$1.85-$1.95 billion in annual OpEx was technically reiterated, but it said that it would likely end up near the high end of this range. That’s all we get for guidance. Its business is too reliant on trading volumes to expect it to be able to effectively guide for revenue. It would be a crapshoot. Smart from them.

EPS is expected to double this year and fall by 3% next year.

As announced at the launch event, Robinhood debuted Robinhood Legend as its new desktop mode during the quarter. This is to drive greater product parity with other trading-oriented interfaces and win over more active traders outside of the mobile market. Whether it’s index options, futures or realized profit/loss tools coming soon, Legend is a large step up in the power of its desktop offering.

“Yes. I would say that the folks that use the Robinhood Legend product tend to be among our most active. So the velocity is quite high.”

CEO Vlad Tenev

Robinhood Gold growth remains solid. Through lower contract fees, IRA matches, free margin up to $1,000, 3% cashback for its nearly 100,000 Gold credit card holders, this is doing well. Notably, revenue for the quarter was reduced by $27 million from paid account funding matches from Robinhood to win over new users. This more than doubled Q/Q to 4% of total revenue. The 3% cashback perks and large bonuses for switching are wonderful for competing in a commoditized sector. But? It’s also very expensive when Robinhood is paying $200,000 for certain large accounts to switch over. So far, leadership thinks payback periods are compelling; it’s confident in making up the difference via a 7x lift to AUC and a 2x lift to deposits that Gold members provide vs. non-members. It needs to be right.

“We expect contra revenues to grow sequentially by a similar amount in Q4 and then grow much slower in 2025.”

CFO Jason Warnick

Back to the Robinhood Gold credit card for a moment. App store ratings are sky-high so far and customer spending behavior and measured deposit uptick are both in line with expectations. This is immensely important, as this is what Robinhood is relying on to pay for the 3% cash back reward and inevitable charge-offs. Finally, most users have it “at the top of the wallet. Robinhood is going very slowly with the rollout despite increasing confidence in the product, which is the correct decision. Losses can pile up rapidly if underwriting algorithms are rushed. The waitlist to get a card reached 2 million users this quarter.

“If someone has the gold card, they love it. If they don't have it, they want to know when they're going to get it. And I hear you. We're working hard to increase the rollout, but we're also being patient and carefully studying customer behavior as we grow so that we manage credit risk to profitably scale over time.” – CEO Vlad Tenev

Robinhood retirement accounts reached $9.9 billion in total AUC during the quarter. Quarter-to-date, it has already added another $1.1 billion to reach $11.0 billion. Good momentum. It crossed $200 million in transfer bonuses paid out to new accounts during the quarter. Again… really good for winning consumers and really expensive.

For overall deposits, growth remained strong at $10 billion, which represents 29% Y/Y growth. This was the first Q/Q decline in net deposit adds in a year, but the result is still undeniably solid. New cohorts are depositing 3x the amounts that cohorts 2 years ago did.

“Net deposits are north of $4 billion and cash sweep balances are more than $25 billion so far through October. As for trading, equity notional volumes had their highest month in over 3 years. Option contracts look to be having one of the highest months ever and crypto notional volumes are over $5 billion, which is on track to exceed the Q3 monthly average.”

CFO Jason Warick

I’ll split this into my views of the quarter and then the business overall. The quarter really wasn’t bad. The ugly stock reaction comes after explosive year-to-date gains that likely just needed to be digested. Very small misses vs. consensus were an easy excuse to justify that. Margins continue to head in the right direction, growth remains rock-solid, its balance sheet remains a large strength and it continues to attract more users at higher lifetime value. As a non-shareholder, I’m also excited to see what kind of impact Robinhood Legend can have on its retail desktop marketshare. It has been the largest share-taker in mobile for years… so why not desktop too? We shall see.

Now for the investment case and company. I hesitate to assume that this brokerage business can morph its suite into the kind of offering that can overcome inherent industry cyclicality. There’s a reason why sell-siders expect EPS to fall Y/Y next year — its results are mightily tied to asset volatility and trading volumes. I’m also skeptical that its large deposit and asset transfer bonuses can be a sustainable way to profitably grow users while expanding margins. It has done admirably in proving me wrong thus far, and I truly hope that continues for the bulls here. Finally, I was also worried about the credit card’s 3% cash back and potential for mounting losses considering HOOD’s lack of underwriting experience here. I am very glad to see the leadership team going so slowly with the rollout. Great decision.

All of this is to say that this is not my favorite investment. My thumb is pointed firmly sideways. I applaud the team and company for doing as well as they have over the last several quarters, but question whether this momentum can prove to be structural, rather than temporary. Hopefully, in a few years time, optimists will be able to come back to me and say “I told you so.”

In a letter sent to SMCI this past week, its auditor, Ernst & Young (EY), had some pretty damning things to say. EY wrote that they’re “no longer able to rely on management or the audit committee’s representations” and they’re “unwilling to be associated with the financial statements.” Hard to get much worse than that. While the valuation is dirt cheap, its liquid cooling niche is interesting and growth is rapid, I personally wouldn’t touch this with a 10 foot pole. The same team that ran into accounting and revenue recognition drama years ago is still there and now pulling the exact same stunts today. As public market investors, we rely on the honesty and reliability of a management team’s disclosures to gauge an investment’s appeal. If they can’t be trusted, which in my view is the case here, nothing else matters. I think getting delisted is more likely than recovering for the company.

CFO Susan Li hosted a Meta earnings follow up call for sell-siders as she always does. Here were the highlights.

In addition to being happy with overall Meta AI adoption, Li added that the product is especially popular in India thus far. On the earnings call, Meta talked about building consumer scale before monetizing its Meta AI work with AI agents like lead generators, custom service automations etc. As I spoke about in the review, this opportunity may not be very far off. It integrated Meta AI into click-to-message ads on WhatApp. The focus early on is in Asia, where it’s currently in experiment-and-learn mode.

My earnings review focused a lot on GenAI’s impact on engagement and monetization. We got some more examples during this call. 

During the quarter, it consolidated recommendation algorithms on Instagram to narrow its focus and “drive higher engineering efficiency.” The model work that Meta has done on Reels and the switch from general compute CPUs to high performance GPUs to run that content matching have yielded large engagement gains for the company. It’s now evaluating every other part of its apps business to see where these same changes can be implemented. There’s a long way to go in terms of optimizing engagement and monetization through constant, GenAI-inspired improvements.

Meta is the only mega-cap tech name that uses a 5-year depreciation schedule for its short-lived servers. All others use 6 years, which means they incur lower apples-to-apples depreciation expenses than Meta does. Switching to 6 years for Amazon, Microsoft, Alphabet and others was a large source of short-term profit outperformance and this is a future profit lever that Meta can pull whenever it feels it needs to. There are no plans to do so, but Li hinted at being open to this idea.

Meta is seeing strong adoption for its Advantage + campaign building automation suite across shopping campaigns, app campaigns and soon-to-be lead generation campaigns.

Celsius bought Big Beverages Contract Manufacturing for $75 million in cash on hand. With the purchase, Celsius gets a 170,000 foot manufacturing facility and broader control of its supply chain and manufacturing. It will also give the beverage company more opportunity to innovate and run limited time offer products.

Big Beverages Contract Manufacturing has been a long time co-packing partner for Celsius so the integration should be seamless; the facility will just keep making the same thing. The vertical integration is expected to drive input cost savings and operating leverage over time. It also provides easy capacity expansion opportunities as Celsius continues to grow. 

This makes sense to me. Celsius is working hard to preserve margins amid the weak backdrop of declining convenience store traffic and negative overall sector growth. Owning a larger piece of its supply chain, without diluting shareholders or raising debt, is a great way to do so. Importantly, this does not resolve any piece of the Pepsi inventory right-sizing headwind for Celsius. This simply means the product that Celsius is delivering to its main North American distribution partner will be made internally more frequently.

By segment, home and renters insurance IFP rose 15.9% Y/Y, pet IFP rose 53% Y/Y, car IFP rose 2.6% Y/Y (still working on underwriting models), Europe IFP rose from $8 million to $19 million Y/Y.

Note that gross loss ratio (GLR) is Lemonade’s total losses stemming from its book of business. Net loss ratio considers losses after the impacts of its reinsurance contracts.

One more note on the EBITDA guide. Lemonade enjoyed outperforming loss ratios and CAT impacts during Q3. As Hurricane Helene happened days before the end of calendar Q3 and Milton happened in October, it wanted to take a “conservative” approach to loss ratio modeling and not assume that its outperforming strength would carry into Q4. That was prudent and, to me, simply positions them to continue outperforming (lower loss ratio = more profit).

“For Q4, we do have our typical conservative assumptions on loss ratio.”

CFO Tim Bixby

Lemonade reiterated its path to positive EBITDA at some point during 2026. FCF should come a few quarters before that, thanks to favorable working capital dynamics. It continues to expect the net cash flow inflection to be permanent. The chart below uses its enterprise value to gross profit (EV/GP) multiple because it has no EBITDA (or EBIT or net income). As you can see below, the multiple has gone from nearly 2x the market to below market in two years.

The theme of the report was “resilience.” For over two years, Lemonade has proactively moved away from high catastrophic event (CAT) risk geographies and products. 

Between slow regulatory premium hike approvals and its older underwriting models writing policies that its newest algorithms wouldn’t touch today, it had some work to do. It needed to let low-quality plans (especially in home) roll off the book to diversify its business. Candidly, it still has more work to do on getting home policy algorithms to a point justifying that segment’s re-acceleration. We’re not there yet, but it’s heading in the right direction (see loss ratio data below); this area of insurance is too large and too important for Lemonade not to be playing here longer term. Near term, it has done well to use 3rd party funders for these policies to ensure it’s still saying yes to more customers, widening top-of-funnel and driving more cross-selling. Eventually I expect Lemonade to take more of the balance sheet risk on its own.

For now, it has pivoted away from these higher risk policies while its algorithms mature. This quarter made it clear why this patience and the coinciding revenue headwinds were worth it. Intentionally slowing growth wasn't a popular move, but it was the right decision. And? It shows this business is focused on the right things… not getting too aggressive just to beat a revenue number for a quarter.

Since it began this proactive pivot, IFP from low CAT risk areas (Europe, Pet and Car) moved from 23% to 44% of its overall business. The GLR impact from this quarter’s hurricanes would have been 40% larger without this. There was significant concern about weather events this quarter blowing up the firm’s strong balance sheet. I’d be lying if I said I wasn’t also a little nervous, which is why I hadn’t been adding to my existing stake. This performance clearly shows that Lemonade is growing up before our eyes. Loss ratio performance is very impressive and that’s hard to overstate. 

The company was asked about the threat of richer incumbents eventually matching its slick, automated, lower-cost processes. Their comments mimicked what they’ve consistently said in the past: Their single, AI-native, interoperable, light weight, constantly learning core operating system (Blender) is unmatched:

“Legacy insurers have been outsourcing data science and AI work for over a decade, yet consumers haven't noticed major changes… since their systems don’t run entire end-to-end processes, they're forced to work with dozens of third-parties. Creating a seamless AI-powered experience that delivers personalized customer interactions, improves efficiency and drives better underwriting and pricing requires a unified full stack system with AI at its core. Unlike traditional insurers, the Lemonade platform was designed, built and maintained in-house, and I believe this is our secret weapon. Blender, our insurance operating system, integrates everything… I believe this gives us an unfair advantage.”

Co-CEO Shai Winninger

Lemonade is a 9-year-old company with less than $1 billion in total insurance premiums. Considering this, the company has far more control over its growth and profitability than anyone should expect. As the pandemic bubble deflated, investors demanded that Lemonade expedite its path to profitability. At the same time, rampant inflation and very slow premium hike approvals from regulators meant previously profitable business was no longer compelling. 

So, when inflation sky-rocketed, insurer bottom lines suffered. Because of this (and the aforementioned need to cut CAT risk), Lemonade made the calculated decision to pull back on growth spending, right-size its operating expenses and boost its pace of leverage. For years, it has stuck to a strict schedule to profitability and has even expedited that schedule in recent quarters. Again… it’s not normal for a tiny, brand new company to know they’re going to be profitable on X date years down the road. And again, executing on that plan required significant scaling at favorable cost to provide the leverage needed to make money. That was FAR from certain… yet it’s delivering.

And now? Cost of capital is falling, Lemonade’s large cash pile is no longer shrinking, higher rates have been approved and it boosted growth spend meaningfully this quarter. Right on cue, Lemonade delivered the $10 million premium growth outperformance amid a re-acceleration in growth spending.

Loss ratios improved Q/Q and Y/Y for every single product. It attributed this strength to the “tight feedback loops from its LTV AI models,” which enable rapid iterating and improving, as well as aforementioned rate approvals. Lemonade was especially encouraged to see a 12 point improvement in its car loss ratio, which is providing more confidence that this is a key future growth outlet. We got more granular disclosures for loss ratios this quarter. Note that lower is better for all metrics in the table below.

As Lemonade revs the growth engine once more, investors should not expect margins to start to sour once more or for cash burn to resurface. It will get more aggressive on growth while sticking to its path to profitability (reiterated this quarter). The answer to how it's threading this needle is encouraging.

While growth spend did grow from $13 million to $40 million Y/Y and led to 27% Y/Y OpEx growth, OpEx excluding this line item has now been flat for two years while the business scales with fewer people and lower fixed costs. This is the power of building a digital-and-AI-native insurance company. Lemonade leans on R&D and constant algorithm improvements to automate any part of its business that can be automated. It continues to raise the proportion of fully automated onboarding and claims processes, which means lower call center reliance and a higher margin ceiling… not to mention its lack of insurance agents to pay commissions to. It expects the relative efficiency and automation leads to sustain and blossom into a durable cost moat.

As an aside, its growth spend is also 80% financed by General Catalyst (GC), which means it’s not funding it with cash on hand. GC does get a piece of the profit pool, but this arrangement will likely be temporary. As Lemonade grows more comfortable with underwriting, its liquidity and its scale, it will likely no longer use arrangements like these, which would raise the overall potential margin ceiling. GC has been instrumental in plugging near term cash burn issues, but those issues are now quickly going away.

Home Insurance is on track to shed $25 million in low quality premiums by year’s end. About $15 million of it was removed in Q3.

Leadership called Auto Insurance progress “remarkable” (partially thanks to needed regulatory premium increase approvals) and is confident that the product is “positioned for rapid growth.” Loss ratios are dropping and Lemonade is nearing readiness for primetime. That will commence in 2025 when it gears up to launch in several more states. 

This was the quarter when Lemonade’s investment case could have been dismantled. It was the period in which its balance sheet and underwriting could have been put to the test. Instead? It continued to grow its cash pile and improve loss ratios across all products. As I said, this quarter is Lemonade’s coming out party. It’s the firm’s “well maybe this company is actually real” quarter. It’s the quarter where Lemonade showed the world that it’s quickly growing up.

That said, this is still a very young company in a sector that requires large scale to profitably compete. It’s also living in a rapidly evolving world in which building that scale as quickly as possible matters. So? This team has not taken the bootstrap approach to growth. This is a classic venture capitalist-funded darling, with the mindset of sacrificing near term margin to grow its book large enough to generate sustainable profitability as quickly as possible. Its pristine balance sheet makes that a sacrifice worth making and so does the potential upside if it executes. This is by far the most speculative name in my portfolio and the only name without EBITDA or free cash flow (FCF).  It remains risky as all companies do before they prove their models are durably profitable.

Starbucks already pre-announced very ugly results that we covered here. The focus now is on what new CEO Brian Niccol will do going forward. For context, Niccol is essentially doing a total pivot on every strategic initiative the old CEO was running. It will take time for all of this change to bear fruit, and for now, financial results are not super important.

“Traffic declined across all channels and dayparts with the most pronounced decline in the afternoon daypart.”

CFO Rachel Ruggeri

Niccol reiterated a lot of what he said in his opening letter. Starbucks has drifted from its core, struggled with throughput and product consistency, tightened marketing focus too much, leaned too heavily on discounting, fostered channel clutter and created irrational menu complexity. Essentially everything that the old team has worked on besides Siren Craft (more later) is being entirely undone by Niccol. To that I say “go for it”. He has a clear track record of fixing similar issues at Chipotle and a view that “all of the issues are fixable” here too.

Starbucks is going back to a lot of the old practices that made its brand more personal. It will resume hand-delivering all beverages and will re-implement barista sharpies to add more personal touch. It will also serve in-store beverages in ceramic mugs. The printed labels and mess of drinks sitting at the end of counters just aren’t good looks and don’t represent the brand that Starbucks wants to again be. Next, improving store ambiance and warmth is part of the roadmap. As part of re-thinking store design, it will intentionally slow the pace of new openings next year while it works through ideas.

Niccol is also determined to drive customer value through structural changes, rather than constant discounts that dilute the brand’s quality. Starbucks will eliminate non-dairy milk upcharges (personally celebrating) and is committed to zero price hikes through the end of 2025. It’s also greatly simplifying customized drink ordering on its app. Today, every customization option is available, which creates a mess of choices and longer processes. Going forward, customers will receive more simplified customization options based on specific items ordered. It will also offer suggestions of frequently ordered combinations. 

Along similar menu simplification lines, Starbucks plans to cut several food items that drive more waste and delays than demand. Astoundingly, old leadership abandoned the carefully built stage-gate process for new food items that Starbucks had developed over several decades. They decided to prioritize rapid roll-out of new items with little testing and little understanding of what would work. To that I say, “Are you kidding me?” A large-cap, blue-chip, iconic consumer brand willfully ignoring its massive set of data and picking new items based on hunches? What? Shockingly, this did not go well. The pace of newness is going to slow as Starbucks re-implements these same procedures. Niccol couldn’t have come soon enough.

Throughput improvements go hand-in-hand with improving the overall customer experience. Starbucks now has a clear goal of hand-delivering consistently excellent products to all customers in 4 minutes or less. 50% of its stores have already achieved this goal, with its top franchise partners routinely offering service in under 3 minutes. Again… the issues are all fixable. It just needed a lot more operational focus.

First, Starbucks is optimizing labor hours per store. It’s making good progress but there’s more work to be done. This is a big reason why it aims to become one of the best places to work in quick service, with a goal of retail leadership positions being 90% internally filled. Next, it will fully separate the mobile and in-store ordering processes and it’s working to modernize order prioritization algorithms to help here as well. This is the same playbook Chipotle ran under Niccol to raise its average unit volume beyond where anyone thought it could go. Starbucks will also bring back its condiment bars to take some of the customization burden off of baristas, allowing them to deliver beverages more quickly.

The one thing that Niccol isn’t abandoning is the Siren Craft system project. This simply refers to upgraded workflows and tools (like a new blender and food warming systems) that make service more efficient. Niccol thinks these ideas are good, and the implementation has just been too slow and disruptive to operations. He wants to make store renovations (which these systems often require) a lot quicker and cheaper, which could enable faster implementation. This is a big piece of realizing that sub-4 minute goal, and Starbucks will prioritize onboarding these systems at its lowest performing throughput stores first. Some of the throughput issues also don’t require total implementations of Siren Craft, which could mean lower costs.

Niccol wasn’t ready to comment on China. He called the competitive environment “extreme,” macro “tough”, and also explicitly mentioned being open to future partnerships there to drive growth. Outside of China, the international opportunity was a “pleasant surprise”.

Starbucks walked back previous annual cost savings commitments.

I am confident in Niccol turning this around. I’m not adding, as markets also seem very confident, considering the 30x forward earnings multiple and nearly 2x PEG ratio. But I am quite optimistic that Niccol will run an eerily similar playbook at Starbucks that made Chipotle so wonderfully successful. Give it time. He’s a superstar.

YipitData came out with new information Friday on how Lululemon’s quarter is shaping up. Encouragingly, it witnessed an acceleration in the struggling U.S. business to mid-single-digits Y/Y growth. This compares to 0% Y/Y growth last quarter. China has also materially accelerated quarter-to-date. This indicates that Lulu’s struggles are a combination of macro and inventory assortment, rather than structural brand decay. Good news (U.S. especially).

In other news, Lulu’s brand partnership with Fanatics is going well. The two worked together to debut NHL apparel and accessories (like the Lulu Belt Bag), with opening day sales called “explosive.” Many items are sold out and despite this only being available for 11 teams, the duo clocked 80% of NHL merchandise sales overall during the period. I would love to see this expand to the rest of the NHL and the other three major professional sports leagues in North America.

Mastercard beat revenue estimates by 1.4%. U.S. volume rose 7% Y/Y vs. 6% Y/Y last quarter. This was its fastest Y/Y U.S. volume growth in 7 quarters. Rest of world volume rose 12% Y/Y vs. 11% last quarter and worldwide volume rose 10% Y/Y vs. 9% last quarter. Really good to see a Q/Q acceleration for one of the biggest consumer spending players on the planet.

“Results were underpinned by healthy consumer spending, including strong cross-border volume growth of 17% year-over-year on a local currency basis… the macro environment remains supportive and continues to underpin strength in consumer spending. The labor market is strong, even if slightly below historically tight levels. Inflation has moderated… overall we remain positive about our growth outlook.”

CEO Micahel Miebach

Quarter to date, switched transactions are stable Q/Q and “spending remains healthy.”

Bank of America channel checks point to an outperforming quarter and a guidance raise for Duolingo this quarter. We shall see – alt data is often noisy.

Disney formed an AI and Mixed Reality business division.

Employment and Consumption Data:

Output Data:

Inflation Data: