnet model on Amazon Bedrock. As a reminder, Bedrock is Amazon’s fully managed environment where customers can access what Amazon calls a world-class supply of 3rd party models like Claude and Amazon’s own Titan models. The model is extremely impressive across several key performance indicators (KPIs). For example, Anthropic was already world class in SWE-Bench Verified (measures AI model ability to code and “solve real-world engineering problems") and took another large leap forward (score jumped from 33.4% to 49.0%) with this announcement to extend the lead.

Per Anthropic, the model also excels at vision and writing tasks. Specifically for vision, the company calls Claude 3.5 “industry leading,” with that being especially accurate in its ability to read charts and graphs (even very low quality images). Next, its “enhanced reasoning” makes it perfect for “solving complex cognitive tasks like understanding nuanced instructions. Along those lines, the model is the perfect conductor of Agentic AI tasks. As a reminder, Agentic AI is goal-oriented AI where the models themselves are actually tasked with realizing that goal in the best way they know how to (less rigid instruction; more “just go do this for me”).

All in all, here’s how favorably the newest Claude 3.5 model compares to others (note that Gemini is Alphabet, which again is an investor here):

Claude 3.5 also includes a brand new capability called “computer use.” With this tool, Claude 3.5 can “generate computer actions like mouse clicks” to automate tedious tasks like online checkout. I know every single payment processor would love to get their hands on something like this. Fewer clicks; higher conversion rates; more revenue. This is a brand new capability in public beta testing. It is pretty limited at this point, but will improve over time. DoorDash is using this model to reduce app development by 50%, cut customer service response times by 50% and improve overall customer satisfaction levels. This uses Amazon Bedrock Knowledge Bases, which is the firm’s tool for efficient, scalable first party data onboarding and querying (through Retrieval Augmented Generation (RAG)) for GenAI apps.

In other news (also depicted in the chart above) Claude 3.5 Haiku will soon be released. This won’t be multimodal to start, but the text-only product will eventually add image processing too. This is purpose-built for use cases reliant on minimum cost and maximum speed. It will offer chatbots, data querying and processing and code completion tools too.

Needham came out with some encouraging advertising channel checks for Amazon this week. They included two large advertising agencies and their commentary on client spend levels with Amazon. For the first, 16% Y/Y client spend growth should accelerate to 18% in Q3 and “even more” in Q4. Amazon’s unique ability to connect streaming ad impressions directly to its marketplace makes it arguably the best streamer for driving more business at high return on ad spend – especially during the holidays. That’s what the agency says, and I wholeheartedly agree. 

The vertical integration associated with owning the marketplace and the impressions creates opportunities for custom impressions and more granular targeting. This is all driven by Amazon’s highly-regarded ad tech stack and its massive consumer base. The other agency talked about Amazon inundating the streaming market with more advertising supply since its Prime Video change earlier in the year. Cost per impression across the industry is suffering a bit as supply gluts push this to more of a buyer’s market. Still, he sees Amazon winning through this market reset as clients boost Prime Vide spending from 0% of budget to 1%-3% this year. That should lead to about 13% ad growth in Q3 and 18% in Q4 for that specific agency’s clients. Amazon ad impressions come with sky-high margins. They’ve already purchased the content and built the traffic. Now it’s time to monetize. Finally, Needham, as part of a buy rating reiteration and a $210 target, also thinks that cost savings from previously announced layoffs will offset added Project Kuiper expenses. 

This is important to emphasize. The main debate for Amazon this quarter is on near term margin pressures. Skeptics think EBIT estimates are too ambitious as AMZN leans into GenAI investments, Project Kuiper, and seller fee changes while the consumer weakens. The last item is why Bank of America sees mixed results for Amazon next week. I’m more optimistic. I think recently announced layoffs, ad growth, continued cost to serve reductions and accelerating cloud trends make these fears “overblown” like Jefferies also said this week.

BMO also boosted its Q3 2024 AWS growth estimate from 19% to 20% and added that it expects growth to accelerate more through 2025.

As a reminder, here’s a list of the notable partnership news PayPal has delivered during the last several weeks:

Under new leadership, PayPal has moved away from prioritizing low-quality, low-margin revenue to fixating on profitable growth. To do that, it has greatly accelerated the pace of innovation on things like Fastlane and mobile checkout. It also halted irrational predatory pricing practices being used at Braintree to only pursue deals that meaningfully contribute to transaction profit dollars. It hiked prices, let some customers churn and turned Braintree into a profitable growth engine in very little time. By making itself less of an enemy amongst white label payment processors, embracing partnerships and creating products with actual value, PayPal has opened the floodgates for new deals to be struck with important ecosystem participants.

This week, Global Payments Inc. (GPN) ($25 billion company) and PayPal announced a new Fastlane integration for GPN’s base of clients. Apparently the 50% boost to guest checkout conversion rates and 28% faster time-to-completion were compelling to them. Shocking. That will be another shot in the arm, adding new product adoption in large chunks (like Fiserv and Adyen will be). Beyond Fastlane, PayPal will offer GPN’s U.S. merchants “enhanced PayPal and Venmo checkout.” To me, this just sounds like PayPal will bring its latest and greatest flows to the GPN product offering.

The Venmo note is especially important. Aside from fattening up Braintree margins, PayPal has the same emphasis at Venmo as well. While that brand is beloved by young, affluent, valuable customers across the USA, it has done an awful job with monetizing. Non-peer-to-peer products like cards weren’t prioritized at all. I’m the only person I know with a Venmo card and I use it for two categories to earn 2%+ cashback. Not bad. I think a lot of others would find this compelling if they knew about it; PayPal is working on visibility.  Aside from cards, Venmo business accounts give merchants direct access to the powerful localized payment feeds to purchase impressions. Merchants can directly tie themselves to a user purchasing a donut from a bakery to tell their friends about it and to attach a promotion. This means they’re connecting impressions to communities that are actually voting with their wallets, rather than saying “trust me this product is good.”

Cards and business accounts are key pieces of monetization, but checkout integration is the true, high margin prize. Along with cards, it is how Venmo will reduce the 90% of funds entering its ecosystem and immediately leaving it by giving customers more reasons to keep money with them. That, in turn, will drive net interest income, interchange revenue, checkout revenue etc. Good piece of news here.

Walmart launched same-day prescription deliveries in 6 states. It will be in all 50 by January. Walmart+ members will get free delivery, with non-members paying $9.95 per order. This is how Walmart (and Amazon) can make itself an indispensable staple in the day-to-day lives of Americans. It’s how the company can build value, pricing power and retention for its subscription and unlock another massive piece of the U.S. consumer economy.

Walmart (and again, Amazon) are the two best positioned to capture this opportunity. They have several other tools to cross-sell, juice customer lifetime value (LTV) and justify offering lower prices to win more market share. They have massive fulfillment capacity, with retail pharmacy operations already in place. They are also able to tie other purchases to these prescriptions to create more convenient customer one-stop shops. They have the trusted brand to win this sensitive market.

I think when we look out 10 years from now, pharmacy is going to be an oligopoly dominated by these two players. Costco, Kroger and a few others will likely carve out pieces too, with CVS hanging on by a thread. This category is quickly being disrupted and these are the two clear winners in my mind. Not stand-alone point solutions with claims to data moats; not legacy pharmacies… Walmart and Amazon. 

Loop Capital expects strong Microsoft results. Channel checks point to favorable client spending levels (including for Azure). Workload optimization continues to level-off as new workload demand builds. System Integrator channel checks also show strong Microsoft Copilot adoption, which is encouraging considering the negative press that product has recently received. Interestingly, it said monetization of Copilot was lagging adoption. That surprised me, considering Microsoft has been the most aggressive GenAI monetizer in software. It still continues to call Microsoft a top GenAI play and sees a 20%+ 7 year FCF CAGR. That’s insane at this scale. It also had this to say about what I think is MSFT’s most underrated product in Microsoft Fabric:

“One surprising data point we uncovered during our industry checks was the increasing market awareness regarding the company's Fabric (AI-driven analytics platform), which we believe is a reflection of the company's M365 and GenAI go-to-market efforts. We will be closely monitoring this development, which could further disrupt the overall data lake and analytics market.” – Loop Capital 

RBC came out with a note calling Tesla robotaxi plans a risk for Uber and Lyft in California next year. We’ll see what the actual timeline for any rollout is, but the analyst is not wrong. The autonomous wave creates some uncertainty for the positioning of these two firms (especially Lyft). Uber’s network effect as the demand aggregator is a massive plus and is why Waymo continues to work more closely with them, but we’re still moving from guaranteed duopoly to something less certain. This is why I’ve lightened up on the holding and have taken considerable profits in recent weeks. I remain optimistic that Uber will carve out its niche here in the coming years. The probability of that happening is high, but lower than it has been in a long time and I wanted to harvest profits. Max readers have already been updated with all of those thoughts.

Honeywell and Google announced a new cloud partnership to accelerate AI tool adoption for industries. It also partnered with Qualcomm in similar capacity for the automotive market.

Morgan Stanley is bullish heading into Shopify earnings. It sees strong GMV trends, strong product uptake on newer tools like Shopify Audiences and continued cost discipline. It sees new returns-based marketing initiatives propping up operating expenses a bit, as Shopify finds more productive places to spend. At the same time, it sees the overall margin trajectory overcoming this and staying positive.

Snowflake and ServiceNow expanded their partnership to include zero copy data sharing with NOW’s new data tools. It also announced a new ServiceNow CortexAI integration.

The rumored issues with Nvidia Blackwell chip yields have been resolved.

Apple is supposedly considering shuttering its Vision Pro project. I’m sure Meta wouldn’t mind. Popular Taiwanese Apple analyst Ming-Chi Kuo cut iPhone 16 order estimates by 10 estimates via weak channel checks. iPhone alt-data is notably noisy.

Roch MKM sees continued Celsius revenue growth slowing (like everyone else), but is hopefully that forward-looking commentary being celebrated by investors.

Jefferies sees Meta beating results and $30+ in 2026 EPS as potentially in reach. That would represent 22% EPS compounding from now to then at a massive scale. Bank of America sees Meta delivering beats this quarter too.

Jefferies also sees the setup for Alphabet as the easiest among mega-caps and strong advertising and cloud trends within its channel checks. Speaking of the Search King, Waymo closed a $5.6 billion funding round.

James Gorman was named as the new Disney Board Chairman. Disney cut Blade from its 2025 film schedule to focus on higher quality projects. It’s these decisions that have led to its last several releases being fantastic hits. Focus on quality. Disney also pulled its apps from the Apple store to avoid the 30% fee paid to that tech giant.

Consumer and Employment Data:

Output Data:

Inflation Data: