In an effort to continuously augment the value we provide, we’re adding a new section to News of the Week articles. We’ve purchased access to sell-side research reports from several different institutions, and will be summarizing the key updates in an “analyst updates” section every week. 

Needed Notes:

Needed Notes:

OpenAI is using Alphabet’s Tensor Processing Units (TPUs) for some of its ChatGPT compute needs. On Friday, news came out confirming that the cloud deal announced between the two earlier in the month includes usage of the tech giant’s high-performance chips. Along with ChatGPT gearing up to use AMD’s new GPUs during the 2nd half of the year, this marks the first and second instances of OpenAI diversifying away from NVDA’s chips. Similarly to Application-Specific Integrated Circuit (ASIC) chips, TPUs are custom chips, purpose-built for specific workloads, such as Google Cloud. ASIC players like Broadcom and TPU vendors like Alphabet continue to find more product-market fit for their hardware, more meaningfully complementing general-purpose GPU juggernauts like Nvidia and AMD.

The company launched a new, multimodal small language model (SLM) under the Gemma family called Gemma 3n. The 4 billion parameter is the first sub-10 billion parameter to record an LMArena Elo Score (popular ranking metric for models) over 1300. This should build on a whopping 160M downloads since Gemma debuted a year ago, as Alphabet just continues to convincingly lead the model innovation pack.

Similarweb published some alternative data this week on M/M growth for Alphabet’s AI-focused sites and on how much chatbot traffic is being directed to YouTube.

Uber’s charismatic founder, Travis Kalanick, was forced out of the company about 8 years ago. Considering this, one may be surprised to learn that Uber and Kalanick are exploring a purchase of PonyAI’s U.S. business. Pony, which is already an Uber partner, is a player in autonomous driving with a virtual driver software platform and a sensor business too. The AV player has an important manufacturing partnership with Toyota and this partnership could lock in a lot of future supply for Uber’s network.

Micron sells semiconductors for memory and storage. Its “Not And” (NAND) chips offer non-volatile data storage, which maintains stored information when a system’s power is turned off. Separately, its Dynamic Random Access Memory (DRAM) chips offer volatile memory storage for personal computers, data centers, and mobile devices. Volatile means that storage isn’t maintained when a system’s power is turned off. DRAM helps processors access needed real-time data to minimize processing latency. These chips are considered to be somewhat commoditized at this point, with Micron’s cost advantages providing its edge.

These chips provide the foundation for its solid-state drives (SSDs), which are used in computer data storage and things like USB flash drives. Micron sells standalone DRAM and NAND chips and also SSDs with these chips in them. SSDs replace hard disk drives (HDDs), as they’re more power efficient, durable and resilient. It provides basic memory cards for things like gaming devices and cameras as well.

Perhaps most interestingly, Micron offers a type of DRAM called high-bandwidth memory (HBM) used to enable massive data needs of GenAI. It sharply improves data processing capabilities and facilitates improved data sharing between CPUs & GPUs. Nvidia is a big customer, using Micron’s HBM in its Blackwell and future Rubin systems. It also offers higher-capacity SSDs to help with LLM storage.

As we work through this piece, keep in mind that Micron is a hyper-cyclical business model. Demand violently fluctuates as backdrops evolve and the macro environment changes. Margins do too, as pricing & utilization rates can experience hefty swings.

Micron beat revenue estimates by 5.1% & beat guidance by 5.6%. DRAM bit shipments rose 20% Q/Q and average selling price (ASP) decreased in the “low-single-digit percentage range” Q/Q. NAND bit shipments increased by roughly 25% Q/Q. ASP decreased “in the high-single-digit percentage range” Q/Q.

Compute & networking revenue rose 11% Q/Q thanks to a 50% Q/Q rise in HBM memory as well as general strength across DRAM. Storage revenue rose 4% Q/Q, mobile revenue rose 45% Q/Q and embedded revenue rose 20% Q/Q.

For the year, it sees DRAM bit (unit of memory) demand growing in the high-teens percentage range and NAND growing somewhere around 11%-12% Y/Y. Micron’s supply growth will lag this unit demand growth, which will be good for pricing. Over the medium term, it sees DRAM and NAND both compounding at around a 15% clip. With insatiable DRAM demand and NAND improvements driving more output per square foot, it is shifting some capacity to DRAM and will end up with a 10% lower NAND manufacturing footprint by the end of the year. Finally, it reiterated its $14B in annual CapEx guidance from last quarter.

All of this sector-level growth is supporting healthy customer inventory levels, which should support strong demand and pricing through the end of the calendar year. There could have been some demand pull-forwards related to tariffs, but if that activity happened, it was reasonably modest in nature. 

“We're not going to provide Q1 guidance. We are positive on the trajectory of the business. The market environment does remain constructive, particularly in DRAM versus NAND. We're focused on pricing and making sure to put our bets in the right places.”

CFO Mark Murphy

Micron trades for 11x forward EPS. EPS is expected to 5x this year, then grow by 54% and 13% in the following two years.

DRAM – Data Center & HBM Performance:

Micron delivered 100%+ Y/Y data center revenue growth thanks to continued rapid expansion for the HBM subsection of the category. Specifically, data center HBM revenue again rose by 50% Q/Q, as Micron continues to lead the pack in terms of memory efficiency and performance, while remaining the “sole supplier in volume production” for its low-power data center DRAM business. They’re dominating entire portions of the HBM market like this one, taking market share across all other pieces of the opportunity and quickly approaching HBM market share in line with its overall DRAM market share – despite being a later entrant into the HBM market vs. SK Hynix and Samsung.

Last June, Micron excited investors by saying supply was sold out and pricing was locked in for its HBM products through the end of 2025. They offered 18 months of visibility and have stuck to that every step of the way. This quarter, an analyst requested that same color through 2026, but leadership declined to offer it. When taking into account strong demand signals through the end of this year and the upbeat channel inventory commentary, that isn’t overly alarming. Two ideas for the reasoning:

“We are very much focused on addressing the '26 needs for the customers and executing well and remaining extremely focused on our execution for 2026.”

CEO Sanjay Mehrotra

Micron is poised to reach its market share goal for HBM “during the 2nd half of this year” vs. “by the end of the year” previously. I think this is a modest improvement to that timeline.

DRAM – HBM Tech Roadmap

Its latest HBM3E 12H chip is seeing “yield and volume progress extremely well,” as it expects shipments of this product to surpass its predecessor by the end of calendar 2025. From there, HBM4 is expected to deliver 20% better power efficiency than the previously-market-leading HBM3E 12H chip and 60% more bandwidth, keeping the pace of product improvement rapid. HBM4 samples have been delivered, with a production ramp scheduled for 2026.

“We have deep relationships with practically every major customer of HBM, and have earned their trust with our execution, delivering the world’s lowest-power, highest-performance HBM.”

CEO Sanjay Mehrotra

HBM4 is unsurprisingly built with inference scalability and reasoning model performance in mind. This continues to pull from Micron’s 1-beta node technology to power production. A node in this context is a manufacturing process. It’s important to highlight that all of these products are still based on its 1-beta DRAM technology. 1-gamma is quickly coming after 1-beta and has already been included in some smartphone end market shipments. This production change is already netting a 25% cadence boost to recommendation throughput and 20% better power efficiency vs. 1-beta.

It will be infused into datacenter manufacturing over the coming years, with 30% better bit density, 20% better power efficiency and 15% better performance than similar products built with 1-beta.

Manufacturing Footprint Updates:

Earlier this month, Micron announced a $200B investment over 20 years in U.S. manufacturing and R&D capacity. It added $30B to this plan to create a budget for its 2nd leading-edge memory factory in Boise. This facility will benefit from learnings at the first location and will share R&D capacity to create more cost efficiencies. The first Boise facility is expected to produce DRAM chips later in calendar 2027, with production from its New York facility coming after that. Finally, it’s expanding capacity at its Virginia plant and moving some advanced packaging capabilities to the USA to support HBM production in the coming years.

NAND notes:

NAND set new records for data center SSD market share and also client segment SSD market share. According to independent 3rd-party data, it’s now the 2nd-ranked company by market share in NAND data center SSDs. There are two important products to know here. First, the 9550 SSD is purpose-built to maximize data center performance. This recently finished key customer qualifications across several potential vendors, signaling more growth. Secondly, its slightly newer 6550 ION SSD is meant for higher compute intensity needs. These two represent “industry-leading performance and energy-efficient data storage for AI servers.” They are routinely a key complement to Nvidia and AMD GPUs, as well as Broadcom switches and ASICs.

Micron just announced a brand new SSD which quadruples performance vs. previous models.

Notes on End Markets:

The headline numbers were great. They’re rapidly growing, taking market share across key AI demand buckets, finding pricing upside to facilitate gross margin beats, exercising OpEx and CapEx discipline and, simply put, executing. Everything in these financials points to its GenAI/Agentic AI demand being red hot and that momentum continuing through the next 6 months. Some skeptics are picking on the lack of confirmation that HBM is sold-out through 2026, as that could point to early signs of this cycle finally slowing down. We’ll have to see how things progress into next year, as it doesn’t look like these preliminary signs will hold back demand at all for the next two quarters. The near-term continues to look quite favorable for this cyclical business model.

“While our financial results are in line with our expectations, they are not where we want them to be. Moving forward, we expect our business to improve as a result of the progress we're making through our Win Now actions.” – CEO Elliott Hill

For next quarter, Nike guided to roughly -5% Y/Y revenue growth, which was about 1.9% better than expected. Gross margin guidance was about as expected and includes a 100 basis point tariff impact (more later). When combining this with SG&A guidance and other items, EPS guidance missed by about $0.03-$0.04. They continue to prioritize investing to fix the brand and product assortment over near-term margin optimization.

For the full year, they again refrained from providing tangible targets. Still, they expect the inventory mess to be cleaned up in two quarters. Air Force 1 inventory resets are basically done, but it still has a lot of work to do with shrinking Dunk inventory on hand. More discounting ahead (especially on the lifestyle side of the business), which is why GPM will remain so challenged next quarter. As we’ll explore below, headwinds are now set to ease in the coming quarters, but won’t fully go away until later in FY 2026. That’s when sell-siders expect the company to resume low-single-digit Y/Y revenue growth and accelerate from there.

Nike trades for 48x forward EPS. EPS is expected to fall by 20% this year and then grow by 45% the year after.

Tariffs:

Nike expects tariffs to raise input costs by about $1B for FY 2026 and reduce GPM by roughly 75 bps. That led to 2026 profit estimates falling after the report, but revenue estimates rose, and investors seem to be paying more attention to that. That makes sense. It’s harder to re-ignite demand growth for a brand some view as decaying than it is to juice margins a little with some cost controls. Nike is focused on the former, and shareholders are rightfully confident in the latter coming if they can deliver a durable acceleration in top-line results.

The impact of the tariffs will be weighted towards the beginning of the year, as Nike has a few levers to pull. First, it plans to shift where some U.S. inventory is sourced. This year, it expects U.S. footwear sourced from China to fall from 16% of total to below 10% of total. It’s safe to assume Vietnam and Cambodia will be picking up a lot of that slack. It’s “partnering with suppliers” to reduce the consumer impact, which to me is a nice way of saying they’re demanding that supply chain participants absorb some of this price impact. From there, Nike is planning select price hikes in the USA and evaluating corporate costs that can be slashed. They anticipate recovering all of this incremental cost. That will just take time.

Changes:

Since Hill took over Nike a few quarters ago, the key theme of the turnaround effort has been “Win Now.” To accomplish this, Nike has embarked on a series of major changes that are beginning to deliver subtle signs of progress. Emphasis on “subtle,” but that’s to be expected. This isn’t a case like Meta in 2022, where the core business was a few straightforward fixes away from being healthy again and unproductive spend could be easily cut. This business was in significant decline when Hill took over, so expecting that to change overnight (or even in 3 quarters) is too ambitious in my mind. So, yes, subtle… but understandably so.

Perhaps the most important part of Win Now is placing sports and athletes at the center of every decision Nike makes. They’ve perhaps gotten too distracted pursuing other parts of the apparel and footwear opportunity and have strayed from their core. That has cost them dearly, and they’re determined to right the wrong. They call this “Sport Offense,” which essentially means further segmenting teams not only by brand and gender, but by specific activity too. Nike competes with a plethora of different companies across running, training, golf, basketball, soccer etc. It needed to organize teams with a more concrete understanding of competitive landscapes for specific sports, and now it has.

Beyond the reignited sport-centric strategy, they’ve made a series of other changes to reverse ugly financial trends. They’ve streamlined leadership and reporting structures, which they think is already leading to better assortment decisions and better product portfolio balance.

They’ve significantly reduced promotional days across their marketplace to make its digital business a more full-priced and “premium” destination. That’s not showing up in margins yet, as they’re having to aggressively discount elsewhere to work through the mountain of stale inventory from the last couple of years. It’s also not showing up because this channel is shrinking as a % of overall inventory, as the price hikes are still leading to traffic declines. It will take a few more quarters for comps to normalize.

Nike got too aggressive in chasing quarterly revenue targets with aggressive discounting and diluted brand quality as a result. Per the team, DTC channels are “showing early signs” of garnering that more premium reputation. The aggressive discounting also burned the trust of its key wholesale partners, which it’s determined to claw back. Not undercutting these partners at every turn, while unlocking higher-quality inventory for them are the two best ways it can do that. Hill is actively meeting with notable partners to voice this commitment to taking better care of them. It is no longer fixated on directing maximum revenue skew to its own channels. It simply wants good product placed wherever the customer is. Sometimes that’s on Nike.com… sometimes that’s elsewhere. More on wholesale later. 

Finally, Nike sought to greatly reduce large inventory gluts among classic footwear franchises, including Air Force 1, Dunk and a few others. This is where the most aggressive discounting is coming from. 

It is determined to rev the product innovation engine and accelerate the cadence of impactful product launches. They see some glimmers of progress, such as its holiday order book looking healthier than it did a year ago.

A Reminder on Impact From These Changes:

The most pronounced impact from the “managing down” of its classic footwear came this quarter. This is when the revenue headwinds were expected to peak and Y/Y revenue growth was expected to trough, with performance improving from here. Thankfully & vitally, that expectation was reiterated. This is still expected to be the worst quarter in Nike’s turnaround plan, with classic footwear revenue falling 30% Y/Y in Q4 vs. 20% Y/Y for the full year. Without this intentional decision to greatly pivot its product focus and right-size inventory, revenue would have fallen by 4% Y/Y, rather than 12%.

Brand Marketing:

A large part of rejuvenating Nike’s momentum entails leaning into marketing and demand creation expenses to take back control of its brand and storytelling. They are determined to communicate their commitment to making athletes the most important stakeholder in every decision. Relatedly, they’ve heightened prioritization when it comes to inserting the brand in major sporting events. It’s doing a better job of tying its brand intimately into these globally relevant occurrences.

Its Nike After Dark running event led to its best sell-through day at its location in The Grove (LA) since 2022, while its Breaking 4 running campaign with Faith Kipyegon was deemed a success too. Outside of internally cultivated sporting events, it used the wildly entertaining French Open final to rev its brand marketing engine and juice sell-through rates by 30% on that day. It was a central brand in Rory McIlroy’s (also wildly entertaining) Master’s win and at the Champions League Final. Speaking of soccer, it has 3 new cleats planned for the 2026 World Cup. And, in a way that only Nike can, it flexed its famous athlete partnership muscle with the new A’One shoe for women’s basketball.

All of this focus, and hefty spend, is helping its DTC channels recover premium positioning as it’s getting “into a better rhythm of aligning sport moments and key launches.”

“We’re investing heavily in big sport moments and key product launches to win back our brand voice.”

CEO Elliott Hill

Quick Notes on Product Categories:

More on Wholesale:

This quarter, Nike had a few interesting wholesale highlights. First, with Dick’s Sporting Goods, a series of training products facilitated rising overall Nike sell-through rates with that partner, while its Air Max 95 offering through JD.com did too. With Urban Outfitters, a Gen-Z focused program led to Nike reaching the top brand at some of its stores, while it added 200 Aritzia stores to further expand its wholesale presence. And finally, it is re-adding some merchandise to Amazon through a new (again) partnership with that giant. They have a planned brand destination within Amazon.com that will launch in a few months. While others like On Running are focused on growing with existing wholesale partners, Nike sees a real opportunity to do that, while also expanding to new retailers. It does have a wide array of products for a diverse set of occasions that can work across many different store concepts. To keep all of this progress humming, it’s adding new retail marketing, merchandising and account management talent to support its renewed focus on wholesale relationships.

Wholesale partners are reacting favorably to the new assortment and that positivity is leading to Q/Q growth in its orderbook and Y/Y growth for holiday orders among these partners. They’re “finding better balance” within the portfolio, and expect newness to help offset the large declines in classic footwear, while moving beyond the sharpest part of that demand headwind will help too.

While this was better than their last few quarters, we still can’t call these results anything but bad. At the same time, that’s as expected. This is what happens when you are correcting massive ordering errors within your largest financial drivers. And when looking beyond the surface-level numbers, there are some signs of subtle progress. These greenshoots are just not yet enough to overcome the core franchise headwinds. It will be a few more quarters before that happens, but there’s reason to believe things will look much better once we get to calendar 2026.

As I’ve been saying in recent months, I don’t love apparel or shoes as an investment category. I do find this to be more intriguing than most other options in the space from a potential fundamental turnaround perspective, but at 40x forward EPS and no EPS growth for another year, it seems like everyone in the world is already assuming that. I’d be interested in this at 20x EPS, not 40x. There’s too much speculating to do and too many bad quarters ahead before things look good enough here to justify that price tag.

Evercore ISI upgraded Trade Desk to outperform with a $90 price target this week. It cited improving satisfaction with the new Kokai platform, thanks to recent launches like Deal Desk. 

Beyond improving TTD execution, the analyst talked about channel checks pointing to outperforming ad spend for Q2. Finally, they excitedly mentioned many 2026 tailwinds, such as the ramping Netflix partnership, midterms, the World Cup, the Olympics and Microsoft sunsetting its competing Xandr business.

On the other hand, Wells Fargo downgraded TTD to neutral, citing Amazon-based competitive concerns. The company shared some notable surveys in which most respondents were shifting some level of budget away from their connected TV (CTV) spend with TTD. As I’ve talked about a few times now, I think it’s inevitable that a new entrant in the market like Amazon will win its fair share of the pie. I just also know that while Amazon is getting stronger, Alphabet is actively deprioritizing their own competing offering, which TTD has outcompeted for many, many years. 

And two more ideas here. First, digital advertising is a nearly trillion dollar market. Amazon taking some of that opportunity will in no way prevent TTD from finding durable growth. Its future is firmly in its own hands and the hands of a great CEO. Secondly, Amazon struggles with the exact same conflict of interest issue that Alphabet always has. It owns supply… yet wants to represent the demand side… and makes more money when impressions end up placed with its own properties. Green always says the world needs a scaled, independent demand-side platform that doesn’t possess that same conflict of interest. By not owning inventory, TTD avoids that conflict and can be that objective player. So, while I think Amazon’s ad business will keep doing very well, I think TTD’s value proposition is in no way threatened; I think its growth engine will keep marching along.

SoFi officially announced that crypto transacting will return to the SoFi ecosystem. They’ll offer Bitcoin, Ethereum and Solana, with stablecoins, secured lending, payments and staking all planned thereafter. When SoFi got its banking charter, it was forced to shut down its crypto business. Regulations have since opened the door for chartered banks to offer this service, and SoFi is taking full advantage. 

For firms like Robinhood, crypto powered over 40% of their total transaction revenue in the most recent quarter. I do not think SoFi will ever reach that kind of crypto business skew. Robinhood’s client base is made up of the people who tend to gravitate to this asset class, while SoFi’s is simply not. There will be less, but still material, interest in the product for SoFi. Before shutting down, the business was doing about $6M in quarterly revenue. Since the business was shuttered in 2023, SoFi has delivered about 70% cumulative member growth. Considering this, it’s easy to see how this could quickly ramp to $40M in incremental 2025 revenue. That’s just 1.2% of revenue estimates, but we’ll take upside whenever we can get it, and that revenue should come with a strong margin. The other product announcement was international money transfers among 30 countries. According to SoFi, this product will be faster and cheaper than competing offerings. Yet another way to generate member and engagement growth in a competitive field.

Jefferies cited Duolingo data pointing to June daily active user (DAU) data slowing to 37% Y/Y. 37% Y/Y growth is still surely excellent, but it’s also slowing faster than it’s supposed to – according to this data. The information Jefferies is citing comes from Sensor Tower data, which is often noisy and imperfect (just like all alt data is). Still, if it is accurate this time, this is below their current targets. The note is what led to the weak stock price action we saw this week. Some think this rumored slowing is related to souring social media sentiment stemming from the AI-first pivot. I heard the same things when Duolingo changed the layout of its home screen and its users supposedly boycotted as a result. Through all of the home screen noise, Duolingo continued to outperform. I expect the same to be the same this quarter. We shall see.

Mercado Libre’s highest-margin market is rocking from a macro perspective. Argentina just delivered 5.8% GDP growth while inflation continues to fall (but still remains very elevated). This nation flipped from large financial drag to the highlight of Meli’s financial results last quarter. News like this, as well as commentary from leadership during the quarter, points to that tailwind continuing to blow.

Bank of America sees a new off-platform advertising deal with Alphabet driving material ad revenue growth. I hadn’t heard about this deal until I read this note, and it’s an encouraging piece of news. MELI wants to be the largest digital advertiser in Latin America, and off-platform (outside of its own marketplace) traction is needed in that pursuit. Good news.

Morgan Stanley upgraded Snowflake to overweight based on improved product innovation building great traction.

Oppenheimer raised their profit estimates for Amazon based on brightening tariff assumptions. They maintained their outperform rating.

Oppenheimer is seeing a bet volume acceleration during the month of June for DraftKings. They slightly lowered 2025 and 2026 estimates due to tax hikes in key states like Illinois and Maryland, but, like me, don’t think that will even matter. When something trades for 20x forward FCF with an 80% multi-year FCF CAGR, there’s way more than enough room for falling estimates. Their hold rate data also points to somewhere around 10.5% for Q2. The firm reiterated their outperform rating on DKNG.

Morgan Stanley surveys revealed a resilient share of digital wallets for PayPal, with retention rates in line with best-in-class peers and a reversion in market share losses since 2022. In this same report, the analyst talked about sizable gains in Shopify’s Shop Pay momentum among respondents as well.

RBC has “increased conviction” in PayPal as Venmo monetization builds, the branded checkout overhaul resonates and Braintree’s profit-focused pivot bears fruit. They also see PayPal hitting their 2027 financial targets, which are ahead of consensus expectations.

DA Davidson called out an encouraging ramp in Experiences growth for Airbnb. So important for the durability of the growth engine.

Piper Sandler raised its 2026 and 2027 Meta revenue estimates by 1% and 3%, respectively. They think Meta’s AI investments are leading to strong improvements in ad performance and reporting. They’re noticing tangible conversion uplifts in their channel checks, which mesh very well with what Meta leadership says.

Morgan Stanley research is showing Google search traffic growing at a steady pace similar to when ChatGPT launched. Resilient.

PayPal:

Nu launched support for physical SIM cards as part of its NuCel telecom business. Telecom and banking are two of the largest industries in Latin America with two of the lowest customer satisfaction scores. Nu is rapidly disrupting banking and wants to do the same with telecom.

Amazon is expanding same-and-next-day delivery to 4,000 small cities by 2026. This is how you take advantage of the billions upon billions you’ve invested in carefully building out a logistics footprint. Amazon, through things like proximity and routing algorithms, has the ability to offer this level of service delight while others can’t. This should engender continued customer growth and engagement.

Robinhood shared June trading data, with equity and options volumes above consensus and crypto modestly below consensus.

Chipotle added a new $500M credit revolver.

Meta hired 3 OpenAI researchers away from that company. It’s also raising $29B to fund more data centers (supposedly $26B debt; $3B equity).

Macquarie Equity Research sees DraftKings and Flutter both beating Q2 expectations.

Output Data:

Inflation Data:

Consumer & Employment Data: