The slow season is over and earnings season starts next week! There are about 40 reviews to catch up on from last season, and I cannot wait to publish 40 more in the coming weeks.

To access my current portfolio & performance vs. the S&P 500, click here.

I’ve been doing weekly podcasts with two colleagues. I’m going to start sharing the links to those for investors who want to listen/watch. The latest episode can be found here. Topics include Tesla, Google, SoFi and Nvidia. The 2nd most recent episode can be found here.

Table of Contents

• 1. SoFi (SOFI) – Miscellaneous News

• 2. Meta (META) – Wearables & AI

• 3. Mercado Libre (MELI) & Nu (NU) – Tariffs & More

• 4. Alphabet (GOOGL) – Bullish Search Data & More

• 5. Amazon (AMZN) – Prime Day Data

• 6. Lemonade (LMND) – Indiana

• 7. DraftKings (DKNG) & Flutter (FLUT) – FanDuel

• 8. J.P. Morgan – Data

• 9. Analyst & Credit Rating Notes

• 10. Macro Data

1. SoFi (SOFI) – Miscellaneous News

Anthony Noto conducted a brief interview with CNBC this past week. In it, he shared excitement about expansion into private market investments and how the new federal budget will support continued normalization in student loan refinance demand. There’s zero additional forgiveness coming and borrowers will need to make payments this year. This past week, the Department of Education announced that interest accruing for many borrowers and monthly payments for many more will resume in a few weeks. It’s so easy to see how they’ll flock to refinancing options to lower their payments. And? SoFi has a 60%+ market share of this refinance demand within its credit buckets. Most Noto-bly (sorry, couldn’t resist), he reiterated expectations for 30%+ member and product growth going forward. That bodes very well for Q2 results and updated 2025 guidance coming in a couple weeks.

Next, SoFi added more investment funds that provide access to private companies like OpenAI, SpaceX, Epic Games and many more. The funds include Cashmere (high-growth strategy), Fundrise (real estate and venture capital access) and Liberty Street Advisors (partners with seasoned fund managers to provide structured private market access). This follows a Templum partnership that initially unlocked access to high-profile private firms like SpaceX late last year.

In other news, SoFi’s GenAI banking assistant (Konecta) is yielding encouraging results for customers. Per a new press release, clients are enjoying 65% faster response times and 50% lower chat abandonment rates. That’s how you create real value and drive real traction for your products. Monetization follows utility.

2. Meta (META) – Wearables & AI

Meta bought a nearly 3% stake in EssilorLuxottica for $3.5B this past week. The two are close partners on the Ray-Ban Sunglasses, and now they’ll presumably get even closer. Direct ownership in the company provides Meta more flexibility to iterate more freely and frequently. It should expedite the product roadmap via improved collaboration and asset sharing between the two companies. Bank of America came out with a note saying large traction for this hardware (which is already clearly building) is probably not priced into the stock and could yield multiple expansion as it comes. They’re excited about that and, like me, also excited about Meta creating less dependence on Apple hardware for its own distribution.

Meta is also reportedly using Antrhopic’s Sonnet models for coding instead of its own Llama models. This offers more evidence of Meta perhaps falling a bit behind in the model race. That’s why it’s spending so aggressively to hire world class talent away from Apple, Alphabet, Anthropic and OpenAI and why Alexandr Wang was put in charge of Meta’s AI endeavors. Along those lines, it just hired the head of Apple’s AI teams (Ruoming Pang) for a $200M comp package. I’m sure the majority of that is tied to equity and performance. Some thought this meant he’d be getting an annual check for $200M and that’s just not the case. I continue to think LLM providers will constantly leapfrog one another over the coming years. Right now, X and Alphabet are considered the leaders, with X’s brand new release leading ranking boards and Gemini’s upcoming release most likely reclaiming that lead. I’m quite confident that Meta will continue to effectively innovate in the space as well. I view modest Llama 4 disappointment as an anomaly, with the company’s obsessive founder determined to make sure it doesn’t recur. Not ideal but motivation for a somewhat crazy and world-class founder/CEO is a beautiful thing. This is not overly concerning to me either.

3. Mercado Libre (MELI) & Nu (NU) – Tariffs & More

There’s a new 50% tariff looming for Brazil that was just announced by the U.S. administration. From a direct point of view, Mercado Libre and Nu are not vulnerable to these taxes and their cost structure won’t materially change. Nu has no import/export business and so will not be paying these tariffs if they remain in place. Mercado Libre does not source goods from Brazil to export to the USA. They have one USA fulfillment center, but that’s used to service U.S. sellers shipping goods to Mexico specifically… not the other way around and not predominantly to Brazil (which matched the USA’s 50% tariff). From this perspective, the news is irrelevant for both companies.

But? There is a potential indirect headwind. Tariffs could lower demand for Brazilian goods, as well as impact that consumer, economy and currency too. Neither of these companies is immune to worsening macro, so both would suffer to a certain degree. For context, total Brazilian exports to the U.S. in 2024 were roughly $40 billion, with the total economy at $2.33 trillion.  That is a roughly 2% hit if the entire export cohort vanishes (which it won’t).

At the same time, that suffering would likely be ephemeral if these new tariffs even stick. The companies would endure slower growth, but would keep taking market share, fortifying their moats and preparing for faster expansion as soon as macro once again brightens. Tariffs would not impact the long-term investment case or the relative value propositions for these companies, and I’d likely use any material sell-offs stemming from them to add to these positions.

Separately, S&P Global upgraded Mercado Libre credit to BBB- yesterday. MELI is officially investment grade in the eyes of that firm with a stable outlook. That will be great for its future cost of capital.

4. Alphabet (GOOGL) – Bullish Search Data & More

Oppenheimer published a bullish survey on Alphabet’s AI Search products as part of an outperform rating reiteration this week. The survey included 1,416 U.S. adults and focused on their usage of and satisfaction with AI Mode. The results bode very well for Alphabet:

• 60% of respondents found it more valuable than ChatGPT.

• Of the 263 respondents who are paid ChatGPT users, 75% of them find Google’s free AI Mode to be more useful.

• 53% of respondents click an AI Mode link often or very often. Great news for monetization.

• 65% of AI Mode users are downloading Gemini.

• AI Mode makes it less likely for respondents to pay for an AI agent.

• ChatGPT and Gemini are by far the two most popular chatbot products.

Separately, Oppenheimer sees Waymo (self-driving cars) compounding at a 100% clip over the next 5 years to reach $12.5B in total revenue. That’s when that unit is supposed to breakeven from an EBITDA perspective. By 2035, it expects Waymo to be doing $25B in annual EBITDA, representing a high-single-digit percentage of total profitability.

Going back to search for a moment, OpenAI is releasing an AI web browser. Many think this will be a monumentally new product for Google Search. I say, “not so fast.” Like Perplexity’s similar offering, this will be built on Alphabet’s own open-sourced Chromium platform. Hard to imagine this will be all that disruptive to Alphabet’s core business, especially if the disruption is built on the exact same foundation and with far inferior access to data vs. what Alphabet enjoys. The company has fended off several search entrants in the last two years and has maintained robust and improving revenue and engagement growth despite that. I expect more of the same.

• Alphabet is up to 50% of company code written by AI vs. about 35% 2 years ago.

• The company will reportedly offer material discounts to the U.S. Government to secure a large cloud deal in the near future.

5. Amazon (AMZN) – Prime Day Data

My social media feeds were dominated by proclamations of Amazon Prime Day spending being down 41% Y/Y over the first day of the event. These posts were mainly published with complete arrays of capital letters and sirens, but while lacking needed context. The Prime Day event last year was two days long. This year? It’s four days. When you give people double the time to lock in discounts, there's a lower sense of urgency to do so at the beginning of an event. 

Let’s run through an overly simplistic example. Last year, assume Amazon did $1 in revenue for day 1 and day 2. This year, that $1 shrank to $0.59 for the first day. Assume revenue stayed at $0.59 in each of the next 3 days. That leaves us with $2.36 in event revenue +18% Y/Y. Still, by adding two days to the prime event, Amazon loses 2 days of normal sales in its quarterly results. Considering Prime Day volumes are usually about 7x a normal day, we should therefore reduce each day of revenue by $0.08. After doing so, we’re left with 10% Y/Y Prime Day event growth – right in line with annual marketplace growth expectations for 2025. That just doesn’t make for nearly as sexy of a headline but this is why the news was entirely shrugged off by the stock and why marketplace estimates for 2025 didn’t budge either.

6. Lemonade (LMND) – Indiana

Lemonade launched its car insurance offering in Indiana, crossing 40% U.S. population coverage for that product. Lemonade has 500,000+ people just waiting for car insurance to be offered in their home state. These customers already know and love the Lemonade brand and are clamoring for more. Not only does that mean more revenue as this rolls out, but it should mean high-margin revenue. There will be virtually zero customer acquisition cost for cross-selling as they’ve already won over the customers. Furthermore, these customers will come with extensive data profiles, which should make underwriting policies for them much easier.

I’m excited for this product to scale through 2026. Lemonade leadership is increasingly confident that the underwriting is ready for national expansion. If they’re right as usual, the opportunity for this company is about to get much larger. That’s a big reason why growth is expected to keep accelerating towards 30% Y/Y even as the revenue base gets larger.

Interestingly, the Indiana launch comes with something called “LoCo” or its no-code, “LLM-first” playground for AI application development. This will foster more freedom to seamlessly tweak products for faster debuts across states and make things like updates to its underwriting algorithms (through LoCo Core) and personalized plan pricing (through LoCo Rater) more malleable and real-time. This is the luxury of Lemonade being an AI-first company. It has one technological core from which products like LoCo are derived. It is not stitching together dozens of disparate tools and 3rd-party vendors to try to force things to work well together. It is one, lean and mean insurance machine… and that means faster iterating, faster innovating and faster underwriting improvements than everyone else. This is the true tech company in insurance and LoCo is simply another example of that being the case.

7. DraftKings (DKNG) & Flutter (FLUT) – FanDuel

Flutter purchased the 5% stake in FanDuel that Boyd Gaming held to secure full ownership of that gambling brand. The most interesting part about this news is that the deal values FanDuel at $31B, while DraftKings carries a roughly $21B valuation. In a key state like New York, FanDuel and DraftKings are roughly tied in terms of year-to-date gambling volumes. FanDuel is ahead for revenue, with $137M vs. $108M for DKNG, but that’s a byproduct of a higher hold rate and DKNG is quickly closing that gap. As that happens, there’s more upside to revenue growth than for FanDuel. 

Beyond this, DraftKings (including Golden Nugget) and FanDuel are neck and neck in terms of iCasino market share, with miles upon miles of legalization left to go and leaders in states like Illinois confirming legalization is eventually coming this past week. Finally, DraftKings has the largest online lottery brand in the USA with Jackpocket and FanDuel doesn’t have a competing offering there. When contemplating all of this, I think the valuations of these two assets should be roughly similar. Maybe one could argue the revenue lead for FanDuel in sports gambling should fetch a tiny premium (even though DKNG has more growth upside due to more expected hold rate progress), but certainly not anything close to 50%. Just another example of DraftKings… at 25x FCF and an 80% multi-year FCF CAGR… being grossly undervalued. Just keep executing and the stock price will take care of itself.

And finally, in other gambling-related news, there’s bipartisan support to reverse the recent decision to cap gambling tax deductions at 90% of total losses. Last week, I went into why I’m not overly concerned about that cap, but I’d still prefer it if it were removed.

8. J.P. Morgan – Data

J.P. Morgan will start charging companies like PayPal, (and Block, Affirm, Coinbase, Apple etc.) for access to customer data to facilitate transactions. When a J.P. Morgan account is used for an online transaction, companies like PayPal connect to that firm’s customer data to power verification and authorization. Now, doing so will become more expensive. This will be a modest margin headwind for all of the companies involved, but I don’t anticipate this having a drastic impact on PayPal profitability. The company has so many margin tailwinds raging right now and so much opportunity for continued cost cutting that I think they’ll be able to seamlessly overcome this with no sizable impact. I think things like Venmo monetization, Braintree margin prioritization and ongoing efficiency gains will be more than enough to make this unnoticeable.

9. Analyst & Credit Rating Notes

TD Cowen thinks “solid search and YouTube growth amid consumer resilience” will power an above-consensus Alphabet quarter. They see Q2 revenue coming in about 1% ahead of consensus, with profit more than 4% better than expected. Both Search and YouTube revenue are expected to each be 1% ahead of consensus as well. They also think Q3 revenue will be about 1% ahead of consensus, with profit 4% ahead. Optimism is based on rising traction for its AI search efforts, digital advertising resilience and “no major impact from tariffs.”

Last week Scotiabank reiterated an outperform rating on Datadog because they believed rising OpenAI revenue concentration risk was modest. This week, Guggenheim downgraded DDOG shares to sell because of this same risk. The firm sees OpenAI’s shift to cheaper observability vendors and in-house solutions as a pressing financial concern for this company. Guggenheim doesn’t think other customers pose the same risk, but that DDOG will need to more deeply discount OpenAI contracts to maintain a smaller portion of the business going forward. That could lead to Q4 revenue growth of 17% Y/Y vs. current estimates calling for 19% growth – as well as “a big hole to fill in 2026.” OpenAI contributes nearly $200M in annual DDOG revenue (about 4% of 2025 estimates) and the analyst fears that the chunk shrinking will foster 15% 2026 revenue growth vs. consensus estimates of 19% growth. Until then, Guggenheim sees Q2 results and Q3 guidance modestly above consensus. Datadog is becoming more of a battleground stock.  We’ll see who is right.

TD Cowen initiated SoFi with a hold rate and a $21 price target this week. They applauded the company's diversified business and growth vectors, but they see shares as fairly valued for now.

“Based on our understanding, OpenAI has already completed building and testing the internal observability solution onto which DDOG workloads will move.” – Analyst Howard Ma

RBC is upbeat on the payments space heading into Q2. They think PayPal, Affirm and Global Payments provide compelling upside vs. consensus. They also see Visa and Mastercard outperforming as channel checks continue to show a resilient consumer.

For Amazon, Morgan Stanley is gaining more confidence in an AWS acceleration. They raised revenue targets for that segment by 1% for 2025 and by 3% for 2026. The firm sees Anthropic’s contribution to AWS driving great inference demand growth and fostering that growth uptick. This comes as rumors swirl about Amazon boosting its overall investment in Anthropic by another few billion dollars. They also think non-GenAI growth will accelerate based on CIO surveys conducted over the last few weeks. Between this rising optimism, improving macro, observed e-commerce resilience and healthy advertising demand, they raised overall targets for 2026 and 2027. For 2026, revenue was raised by 2% and EPS by 9%. For 2027, revenue was raised by 3% and EPS by 6%. Finally, Morgan Stanley maintained Amazon as a top pick.

TD Cowen sees Meta beating Q2 consensus expectation. They think revenue will beat by 2% and EBIT will beat by 6.5%. They held above consensus 2025 estimates as is. Digital ad channel checks were as upbeat for Meta as they were for Alphabet and Reels traction across the family of apps was notably strong for Q2 – per the firm’s data.

Truist thinks Chipotle same store sales will come in at -3% Y/Y as expected. At the same time, it thinks the concerns over Chipotle’s growth engine will ease following this quarter. The institution observed accelerating growth through Q2 for the food giant. They called out a successful Adobo Ranch debut as contributing to improving momentum.

RBC thinks Chipotle will post outperforming Q2 results based on better-than-expected macro and demand. This optimism stems from channel checks involving 35 stores that enjoyed improving traffic throughout the quarter.

Bank of America reiterated an outperform rating for Uber and boosted their price target from $97 to $115. They think the new federal budget’s tip tax exemptions will lead to better supply dynamics for Uber’s driver network. BofA’s data points to a modest acceleration in bookings growth, which is better than consensus estimates calling for growth stability. Finally, they think Uber-backed Moove raising $1.2B to purchase Waymo cars bodes well for capital appetite to fund these moves. That could drive more industry fragmentation among fleet managers, which is good news for Uber the demand aggregator. Wells Fargo came out with a very similar Uber note calling for modestly outperforming Q2 results and upward revisions.

Needham channel checks with a large Zscaler channel partner (and customer) were quite positive. Partner engagement activity is improving as ZS finishes up go-to-market improvements. SASE momentum in the USA was called very good – despite materially higher pricing than Palo Alto. Competition abroad is a bit tougher.

Needham also did some demand checks for JFrog, GitLab and Datadog. JFrog’s demand trends were stable while things for GitLab are accelerating. Finally, respondents hinted at a strong preference for Datadog observability tools for observability and Elastic for search use cases.

Piper Sandler downgraded CrowdStrike to neutral. This is solely based on valuation, as the company had very positive things to say about this team and company. I think this is a very fair downgrade. As I’ve said in recent coverage of the company, the valuation is probably ahead of itself. I’d love to see some multiple contraction so I can finally start to add to my stake once more.

10. Macro Data

Very light week for macro data releases. As part of the updated Fed minutes, we learned that there are multiple voting members open to a rate cut this month. Other than that, initial jobless claims were 227,000 vs. 236,000 expected and 232,000 last month.