The rest of the Microsoft review, Amazon & part one of the Apple review are all coming tomorrow. The rest of Apple, Starbucks and Cloudflare (maybe just part 1) will come Saturday.

In case you missed it:

1. Microsoft (MSFT) — Brief Earnings Snapshot

a. Demand

  • Beat revenue estimate by 3.4% & beat guidance by 3.7%.

  • Beat 34.5% foreign exchange neutral (FXN) Azure growth estimates & guidance by 4.5 points each.

b. Profits

  • Beat EBIT estimate by 6.8% & beat guidance by 7.3%.

  • Beat $3.37 GAAP EPS estimates by $0.28.

  • Beat FCF estimates by 21%.

c. Balance Sheet

  • $94B in cash & equivalents.

  • $40B in debt.

  • Slight Y/Y share count reduction.

  • 11% Y/Y dividend growth

d. Guidance & Valuation

  • Q1 revenue guidance beat estimates by 1.4%.

  • Q1 EBIT guidance beat estimates by 3.5%.

2. Meta (META) — Earnings Review

a. Key Points

  • Ad pricing pressure concerns throughout the quarter were noise.

  • Massive top-line outperformance for Q2 and the Q3 guide — not driven by FX favorability.

  • Expense growth will remain rapid in 2026.

b. Demand

  • Beat revenue estimate by 6% & beat guidance by 7%.

    • Family of apps (FOA) beat estimates by 6%.

    • Other revenue, which consists of WhatsApp Business Messaging and Meta Verified, beat estimates by 16%.

    • Reality Labs missed estimates by 4%.

    • This beat was not foreign exchange (FX) favorability driven. FX was actually less favorable than expected, so the beat is entirely driven by fundamental outperformance. There was virtually zero FX tailwind vs. Meta guiding to a 1-point tailwind.

    • 21.8% 2-yr revenue compounded annual growth rate (CAGR) vs. 21.8% Q/Q & 22.6% 2 Qs ago.

  • Beat daily active people (DAP) estimates by 1.5%.

  • 11% impression growth beat 7% estimates.

  • 9% price-per-impression growth beat 7.6% estimates. As we’ve covered in weekly articles throughout the quarter, there have been a few notes about weaker price-per-impression trends for Meta. That now looks inaccurate, as pricing significantly outperformed.

  • Beat $12.96 average revenue per person (ARPP) estimates by $0.69.

c. Profits & Margins

  • Beat free cash flow (FCF) estimates by 42%.

    • This is despite spending $17B in CapEx vs. $16.3B expected.

    • On the other hand, FCF estimates have been significantly lowered due to Meta’s aggressive CapEx plans. Still impressive, but they did beat a lowered bar.

  • Beat EBIT estimates by 20%.

    • OpEx rose 12% Y/Y due to 16% cost of revenue growth, 27% G&A growth and 23% R&D growth. Marketing rose by 9% Y/Y as well.

  • Beat $5.89 GAAP EPS estimate by $1.25.

d. Balance Sheet

  • $47B in cash & equivalents.

  • $28B in debt.

  • Share count fell by 1.5% Y/Y

e. Guidance & Valuation

Q3 revenue guidance was 6% ahead of estimates. This includes a 1% foreign exchange (FX) tailwind. It did slightly raise its OpEx and CapEx guidance for 2025, which were both slightly above expectations. For Q4, they expect slower Y/Y revenue growth than the 21% Y/Y growth guided to for Q3. This was as expected (consensus growth estimates for Q4 are 13.5% Y/Y).

For 2026, rising depreciation costs from infrastructure spend and more compensation from the research team they’re assembling will lead to OpEx growth being faster in 2026 than in 2025. This means at least $142B in OpEx for 2026 and is materially above $130B expectations. Also for 2026, it guided to roughly $30B in Y/Y dollar growth (same as 2025). This brings them to $99B in 2026 CapEx, which is far higher than $80B estimates. This is the only thing to pick on from the report, but with revenue and profit outperformance so significant, it’s hard to complain. They see a massive opportunity to spend productive dollars and raise their financial ceiling. Year after year, they’ve demonstrated a world-class ability to deliver fantastic return on investment (ROI) with these costs (outside of Reality Labs). There is no founder on the planet I trust more than Zuck to spend $99B in CapEx in a single year. While that’s a massive number, I say go for it. They’re elite capital allocators; they’re already showing how powerful of a tailwind AI can be for today’s top-line performance. Go. For. It.

Other guidance notes:

  • They plan to fund most of the CapEx next year from their balance sheet, but are exploring some financing options and partnerships.

  • Infrastructure spend will remain highly flexible. They can easily shift it to and from parts of the business as needs evolve.

  • They think new tax laws will reduce this expense this year and in future years.

  • It also sounded like daily active users continued to grow into Q3.

Meta traded for 27x forward EPS entering the report. Estimates are going to rise for 2025 following this report. EPS is expected to grow by 9% Y/Y this year (will be revised) and 11% Y/Y next year. 11% Y/Y growth next year could modestly fall. Maybe not considering how sharp revenue outperformance is. We’ll see.

f. The Call

Building a World-Class Research Team:

A lot of attention was spent on Meta’s recent hiring spree in pursuit of superintelligence (better than human). They have been shelling out massive, 9-figure compensation packages to some of the most talented researchers in the world. They got the ball rolling with a $15B investment in Scale AI, which entailed hiring their founder (Alexandr Wang) as Meta’s new Chief AI Officer and head of the Superintelligence Lab. And they certainly didn’t stop there. They hired GitHub’s former CEO, Nat Friedman, to lead its AI product efforts, added Xingjia Zhao (co-creator of ChatGPT and GPT4) as their new chief AI scientist, and pulled several others from Alphabet, Apple etc.

Why are they all coming to Meta? Well… the giant pay packages certainly help. But aside from that, Meta will offer them access to "unparalleled" compute, including the first gigawatt+ data center later this year. When pairing that with global distribution, Meta lets them shed productivity bottlenecks while giving them a fantastic chance to have their work used by the masses. That’s why.

Zuck thinks this specific team needs to be small and extremely top-heavy in terms of talent. Unlike ranking systems across FOA ads, he thinks it’s vital to have a “small group that can hold the whole thing in their head.” He thinks it’s more productive to have these people working more closely together vs. asynchronous work that’s routine for larger teams across the rest of its business.

There were also reports during the quarter that Meta was abandoning its open-sourced approach to close off this model work in an attempt to monetize it more meaningfully. Those were mostly wrong. No big changes to its mindset here. It will continue to open-source most of its work, while leaving the most advanced and proprietary algorithms for internal use. There may be a modest shift to closed-source, but that’s because "models are getting too big for practical use.” That means releasing impractical models do very little for traffic, while helping the competition.

AI Strengthening the Core Business and Delivering Present ROI:

There has been a lot of talk about AI spending being a waste and this technology being over-hyped in terms of financial potential. While that’s true for most companies, Meta is already turning these hefty investments into tangible, incremental financial success. While most of the monetization to date has happened within infrastructure and GPU-renting rather than software, Meta is quickly joining other leaders like ServiceNow and Palantir in terms of using this technology to bolster its core, software-based business. The large top-line outperformance this quarter was “largely thanks to AI unlocking” efficiency in its ad systems. AI is materially bolstering ad precision and uplifting its performance today. Their Core AI ROI is already “strong.”

In terms of ad system upgrades from AI, there were a few highlights. Its Andromeda algorithm for ad retrieval (picking the best options from millions of potential ads for each user) was upgraded during the quarter and immediately boosted ad granularity and Facebook ad conversions by 4%. From there, its new ranking system called GEM takes this curated list of recommendations and picks the highest-returning options for ad buyers. They boosted GEM’s ability to pull from previous ad conversions by 2x, which led to a 5% boost to Instagram ad conversion and a 3% boost for Facebook (on top of the retrieval upgrade gains). Its Lattice model (which consolidates smaller disparate models for various ranking tasks) raised Facebook conversions by another incremental 4%. These systems will continue to be constantly improved, with new features, usage of newer models and more improvements in ad efficacy. This is why ad pricing is so strong right now.

From a product perspective, its Advantage Plus campaign building automator made it easier and more intuitive to use this product when creating a campaign, which is raising adoption rates. And for its Advantage Plus Creative Suite, which automates campaign content creation and split tests different iterations to uncover the best one, it is now used by nearly 2 million advertisers. It’s also beta testing AI translation to 10 languages to bolster ad reach without adding cost of complexity. All of this should be especially popular for smaller businesses with more finite marketing budgets, as Meta takes the manual labor out of creating marketing materials and running campaigns.

  • WhatsApp ad product expansion is also starting to build. In the USA, click-to-message revenue rose 40% Y/Y as advertisers flock to this ecosystem. 

  • Debuted what it views as the world’s only incremental attribution tool. This is able to credit conversions one-by-one, to make measuring performance more precise and effective.

  • Introduced omni-channel ads to unlock in-store impressions. Based on early testing, this is improving return on ad spend by 15% for advertisers opting in.

More Ad Load:

Ads on Threads debuted during the quarter.

WhatsApp also introduced status ads, sponsored channels and channel subscriptions. For both of these launches, ad load will grow very slowly and likely won’t be large contributors to revenue until 2027. Meta always takes its time with monetization, and this will be no different.

Creating Better Experiences:

During the quarter, more AI work on FOA’s content recommendation systems boosted Facebook and Instagram time spent by 5% and 6%, respectively. For threads, adding large language models (LLMs) to its content ranking systems is leading to sizable engagement gains; it will borrow these learnings for its other apps next quarter. AI is also helping it fix app issues more quickly, as it cut Threads bug reports in North America by 30% since late last year.

And it’s not just matching users with more relevant content that’s helping Meta here. They’re also empowering creators to make better content themselves. Its AI video editing tools and its overarching edits app (for AI-inspired content tweaking) are both enjoying strong progress early on. All of this is contributing to Instagram and Facebook video engagement rising 20% Y/Y. They think this progress is simply the beginning.

Meta AI:

Meta AI is looking to get more personal to grow engagement among its 1B+ active users. This product is still more popular on WhatsApp than anywhere else – as it’s a natural fit for a messaging app – but engagement trends for its other apps are promising. They also think automatic translation for Meta AI will greatly diminish user friction and grow usage.

Reality Labs:

The smartglasses business continues to excel. Despite greatly boosting orders, they still can’t keep enough in stock to service demand. Growth keeps accelerating and popularity keeps growing, with new Oakley and Houston’s partnerships on the horizon. These glasses are designed for athletes, with longer battery life, a great camera, and a sleek fit. Growth for Reality Labs during the quarter was solely thanks to this business unit, as Quest sales fell Y/Y. Still, they remain confident in Quest, as engagement trends are positive and cloud gaming usage grows.

Other Notes:

  • No plans to enter the public cloud space, but they didn’t rule it out. They will have a ton of compute at their disposal. If some of it becomes unneeded, it’s easy to see how they’d start renting it out like Amazon,  Alphabet and Microsoft do.

  • WhatsApp ad proliferation will be a cost-per-impression headwind, as that app is more popular in lower monetizing countries.

g. Take

Awesome quarter. Meta is clearly showing you that their massive AI expenses are already beginning to pay off. Results were fantastic across the board, and this company has so many levers to pull to deliver a lot more growth for a long time. Threads and WhatsApp monetization… more core app engagement gains… Wearables… Meta AI… etc. 

Zuckerberg continues to flawlessly execute and deliver real GenAI value while so many of his counterparts fail to do so. He’s a world-class founder and remains seemingly as hungry as he was when the company IPOed. He is why I am so ok with massive expense guidance for 2026. He sees a once-in-a-generation opportunity and wants to capture it. With the core business in as good of shape as it is and margins as sky-high as they are… I say go for it. This man is an iconic leader who has been proven right over and over again. I think this will be more of the same. Sometimes, companies are just easy to own. This is one of those times.

3. Robinhood (HOOD) — Earnings Review

Robinhood is a next-generation brokerage business quickly expanding into core financial services like savings accounts, event contracts, index options and eventually so much more. It pioneered commission-free trading, has a slick user interface, and boasts a compelling subscription (under Robinhood Gold) that has continued to briskly attract more users.

For a full overview of all the product news announced at the March 2025 Robinhood Gold event, click here.

a. Key Points

b. Demand

  • Beat revenue estimate by 7.2%.

  • Beat transaction-based revenue estimate by 3.4%.

    • Options revenue beat by 8.6%.

    • Equity revenue beat by 7.6%

    • Crypto revenue missed by 1.2%.

    • Transaction revenue fell Q/Q due to lower crypto revenue.

  • Beat net interest revenue estimate by 15.2%.

    • Credit card net interest income rose from $6M to $13M Y/Y.

  • Beat funded account estimates by 1.5%.

    • Funded account growth was actually flat in April and May. It jumped higher in June via inorganic Bitstamp contributions of 520,000 customers. Without this help, Q/Q growth would have been ~1% Q/Q and net new accounts would have been 180,000 vs. 400,000 Y/Y.

  • Beat average revenue per user (ARPU) estimate by 6.3%.

  • Total platform assets rose 99% Y/Y and 69% Y/Y ex-Bitstamp M&A.

  • Annualized deposits rose 25% Y/Y.

More growth metrics:

  • Cash sweep rose 56% Y/Y vs. 48% last quarter.

  • Margin book rose 90% Y/Y vs. 115% last quarter

  • Equity trading volume rose 112% Y/Y vs. 84% last quarter.

  • Options contracts rose 32% Y/Y vs. 46% last quarter.

  • Crypto volume rose 32% Y/Y but fell 39% Q/Q (very cyclical business).

c. Profits & Margins

  • Beat EBITDA estimate by 22%.

    • OpEx rose by 12% Y/Y; adjusted OpEx + stock comp rose by 6% Y/Y. Great expense control.

    • Marketing rose by 54% Y/Y as it leaned back into this expense (as expected).

  • Beat GAAP EBT estimate by 28%.

  • Beat $0.31 GAAP EPS estimate by $0.09.

    • EPS rose by 100% Y/Y.

d. Balance Sheet

  • $4.2B in cash & equivalents (another $9B tied up via regulation).

  • No debt. Pretty balance sheet. It does frequently draw small amounts from available credit lines. This quarter that was $1M (which was already paid back).

  • 0.5% Y/Y share dilution.

e. Guidance & Valuation

As expected, Robinhood raised its annual OpEx guidance from $2.14B to $2.20B due to the closure of its Bitstamp acquisition.

In July, net deposits bounced back to $6B, as promotions worked as planned. This puts it on track to add $50B+ in net deposits for 2025. This was very important to hear, considering deposit growth did moderate this quarter. For April and May Y/Y net deposit growth was $7B vs. $8.8B in June. For the quarter, net deposits rose by just 4.5% Y/Y. This was partially market-related, which is why it’s so good to hear about a great July.

  • Crypto, equity and options trading volumes are all at 2025 highs for July.

HOOD trades for 71x forward EPS. EPS estimates will rise following this report. Due to a massive tax benefit last year, EPS will fall this year. EPS is expected to compound at a 19% clip over the following two years. EBITDA is expected to compound at a 30% clip over the next two years.

f. Call & Investor Materials

Winning Active Traders:

A core priority for Robinhood remains winning the active trader. These users are quite positive for overall engagement levels and Robinhood’s transaction volumes. In pursuit of this objective, they rolled out Robinhood Legend (desktop version) to the UK, with early momentum called strong. They rolled out futures, grew prediction market contracts from 1B cumulative to 2B cumulative Q/Q and began introducing their AI research assistant called Cortex. The first product for this launch is its “stock digests,” which explain reasoning for sudden share price movements and have 100,000+ users already. Index-level options, a key product for more sophisticated active traders, enjoyed explosive 60% Y/Y volume and, generally speaking, several other products continued to deliver growth that was just as impressive.

Where is all of this traction coming from? While they are enjoying some cyclical tailwinds at the moment, they’re adamant that this is more so related to a great pace of innovation. I think it’s a combination of those two things, with a slight skew towards effective product improvement. They keep sprinting and customers keep responding. Market share again rose Q/Q, as its net promoter score (a leading indicator for more market share gains) set a 4-year high.

Growing Wallet Share:

There are several products working to successfully grow Robinhood’s wallet share. First is Robinhood Strategies, which is its robo-advisor. Since launching just a few months ago, it has $500M in assets and 100,000+ customers. The Robinhood Gold credit card tripled issuance year-to-date to 300,000 and is enjoying rising spend levels per user. Credit trends are strong early on, and the company plans to keep accelerating distribution.

Robinhood Banking, where it will offer traditional checking/savings accounts and other tools like automated tax filing, was released internally to its employees. They are working on the finishing touches before launching this later in the year. More account types will be introduced following the initial rollout. Its retirement account progress is also very good. Assets rose 118% Y/Y, with accounts up 80% Y/Y and funds per account up 22% Y/Y. For signs that all of this is working, funds per average account crossed $10,000 for the first time ever during the quarter and rose 2x Y/Y.

Tokenization & Crypto:

The majority of the Q&A was spent on Robinhood’s tokenization announcement at its last product event. Coverage of that event, which delivered a compelling EU engagement acceleration, can be found in section 3 of this article. A lot of this quarterly report was a review of that. The company reviewed the phased launch of this product, with public equity tokens eventually being offered on its own blockchain (thanks to buying Bitstamp) and enabling 24/7 trading and expansion to other asset classes like private equity and real estate. 

 Public equity tokens will be a key unlock for international growth, as those customers routinely want to own U.S. stocks without crazy fees. The other asset classes, I think, will be meaningfully additive to Robinhood’s opportunity in the USA and everywhere else. As leadership put it, many companies are building blockchains for “degenerate traders.” None are building for more responsible, constant access to a diverse array of private and public asset classes. That’s what Robinhood is creating with Bitstamp vertical integration.

  • They also added crypto staking during the quarter, with $750M staked in the product’s first month.

Go Global:

Robinhood will soon add perpetual futures in the EU and eventually the USA when regulation permits it. The company doesn’t currently offer equity shorting, so this should be highly incremental to the overall business, as it allows users to take bearish positions on certain assets.

  • It’s buying WonderFi to expand into Canadian digital assets.

  • Planning to soon launch in APAC and “3 more markets.”

Robinhood Gold:

Robinhood Gold maintained its rapid growth while delivering stable Robinhood benefits from subscribers. For subs vs. non-subs, assets remain 5x higher and retirement account adoption 4x higher. And while a 13% adoption rate seems good, it could go much higher. Its newest cohort opted into Gold at a 35% clip.

Crypto take rates:

Robinhood take rates on crypto are very high and continue to climb. As competition rapidly enters the space, some feel there could be some compression there. Considering it’s about 17% of total revenue, that’s a risk. Robinhood doesn’t seem concerned and thinks it will out-innovate the others entering the space. Still, it also said the trend of rising take rates for this asset class will probably pause at least for now.

Other Notes:

  • They have plans to eventually add more lending products.

  • No plans for a banking charter. They think they can offer everything they need to through partners.

  • They think their business is far less cyclical than in 2021. They’re right. The banking products, credit card and Gold are all less cyclical than their transaction revenue. At the same time, while they are less cyclical, they’re still very cyclical in my mind. Crypto and options still represent 42% of total revenue. And both of those businesses will be violently cyclical.

g. Take

Very good quarter. It’s not surprising to see the stock shrug off the positive numbers, as it has been on a tear for 3 months. That always elevates expectations and makes it harder to positively surprise analysts. That’s not important for investors. The actual data is what’s important and that remains quite positive for this company. The only thing to pick on is a lack of organic funded account growth Q/Q and slower deposit growth in April and May. It’s very easy to see how the deposit note was due to macro and trade drama, which is why it’s great to see things rebound in July. They keep taking share, expanding margins and expanding to new product categories with immediate traction. 

All I will say is that 40%+ of their revenue is still from crypto and options trading. We are in a very fun part of the cycle for those asset classes, which is surely supporting this business to a certain degree. It will be interesting to see how they fare during the next downturn. They do have a lot more products, but the revenue skew remains precarious. Until that comes, this business should perform very well.

Reply

Avatar

or to participate