I sent Zscaler and Okta earnings reviews, Nvidia regulatory news and a GitLab snapshot during the week. I also sent some portfolio updates.

Broadcom creates & manufactures a slew of semiconductor-related equipment within data center, networking and industry-specific use cases. It also offers a range of software tools, which significantly broadened out with its VMWare acquisition. This company does not compete with Nvidia in terms of designing GPUs. Instead, it focuses on networking and connectivity, which competes with Nvidia’s switches and its SpectrumX networking product.

Broadcom’s VMWare acquisition is impacting several company metrics. It’s greatly benefiting overall revenue growth, as well as infrastructure software revenue growth. Finally, the acquisition is greatly hurting GAAP margins due to the M&A-fueled stock compensation, restructuring and integration costs.

It swapped $5 billion in floating rate debt with cheaper senior notes during the quarter. AVGO also used cash from the sale of a VMWare segment to reduce debt by $4.2 billion. It will pay down another $1.9 billion in senior notes in Q4.

AVGO will begin offering quarterly guidance again now that it has moved far enough away from the VMWare purchase.

AVGO trades for 26x forward EPS. EPS is expected to grow by 14% this year and by 26% next year.

The VMWare integration continues to go wonderfully smoothly from a financial perspective. VMware offers virtual, localized layers of software that sit on top of hardware. This allows the centralized hardware to run several different operating systems from the same place. The company, which is now a Broadcom unit, calls these “virtual machines” or virtual private clouds. By reducing hardware requirements, VMWare saves its clients money.

The business generated $3.8 billion in revenue vs. $3.3 billion last quarter, as Broadcom marches to its goal of $4 billion in quarterly revenue there. Broadcom is pushing aggressively to simplify product offerings and shift to subscription revenue. A big part of that is its “VMWare Cloud Foundation” (VCF). That simply refers to its full suite of products that “virtualizes an entire data center and creates a private cloud environment on-premise for enterprises.” This differs from a public cloud approach, like we see from Amazon, Microsoft, Google and others. Maintenance is generally more cost intensive, but private clouds are considered more secure. This quarter, VCF was 80% of total VMWare business as annualized VCF bookings rose 32% Q/Q.

On the cost side, Broadcom cut VMware’s OpEx rate again from $1.6 billion to $1.3 billion sequentially. It now sees reaching or beating its $8.5 billion in EBITDA from the acquisition roughly one year ahead of schedule. It is routine to see financial goals and presumed synergies prove to be overly optimistic in large scale M&A. That was not at all the case here.

Networking revenue exploded higher this quarter, with 43% Y/Y growth. This is being carried by the explosion in demand for high performance compute data centers amid the GenAI boom. Relatedly, hyper-scaler demand remained exceedingly strong. Broadcom’s GenAI niche is predominately in networking revenue. Superchips and high-performance compute (HPC) can’t all be packed into the same corner of a data center. GPUs must be able to connect to one another to drive better bandwidth and performance, with faster, more efficient model training and inference to cut costs. This is where Broadcom thrives. All in all, AVGO sees $12 billion in 2024 AI revenue vs. $11 billion last quarter.

The non-AI portions of networking revenue bottomed for Broadcom as expected. Growth for the segment was 17% Q/Q, and it sees positive Q/Q growth continuing next quarter. It saw an expected recovery in server storage with 5% Q/Q growth. In wireless, revenue rose 1% Y/Y and it sees 20% Q/Q growth for the segment next quarter thanks to industry launches. Broadband revenue fell 49% Y/Y as telecom players and service providers continue to manage spend.

Nothing wrong with this quarter. All of the segments that should be strong for it outperformed and it again raised its AI revenue guidance for the year. Nvidia has just trained investors to demand monster beats and raises for GenAI hardware darlings, and AVGO is in that group. It’s understandable to see the quarterly guidance miss be punished, but I don’t think long term investing bulls should be overly concerned about this report. It was solid for the most part. Not amazing… certainly not terrible.

IOT trades for 230x forward EBITDA. EBITDA is expected to grow by 390% this year, by 84% next year and by 74% the year after.

Yipit, a world-class 3rd party research firm, shared some Amazon data this week that bodes quite well for the current quarter. AWS growth last week accelerated to 21.5% Y/Y to its fastest level in two years. At the same time, North American retail accelerated above 10% Y/Y to its fastest rate of the quarter. 21.5% Y/Y growth compares to consensus estimates of 19%, while North American Retail growth compares to 8.4% estimates. This is exciting, but remember it’s data for only one week.

We’ve written a lot on Amazon’s margin turnaround and all of the levers it has to pull to drive more expansion. Advertising… 3rd party selling… fulfillment localization… offering its supply chain as a service… robotics… etc. One of the most powerful bottom line drivers, however, is top line outperformance. Variable costs don’t scale in tandem with that revenue outperformance. So? Margins and profits outperform too. If this data isn’t an anomaly, Amazon should be gearing up for a very strong showing next month.

The energy drink category continues to be challenged. Per Circana, the overall space has seen roughly 1% Y/Y growth quarter-to-date as macro headwinds persist. One may think this sector is relatively immune from these macro cycles, but it isn’t. Energy drinks are an affordable luxury, and vendors enjoy a large chunk of demand in convenience stores. When times are good, consumers will more frequently pay for convenience. When they’re not, they won’t. 7/11 reporting -4% Y/Y traffic points to the consumer still being fragile. And for energy drinks specifically, Celsius sees hesitant consumers, with halting category growth the result.

Celsius leadership is extremely confident in the sector returning to growth. Why? According to them, consumers are not leaving the category in favor of other options, they’re just spending less on discretionary goods overall. What will happen when things do finally turn around for the sector? There’s reason to be optimistic. For this year’s inventory resets, Celsius secured 39% Y/Y growth in distribution points and 45% Y/Y growth in c-store shelf space. This positions them for more growth as the backdrop brightens a tad and comps get MUCH easier. With this added space, they should be the main benefactor of the sector’s eventual rebound. 2025 inventory reset discussions have so far been “highly positive” to keep this momentum going.

Pepsi is the main North American distribution partner for the energy drink company. Celsius collects revenue as it sends product to Pepsi, not as that product is distributed to retailers. Through 2023, Pepsi supposedly ordered way too much Celsius. It has since been rightsizing the amount of inventory it is keeping on hand. This rightsizing means lower incremental purchases and lower Celsius revenue.

Last quarter, this cost the firm a little more than $20 million in revenue. The team told us it didn’t know what Pepsi would do for the rest of the year but felt that inventory dynamics were in rather good shape. Following this call, it’s clear that the worst of Pepsi’s resetting is not over; Celsius leadership told us to expect a $110 million revenue hit from this decision for Q3. The size of this impact led some analysts to think there’s a demand problem. I don’t think that’s the case for two reasons. First, Circana scanner data remains strong for Celsius and points to 10% Y/Y growth in actual product sales (not growth in distribution to Pepsi). That’s 10x the category’s growth rate despite comping over triple digit growth last year. Secondly, Celsius inventory depletion rates at Pepsi have not slowed. They’re stable according to the team. That lends credence to the idea that demand is fine and Pepsi just doesn’t want to hold as much as it did last year. As long as depletion rates and sales remain resilient, this company cannot shrink inventory held forever. And when these resets finally end, revenue growth should revert right back to what we see with Circana.

I have one other theory about what potentially could be happening here. While the last few years have been very weird for supply chains, it’s hard to believe Pepsi messed up this much on overbuying through 2023. Ask yourself: “If Pepsi wanted to buy Celsius for a lower price, what would it do?” I think the answer is clear. It would overbuy and then end that overbuying to crater CELH’s results. Doing it this way would also ensure there’s no disturbance in distribution to customers. That’s what is happening here at the moment. I find it somewhat likely that Pepsi wants to eventually own all of Celsius to tuck into the brand portfolio. Just speculation. But I don’t think this is much of a stretch.

Celsius was asked about its decision to hike frontline pricing (price offered to retailers pre-discounts) alongside Monster and Red Bull. This timing is somewhat interesting, considering the category is dealing with a weaker consumer. But Celsius doesn’t actually plan to benefit from potential hikes this year. It sounds like for now they’ll simply discount products more deeply to keep overall pricing consistent with pre-hike levels. As macro improves, it plans to actually start selling more of the product at these hiked prices to eventually raise its potential revenue and margin ceilings. The company is dedicated to maintaining its premium pricing and brand position. Its high(er) quality ingredients vs. the competition warrant this.

I didn’t add to my new, very small stake in this name following the investor conference sell-off. I’ve decided that I have very little interest in adding or subtracting shares at this point. Why? Consumer brands come and go constantly, and there’s a small chance that this impressive brand is beginning to peak. Brand popularity and longevity are wildly difficult to model for anyone yet are vital to the success of an investment. For this reason, I love Celsius as my currently smallest holding.

Where does my cautious, guarded optimism come from? The way it stormed onto the scene and stole considerable market share over a several year period is not a fluke. New entrants routinely rack up a couple percent in share with heavy promotions and flashy partnerships and also routinely flake out. These new entrants don’t build a solid 10%+ U.S. market share like Celsius did (with international growth only now beginning). Weakness also perfectly coincided with souring macro and Pepsi (in hindsight) overbuying and speeding up distribution gains. Pull forwards yesterday mean tougher comps today, which is being amplified by Pepsi’s simultaneous decision to shrink inventory on hand. I think this is a perfect storm of transitory headwinds. I see immense upside as monetary accommodation flows in, and the consumer starts feeling a bit better about things IF I am right about this not being structural brand decay. I don’t think I need to own a large (or even medium-sized) position in this to greatly benefit from that vision playing out.

As a reminder, CrowdStrike was the main reason for a global IT outage that happened this July. A massive number of customers and endpoints were affected, and CrowdStrike has been in damage control ever since. CrowdStrike messed up on a configuration update by overloading its agent with more of these updates than were meant. This didn’t involve any core code or software changes and wasn’t a security breach. I covered CrowdStrike’s internal changes implemented to avoid a repeat in section 2f of this article.

CCPs are customer perks offered as concessions for CrowdStrike’s blunder. Perks will range from comped products, flexible payment terms, extended trials, free professional services help etc. Per Podbere, CrowdStrike will “take the temperature” of its customers and their overall impact, with the help of its system integrator (SI) partners. These packages are really only meant for customers like Delta who were harshly affected by the July outage. CCPs are also somewhat malleable, with about $60 million in overall perks (direct revenue headwind) being offered on a “slider” based on severity of damage. It’s “not one-size-fits-all.”

For review, Falcon Flex is a newer CrowdStrike selling approach that offers clients more “flex”ibility in how they use modules and when. It allows committed spend to be drawn down at a client’s desired pace while allowing them to turn on and off whichever modules they need at a specific time. This significantly diminishes buyer friction in many ways. Here are two of many examples:

CrowdStrike enjoys larger contracts, longer commitments and shorter sales cycles, while customers enjoy significant incremental flexibility. Podbere told us during the interview that he wants all customers to be on the Falcon Flex plan and that it’s the module “delivery tool of choice” for CCPs.

Beyond Falcon Flex, CrowdStrike has another strong preference for how these will actually be structured. It wants all of the perks to be comped modules for customers to “seed” them with more product.

To Podbere, CCPs can be used to accomplish two things. First, they can obviously be used to mend fences. Secondly, and less obviously speaking, they can drive more future annual recurring revenue (ARR). How? Falcon is extremely sticky. If churn rate falling Y/Y during the aftermath of this outage isn’t evidence of that, I don’t know what is. CrowdStrike doesn’t plan to give these comped modules away forever. It plans to essentially offer lengthy free trials, with charges coming down the road as customers get used to having this still best-in-class platform. Today’s free samples are tomorrow’s incremental ARR.

Podbere was very candid about how little visibility and certainty he has into the true impact of July’s outage. He “tried to be prudent” when modeling the impact. As a reminder, there were a few layers to the revenue guidance reduction from last quarter. First is the $60 million CCP hit. Second is a $10 million impact from professional services help. Thirdly, it baked in $30 million in impact from more deal scrutiny stemming from this mistake. CrowdStrike always leans highly conservative when offering any forward guidance; I expect this to be no different.

Podbere hinted at $10 billion in ARR by fiscal 2029 still being on the table. Last quarter, it changed its target path to $10 billion from 2029-2031 to 2031.

"There's a bunch of things that can get us there. But look, I think similar to how we guide on a quarterly and a yearly basis in terms of our methodology and the prudence we take, it's the same with the 5 to 7 year range we gave out last year."

CFO Burt Podbere

“Overall, I just think the company is in the best place we've ever been from a customer and competitive perspective.”

Co-Founder/CEO Jason Robbins

DraftKings kills it during football season. It’s the company’s best sport of the year both on gross and competitive terms. Most of its product improvement focus centers on ensuring the user experience is optimal as football season is beginning. This year, a big focus was on improving live betting. For DKNG, live betting is less than 50% of overall volume, while it’s over 70% in more mature markets like the UK. Co-founder/CEO Jason Robbins thinks football is even better suited for live betting than soccer, considering all of the annoying commercial breaks. To help here, the company finished integrating BetVision into its app (owned by Genius Sports) for low latency game streaming and odds updates.

Another big tool for boosting live betting proportions would be adding more live parlays. It’s actually quite complicated to do that, considering multiple sets of odds have to be reset constantly to ensure fair pricing and acceptable uptime. These challenges (and expected cost savings) prompted it to buy live game line provider Simple Bet. Jason Robbins hinted at DKNG moving closer to offering a lot more live parlays and this is the reason.

The rest of the product changes focused on backend improvements. It pushed hard to lower page load times, cash out availability rates and betting line up-times. Today, Robbins thinks DKNG is “best in the industry” for those metrics. While that may seem subtle, it can be the difference between pleased customers telling a friend to try the product or an annoyed one telling them not to.

While DraftKings does very well during football season, it thinks it can do a better job competing for NBA volume. DKNG “trails” FanDuel in this sport, per Robbins. He thinks that’s the largest reason why share gains nearly paused (didn’t reverse) last quarter.

Beyond that, it thinks it can do a much better job on hold rate (turning bet volume into actual revenue). FanDuel’s hold rate is significantly higher than for DraftKings. That means DKNG has a clear opportunity to take more market share with its existing volume if it does things like offer more parlays (higher hold rate) and display those options more clearly. That’s the plan. Achieving it would not only mean more revenue, but also better product parity vs. FanDuel.

“We haven’t focused enough on driving parlay mix and hold rates. We’ve since course corrected.”

Co-founder/CEO Jason Robbins

“I’m seeing excellent numbers to start off this season.”

Co-founder/CEO Jason Robbins

Robbins talked a lot about the firm’s decision to spend more on customer acquisition. Through consistent optimizations, growing national scale (more eyeballs per dollar spent) and strong industry momentum, DKNG delivered 80% Y/Y growth in new users with cratering customer acquisition cost (CAC) last quarter. It saw this outperformance unfolding in real-time and decided to embrace it. This hurt its EBITDA forecast for the year (more promos and marketing) but will mean more revenue and more profit next year and beyond.

This is how Robbins will always act. He will always do what he thinks is best for the company’s long term future, rather than maximize quarterly profits. Still, he did also say he needs to get better at knowing when these “unexpected things” and opportunities will surface. I’m not sure how the team would do that, but to me his language hinted at DKNG getting incrementally more conservative in its guidance methodology as of last quarter. They probably just want a larger margin of safety going forward for a beat-and-raise-obsessed street.

Importantly, the customers DKNG has secured recently are as high quality as older vintages. It is true that its highest quality customers are those that sign up immediately after a state’s launch. At the same time, customer quality degradation doesn’t last forever, and DKNG is seeing quality stabilize in its established states (3+ years into legalization).

As a reminder, DKNG reiterated its 2025 guidance of $950 million in EBITDA despite the Illinois gambling tax hike (more next section). Robbins was asked where this confidence comes from. It’s just a byproduct of a growth mix shift. As the state footprint matures, it expects to get more profit growth from lower promotional and marketing needs, as well as higher anticipated hold rate. Those two levers are far more powerful bottom line drivers than winning more volume.

Last month, DraftKings announced a tax surcharge in high tax states like Illinois. It said these proposals could easily be changed based on overall market reaction. It quickly walked back on the idea immediately after FanDuel announced it wouldn’t follow suit. Robbins did sound a bit nervous about the idea of a few more states raising taxes. This nervousness stems from a lack of total control more than anything. Still, DKNG isn’t entirely powerless. Its lobbying group (which includes FanDuel) is far better equipped to advocate for lower taxes going forward than it has been in years past. The Illinois decision shocked the industry a bit. That shock was used as a needed kick in the butt to add some muscle here. While Illinois raising taxes coincided with two other states rejecting proposals… and despite Illinois relatively pressing budget needs… I think this focus is well placed.

In the eyes of Robbins, either the tax landscape stabilizes, or more states follow suit and the industry eventually responds. He doesn’t envision a future in which taxes go up across the board and the industry doesn’t implement some solution resembling (not identical to) the tax surcharge. For now, the EBITDA maintenance in Illinois will likely come on the marketing and promotional side.

This interview got into the weeds a bit on Azure Fabric’s approach to data amid the GenAI explosion. It’s a somewhat slept-on Microsoft product, so I’m glad we got a brief interview on it.

As countless players race to build the biggest and best model, a data foundation is a prerequisite to winning. Models are only as good as the quantity of relevant data they have access to train on. That’s where Azure’s broad suite of database products, query language options and integrations come in handy. “Fabric” is its overarching platform to tie together all of its data products. Fabric is inherently open-sourced, which creates a more cohesive environment for minimizing data silos and lockage. And shockingly, there’s a copilot for Fabric to ensure customers have an easier time opening their arms to data modernization and GenAI transformation. Again, the two ideas go hand-in-hand.

To deepen the value of this structure, Microsoft unleashes its data across its wildly broad suite of products. It’s not just organized and secure… it’s actually usable in Excel, Teams, LinkedIn etc. That reality has helped Microsoft successfully create intelligent copilots across several individual products. Again, these copilots need data. Enter Fabric.

Ulag offered an analogy of other platforms forcing customers to buy pieces of a car and assemble them on their own. He joked about Chief Information Officers becoming Chief Integration Officers, as they seemingly spend their time ensuring fragmented systems work well together. By housing everything a customer needs in one place (with endless scalability thanks to Azure’s infrastructure), Azure offers the full car, in one package and in a fully managed fashion. Developers don’t need to paste anything together.

In practice, Fabric greatly reduces costs by driving vendor consolidation and broader interoperability. Furthermore, in this case, vendor consolidation removes the need to “create pools of isolated compute” across several point solutions. This means compute spend can be allocated to whichever Fabric use case a customer wants. This minimizes wasted budget. For one customer, this meant cutting data spend from $165 million annually to $45 million. Indirectly, Fabric’s ability to help fetch and conjoin all needed datasets has a way of making companies smarter, more efficient and, in turn, more profitable too.

Ulag is especially excited about Fabric’s developments in real-time intelligence. In his point of view, most firms still work with “batch mode” data on at least a 24-hour delay. This blocks the ability to act on real-time information. Fabric unlocks this ability by processing data as it’s created.

In Shopify’s quest to let merchants seamlessly “sell anywhere,” adding relevant channel partners is vital. Whether it’s YouTube, JD, Meta, its thriving Shop App, TikTok or countless others, Shopify wants to be everywhere for its merchants. It wants them to easily focus on growth and seamlessly turn on new selling avenues with the click of a button, rather than obsess over how to get this done.

This week, Shopify was named as Roblox’s first “commerce integration partner.” Shopify will power the online cataloging and shopping experience for Roblox users and will let its merchants tap into this incremental demand. More channels… more reach… more revenue. More of the same.

Lululemon CEO Calvin McDonald purchased $1 million in company stock during the week. That’s his first open market purchase. His net worth estimates online are widely ranging. Even if we assume the higher estimates of $70 million are accurate, this was a material purchase.

PayPal launched the ability to stack and use rewards in brick and mortar settings. This includes 5% cash back for whatever category a customer wants. Consumers can also now add their PayPal cards to their Apple Wallets.

The Trade Desk is rumored to be building a smart TV operating system to sell to hardware vendors.

Google’s antitrust trial for its advertising business begins next week.

In an interview this week, Intel’s CFO updated us on its cost cutting plans. Everything is on track. Intel had significant cost bloat within basic teams at the company when compared to its competition. Its finance and HR teams were way too large and unproductive. That represented easy ways to cut costs without sacrificing product roadmap. It’s also still full speed ahead on its Foundry segment, and still expects it to reach a 60% gross margin (30% EBIT margin) by 2030. It still expects the foundry business to generate packaging revenue this year and wafer revenue by 2027 (maybe 2026).

Salesforce is buying Tenyx, which specializes in GenAI conversational chatbots.