Articles already sent this week:

Table of Contents

1. Axon (AXON) – CFO Brittany Bagley Interview with RBC Global

Skydio & Approach to M&A:

There was a change to how the Skydio deal will be funded. Instead of a more balanced mix of equity and cash, Axon is scrapping most of the equity portion and paying out most of the proceeds in cash. This is a direct response to their share price, as leadership explicitly said they don’t want to issue equity at a valuation they view as far too low. There's still a small equity portion involved to align interests, but those interests are already in harmony for other reasons. Axon is still just a minority shareholder here, with Skydio maintaining broad independence to pursue several industries and opportunities outside of Axon's reach. Skydio’s team obviously wants to be successful. This deal was not a case of the leadership of an outright purchased entity vanishing to a faraway beach after cashing in. The Skydio team is fully intact, engaged, and now flush with a new cash injection to pursue more of its pipeline and fulfill more of its backlog. Furthermore, Axon is the exclusive distribution partner for Skydio’s public safety drones in North America, which is a massive part of this business. That means Skydio does better when Axon does better, further motivating it to innovate and execute well in this category. 

Generally speaking, this process was a lot like previous ones. Axon makes an initial investment, takes its time to gather information and evaluate management.  Then it  decides if it wants to remain partners or own the entire enterprise. In the case of Skydio, Axon did want them to stay separate, as again there are a ton of exciting drone-based markets to pursue that are well outside of Axon’s domain. Focus is important, and talented teams tend to break up when they’re integrated into larger entities – even one as respected as Axon. Bagley used Fusus as a different example where they did decide to buy the whole company, knowing that the centralized real-time crime center could unify various pieces of Axon’s value proposition and uplift the entire platform.

Going forward, it does sound like there will continue to be a steady stream of investments, although most will likely be much smaller than this one. They see no big gaps in the corporate offering, but are always looking for opportunities to add useful new features.

$130M in Dedrone Delays:

Axon thinks the World Cup created a sense of urgency for a lot of federal deals to close. It sounds like the team got too excited and extrapolated this urgency as a lasting feature of this end market. It’s not. In reality… to nobody’s surprise… the federal government moves very slowly and balances a wide array of items that are constantly rising and falling on the priority list. The World Cup in 2026 helped a lot. The Olympics in 2028 will help a lot too. For now, they “hope” these deals will come in 2027 but aren’t going to bake them into forecasts until that gets clearer. Starting with Q4 guidance, all Dedrone deals will be excluded from forecasts unless they’re certain. Potential wins in the pipeline will serve as pure upside if they happen and won’t prevent AXON from meeting targets if they don’t.

  • Bookings growth excluding these deals is expected to reach 30% this year, which was their goal at the start of 2026. Still, these deals were part of that overall goal and they’ll fall short of 30% because of the delays.

  • Axon will clearly communicate Dedrone deal assumptions in guidance going forward. I think there’s a chance this turns them into a MongoDB-like guider, with quarterly forecasts getting a bit more underwhelming and the size of beats getting larger. They’re going to sandbag more because of this Dedrone unpredictability.

Tasers & Body Worn Cameras (BWCs):

When asked about the Taser runway, Bagley was as optimistic as ever. She discussed a lengthy remaining runway for TASER 7 to TASER 10 upgrades, with that product now in year 4 of availability. There are also still big deals to win in their core U.S. state and local market, while market share across international, corrections facilities, and some enterprise use cases remains very low. Furthermore, they just came out with the Taser Ion, which is smaller, more subtle and perfect for use cases where an officer or security guard needs to blend in. And finally, they’ll be coming out with a new Taser sometime in 2027 or maybe 2028. That, alongside the eventual introduction of their Apollo cartridge, will enhance range, chamber capacity and the ability to work on targets with thicker clothing. All of these upgrades should motivate a lot of refreshes and new demand in the years to come.

  • The AB5 will be coming out next year.

  • Early feedback on the ABW Mini was called “really positive.” As a reminder, this is for enterprise use cases predominantly in healthcare and retail.

  • It’s important to note that Axon still sells a lot of TASER 7s and AB3s. Shelf life is long in both cases, making things like inventory builds less risky.

Margins:

On gross margin, memory inflation and mix shift to newer products are both weighing on this line item as discussed in the review. This will hit GPM by about 50 bps for 2026 and by 200 bps for Q4 2026. That level of quarterly impact is expected to remain in place through 2027. They also expect overall Q4 2026 GPM to be roughly in line with 2027 levels. The goal is to continue generating enough OpEx leverage to drive EBITDA leverage as this plays out, although the severity of the GPM headwinds and the need to spend on growth will mean EBITDA margins are likely flat in 2027 vs. 2026. Expansion should resume thereafter.

For FCF, as discussed on the call, drone and other inventory loading to gear up for strong multi-year demand trends will keep weighing on conversion until the 2nd half of 2027. That’s when FCF generation is expected to quickly recover to its 60% conversion target. The $130M in Dedrone deal delays also led to heavy FCF headwinds during Q3, considering they already built the inventory and now won’t recognize the revenue until later than expected. This is all timing and will probably lead to FCF conversion outperforming for a few quarters as these deals do eventually close.

Automated License Plate Reading (ALPR):

While the Flock drama has brought ALPR into the spotlight for negative reasons, Axon continues to see it as an important part of its platform. Right now, it’s quite small from a revenue perspective, but it has been scaling nicely, with 30% Y/Y growth this past quarter. It is true that some departments around the country have cut usage of this technology from Flock and other vendors. At the same time, those instances are rare, as the vast majority of departments passionately defend responsible usage of this technology as a big part of doing their jobs safely and well. Furthermore, as discussed in previous coverage, Flock has ceded a pipeline of 500 customers (converted or in the process of converting) because of its mistakes. Axon’s pristine reputation when it comes to privacy and transparency, alongside its value-rich platform, makes it an easy option.

Long-Term Targets & Quick Thoughts:

No changes to their $6B revenue guide and 28% EBITDA margin guide for 2028. This represents 29% revenue compounding and 250 bps of margin expansion over the 3-year period. Looking further out, they see “no reason why they can’t keep going at least at a mid-20s growth rate.” Again… all of the issues from last quarter were related to the timing of revenue recognition. Not deterioration of runway, market health or competitive positioning.

2. Rubrik (RBRK) – Cloud Resilience Event Review

Setting the Stage:

This was an event for Rubrik to showcase new products that have been discussed for a few quarters and are finally launching. It was not for discussing the financial side of the business. 

When looking at the big picture, $140B is being spent every quarter on cloud infrastructure, with AI making up a larger and larger part of it. For context, the team discussed a data lake market expected to move from $18B to $52B by 2030, alongside an IDC report on data creation moving from 6.9 petabytes per second to 17.1 in the next 3 years – powered, of course, by agent activity.

The systems surrounding all of this activity are built by brilliant engineers who fixate most on speed and efficient scaling. And while these focus areas are highly important, the tools used for valuable asset recovery when things inevitably go wrong are lacking. That’s the gist of RBRK’s entire value proposition. 

Security systems like CrowdStrike and Palo Alto are world-class but not perfect. Breaches are inevitable, which was especially obvious amid the last few weeks of outages and newsflow. Built-in cloud tools (like snapshots) provide bandaids with a 7-14 day recovery window vs. months or even years of retention capabilities with Rubrik depending on what the customer wants. These tools are also part of the same account as the vulnerable data, so breaking into one with a stolen login means doing so for the other too. Rubrik, alternatively, allows rapid restoration of a customer’s most recent clean state, with enterprise scalability and effectiveness across a wide array of assets and partner environments.

One may intuitively think, “don’t hyperscalers just do this for customers on their own?” Throughout the event, leadership from both AWS and Google Cloud bluntly described their role as protecting the overarching infrastructure and making sure failures here don’t lead to lost customer assets. Their job is not to safeguard a customer’s data and configurations or recover assets that were accidentally or intentionally lost. They’re happy to offload this work to Rubrik, with close long-term partnerships that just keep tightening. As the threat landscape goes from bad to worse, end-to-end tools that RBRK provides simply become more imperative and these relationships get more valuable.

Asset Recovery and Protection:

Rubrik showed off its rapid restoration that makes data immediately available to use after recovery – without requiring customers to manually reconstruct anything. As always, backups are immutable with years of retention as described above. This matters a lot for companies in tightly regulated industries like banks and hospitals, as keeping records dating back this far is often a legal requirement for them. In the last few months, they’ve added support for Apache Iceberg on AWS and for Google BigQuery too. They also unveiled protection for Google Cloud Storage at this event. Again… Rubrik is rapidly building integrations with the most important partners for more of its products. 

For source code, Rubrik added a new cyber recovery offering for GitHub and Azure DevOps. They already were offering immutable backups here, but now, with the new DevOps Protection Enterprise Edition, they’ve added detection, investigation and more granular recovery capabilities. They now continuously monitor audit logs for any weird activity, map a “blast radius” (everything a rogue account or hacker touches) and granularly restore only what was affected. Alongside this end-to-end service, RBRK even generates evidence packs that customers can hand to insurance companies and regulators when need be.

Finally, Rubrik launched the Autonomous Business Recovery (ABR) offering for cloud applications that they’ve been discussing for the last several months. It’s now generally available, as a unified service to “automatically discover and map every dependency” so that customers can protect each app as one unit. This is offered with recovery test drills to prove a full app can actually be restored (and how fast) before a real incident hits. It also comes with something called “Surgical Recovery.” It brings back the newest backup copy, but swaps out any damaged files for older, safe versions. That way, companies get clean data back without losing weeks of work. This is instead of rolling everything back to the last backup that was fully clean, which could be up to 30 days old depending on when the hacker first got in. This is Rubrik’s push to offer a truly platform-wide and full-service suite for business recovery. It’s how Rubrik will drive cross-selling, preserve pricing power, extend its growth and rapidly compound for a lot longer to keep earning its premium multiple.

Below I’ve included the explanation of this product that I wrote as part of the RBRK Investor Day review this past June:

With the foundation of Rubrik AI, the company is extending the reach of its world-class automated immutable data backups to boast end-to-end software application resilience and recovery. And, considering that arguably the most important part of cloud app resilience is recovering all compromised data, Rubrik’s core positions it very well to seamlessly expand into this highly relevant adjacency.

The new offering heavily leans on Rubrik’s Preemptive Recovery Engine to restore a customer’s Minimum Viable Business (MVB) for all cloud applications. This includes full normalization of data environments, app experiences, as well as perfectly reestablished mazes of highly complex permissions and dependencies. And it does all of this in a purely automated fashion that merely requires the click of a button for customers. The innovation turns weeks of hard manual work into a profoundly more efficient process, saving money, downtime and headaches.  

Alternative solutions require “dozens of engineers” frantically patching issues, manually writing code and hoping the sequence of repairs leads to a fully functioning app. That requires a lot of random trial and error, testing and more fixing – all of which means more downtime and higher costs. Rubrik does all of this while allowing customers to pre-set recovery plans and test their efficacy (whenever desired) before any issues pop up. Importantly, it also features integrations with RBRK and partner security tools, eliminating the possibility of a team initiating a recovery that still has a pressing vulnerability in it.

3. Uber (UBER) – CFO Balaji Krishnamurthy Interviews with Morgan Stanley

Delivery:

As discussed in the recent review, overall demand remains strong. Category positions are stable, and CFO Balaji Krishnamurthy added that they’re now number one in New York, LA and Miami. Going forward, key developments like the Costco partnership expansion from 17 states to nationwide, gaining preferred delivery status for McDonald’s 1st-party orders and adding its Eats inventory to Amazon will all be big pieces of keeping this impressive growth engine going.

Another piece of that will be ongoing developments in autonomous fulfillment. Uber is pleased with early scaling for sidewalk and bike-lane delivery robots. They’re not practical in all settings, and create more friction than convenience in many instances. Still, they can work quite well for local deliveries in urban areas. Furthermore, unlike their nascent autonomous vehicle (AV) mobility program, this was immediately profitable from the beginning, without hefty upfront costs. For many other cases, drones will be a future go-to, with a Zipline partnership now in place that they hope will get to 1M daily deliveries by the end of 2029.

On recent industry M&A, Krishnamurthy said nice things about Instacart as an asset. On the other hand, he also talked about Uber’s ability to develop a lot of what it could have purchased from that entity internally. It sees a great opportunity to naturally get closer to some of Instacart’s biggest partners with this deal now announced. Finally, early communication with regulators tied to the Delivery Hero purchase has been “constructive” and “as expected.”

AV Stuff:

A lot of this interview was spent reviewing how Uber can turn the AV revolution into a tailwind while the world seems so certain that it’s an existential risk. It all comes down to the same things we’ve been consistently discussing. They’re increasingly confident that many, many players will successfully scale in the market. Great software help from Nvidia and about $10B in capital from Uber are both helping make sure that happens. Thanks to this support, many deployments from a plethora of OEMs are planned over the next several quarters. In a world that isn’t mostly dominated by Waymo and/or Tesla, demand aggregation becomes extremely important. These pieces of hardware are quite expensive and incinerate cash if they’re left unused. Uber gives at least 30% higher utilization than 1st-party only fleets. 

The better demand levels and visibility, thanks to its best-in-class consumer scale, allow it to guarantee fleets higher revenue minimums to bring their supply onto the network. It also means Uber can justify charging a take rate of up to 23% (based on 30% utilization boost), which is significantly higher than what it intends to do. They’ll complement better fleet utilization rates with charging, storage and other operating infrastructure, as well as other benefits like enterprise-grade customer service systems. It’s very hard to manage all of this on your own and to do so profitably. It’s even harder to prioritize all of that while racing to broad availability of approved AVs in more places. Uber allows companies to just focus on that extremely hard task while lowering the financial risk everywhere else.

The Rest:

The rest of the conversation was spent reviewing the same human supply issues we covered in the most recent earnings review.

4. NU (NU) – Approval

Nu received central bank approval for the small acquisition it needed to keep using “bank” in its name in Brazil. As a reminder, new rules meant the licenses it had no longer allowed it to call itself Nu“bank,” so it made this move to keep the status quo. The deal was already announced in July and should close in the coming days. This doesn’t change anything in terms of allowing it to offer new products like it did this past quarter in Mexico. Just means it can keep messaging the same.

5. Amazon (AMZN) – Leo

Amazon’s Leo had another launch failure this week. A Vulcan rocket from ULA, one of its launch partners, failed about 10 minutes into flight, and all 40 satellites were lost. This was the first of 38 planned Vulcan missions for Leo, and with Vulcan now grounded pending an investigation, this will likely lead to delays for the program. Fortunately, this is an irrelevant part of the business right now, but it has a lot of potential, and delays like this obviously push out when that potential can be realized.

6. Macro and Headlines

  • CrowdStrike is exploring a Zscaler purchase. This would give it a world-class network suite to join its endpoint-centric umbrella. Together, the combined entity would more closely match the breadth that Palo Alto offers.

  • Meta hopes to start shipping its Muse Charm AI hardware next month at $300 per device. It will come with 1 year of free Verizon 5G as well. They postponed the planned smartwatch launch to next year.

  • Jobless claims data was a bit worse than expected, while manufacturing data from the Philly Fed for November was also weak. On the other hand, retail sales rose by 0.2% M/M for October, which was better than 0.1% growth expectations. Core retail sales rose by 0.6% over the same period vs. 0.4% growth expected. Finally, despite Christopher Waller calling for 100+ bps of cuts from here in a speech this week, the probability of no cut for December rose from 0% to 11%.

  • AppLovin is in talks to buy Criteo.

Reply

Avatar

or to participate