|Shares Out. (in M):||129||P/E||0||0|
|Market Cap (in $M):||575||P/FCF||0||7|
|Net Debt (in $M):||210||EBIT||0||0|
Sign up for free guest access to view investment idea with a 45 days delay.
EXTR is an under the radar, completely forgotten "enterprise network equipment" play that is an industry technology leader that is dominant in certain customer verticals and is transitioning its revenue to be more highly recurring and software oriented, while trading at .75x sales, 1.25x gross profit, and ~7x our NTM FCF estimate.
We believe there are both short term and medium-term catalysts and that the company can ultimately be sold to a strategic if the stock continues to flounder. Finally, if the company is able to achieve their vision of "one network, one fabric", and is recognized by the market as such, the stock is likely to be a multi-bagger over the next 2-3 years.
Extreme Networks is a collection of networking assets that may finally be coming together to form a cohesive product. The company was founded in 1996 in California and IPO’d during the tech bubble, in April 1999. They had a major merger in 2013 with a company called Eterasys Networks--a merger of equals as each company was doing roughly $300 million in sales. In 2015, current CEO Ed Meyercord took over and has spent an additional $400 million on four additional acquisitions. While the stock has technically risen since Meyercord joined (it was scraping along at $2.50 a share vs. $4.43 now), and had a parabolic run from through 2018 as they were seemingly acquiring strategic assets on the cheap (specifically Brocade and Avaya divisions at seemingly distressed prices), since the beginning of 2018 the stock is down ~60% while the NASDAQ is up 70%--not a great three year run, but now creates an opportunity while its still depressed and ignored by active investors.
The most recent acquisition, Aerohive, was Extreme's pronounced entrance into cloud managed networking, the fastest growing area of the market, which we discuss more below.
Extreme Networks operates primarily in the Local Area Network (LAN) market, including the wireless local area network (WLAN) space. Per Gartner, they are one of three industry leaders in this space along with Cisco and HPE. As we understand it, Juniper is also a well-regarded competitor after their Mist Systems acquisition.
Because Cisco and HPE (and Juniper) cover broader swaths of the industry, Extreme can be thought of as the only pure play industry leader in the Wired/Wireless Access space.
They sell a mix of hardware, software, and services; however they are most known historically for their switches and routers (perhaps to their detriment). Their "recurring revenue", which is currently 34% of their total revenue (moving to 40% shortly), is comprised of maintenance contracts on their hardware revenue and software subscription revenue, which includes network management, security, analytics/AI, and other applications surrounded their Extreme Cloud IQ platform. Competitive differentiation is increasingly coming from the software side, especially due to the emergence of software defined networking (SDN).
Where Extreme really seems to excel, and perhaps in validation of their industry leading technologies, is some of the more complex and "extreme" instances of WiFi. For instance, they have contracts with most of the NFL (26 of 32 teams) to manage stadium WiFi, NASCAR, and recently announced a major displacement of Cisco ("heavily contested") with Major League Baseball, which apparently includes multiple Extreme products and covers all MLB teams. They also have relatively strong penetration in other high volume and complex network infrastructures like airports, college campuses/schools, hospital systems, and logistics (Fedex, UPS, and recently Amazon Warehouse). They also do well with Government segments like the CIA that have very high security standards.
They are really an international business, generating material sales in the US, EMEA (Europe, most concentrated in Germany, Austria, and Switzerland), and Asia (mostly Korea and Japan). They seem to have less exposure to China/India and other emerging markets. Perhaps also relevant for a read-through from peers, they don't have much exposure in Financials/Insurance.
Wireless Local Area Networks (WLAN)
We think of LAN and WLAN providers for enterprises as selling all the switches, routers, access points, and other hardware that makes up a network. As a very simple example, you can think of your house as a Local Area Network. Your house will generally have a router that all the internet devices will connect through and then, if your house is big enough, you may put additional switches and/or access points that essentially communicate with the router to expand the total area of the network. On a much larger scale, consider a football stadium that would again likely have one master router, but many more access points (the Green Bay Packers head of IT notes 1,200 access points) and switches that make up a network (still a local area network). Although Extreme does not really compete in this space, the Wide Area Network (WAN) is essentially connecting multiple networks together. So hypothetically, if all football stadiums needed to be controlled at an aggregate level, that might be a WAN. Or perhaps if a company has several headquarters or regionally different offices, that would be a WAN.
Campus vs. Data Center
We can also break the Enterprise WLAN industry down further by "campus" and by data centers, which require different types of hardware. While campus networking and data center networking have similar end goals (providing compute and storage across devices), they might as well be comparing a sports car to a pickup truck. The best way I've heard the difference described is the campus network is where USERS access the network (like say a literal college campus), whereas the data center is where DEVICES access the network. Data centers will be almost completely wired, while campuses will be a mix of wired and wireless. Data Center speed/throughput is significantly faster than Campus throughput, e.g. good data centers can move data at 100 gigs per second while Campus will be more like 1-5 gigs per second. While Extreme has Data Center assets from the Brocade acquisition, that space has a different set of leaders. For instance, Arista Networks is dominant in Data Centers but weak in Campus.
Finally, and perhaps most critically, we can break the industry down into cloud and non-cloud. Extreme is now the #2 Cloud Networking player by revenue and tied with HPE for number of Devices Managed, per Omdia, and this is almost entirely because of the Aerohive acquisition. This is clearly the fast-growing segment of the market, as Omdia forecasts a 5 year revenue CAGR of 23% vs. 3% for the total enterprise network equipment market.
What do we mean when we say "cloud-managed networking?" Ultimately, we mean centralized control of the network in how it is configured and modified. For instance, if making changes within the network one could make those changes through cloud enabled devices, rather than sending a technician to each individual switch, router, and access point. Extreme IQ, Extreme's cloud platform, is effectively a control panel that allows users to manage all of their network devices through software, similar to a Google Home setup where you can turn on your lights, control your TV, or manage your thermostat. Just as your Nest thermostat can be given automated rules to maximize temperature efficiency, "network automation" allows for remote control of the network device. The "automated network" is the holy grail of networking. Below is a snapshot from a recent webinar on their Extreme IQ platform. This dashboard allows you to not only look at the network at macro level, but also to drill down to individual devices on a company's network.
Cisco, through the acquisition of Meraki, effectively invented the industry 10 years ago by offering alternatives to an on-premise controller. Whether it was managing the large number of network elements, coordinating radio frequency, load-balancing access points, setting consistent user policies, or providing security and scanning for hackers, this was historically done onsite with a hardware-based controller. The idea behind cloud is that the controller is moved to a centralized location and can then be managed from anywhere.
An important thing to understand is this does not move ALL the hardware to the cloud. There will still be a number of switches, routers, and access points, but for them to be "cloud" they need to be cloud enabled and be "managed" remotely. When Extreme talks about how they have over one million managed devices they are talking about cloud enabled devices that are being controlled by the Extreme IQ Platform. Access Points (APs) seem to be the most common to have as cloud managed, although switches and gateways are also moving in that direction. Access points will probably always be the highest percentage, simply because there are so many of them:
There is still a long runway for movement to cloud. Omdia estimates that there are 100 million enterprise WiFi access points, but only 7% of them are cloud managed at the end of 2019, up from 6% in 2018. Omdia further estimates Cloud Managed Networking will grow at a 23% CAGR through 2024.
What is the Market Missing?
The stock is trading at 1.4x consensus gross profit, which is really bottom of the barrel for any technology company (in fact, this the 3rd cheapest across the entire North American tech universe), let alone a niche industry market leader. We believe the market is treating the company as a levered Frankenstein collection of mishmashed assets with no real strategic direction and no competitive differentiation. Looking under the hood, we believe the market has overlooked a positive transformation and is missing several important factors/catalysts that can, at the least, get Extreme trading more in line with the bottom end of its comps group and closer to where it traded pre-COVID ($7-$8).
It's critical to understand that the growth in cloud-managed devices described above is also introducing an attractive new revenue model for Extreme: subscription revenues. In the non-cloud model, the company would sell their hardware, and then try to wrangle a hardware maintenance contract on that hardware. In the new subscription model, the company sells the hardware, has hardware maintenance contracts on the hardware, and tries to layer on a subscription package under its Extreme IQ brand on a fee per device basis. The different tiers are below:
They have a basic free version, then more advanced versions they charge $175/year/device for the Pilot version (plus a cheaper Navigator version). As you can see above, they have a product roadmap to roll out even more advanced (and presumably more expensive) versions of the their ExtremeCloud IQ.
The company has just started reporting some KPIs on their cloud products, as shown below.
It's our understanding from IR/management that the blended average revenue per device (ARPD) is a little under $60 now, so with 1.14 million managed devices at say $58 they would be running in the mid-$60s million in recurring revenue from the current installed base. The reason it is so far below the $175/unit/year ARPD is because Extreme is doing more upfront discounting and has a certain percentage of devices that are under the free model.
They have also started giving out commentary/guidance on the cloud business:
"Specifically, on the subscription bookings as pointed out, they're up 42% (sequentially). We would expect this pace to continue and actually as I mentioned in the prepared remarks, we feel like cloud adoption is accelerating with COVID, so the pace of growth in subscription bookings will remain very strong."- CFO, Q4 Call
We would also note that the CTO, mid-quarter, said they were at 1.2 million managed devices, and on a recent webinar the Extreme product person said they were adding 30,000 devices a month (e.g. 360,000 a year). We would also point to an Omdia survey that suggests the transition will move rather quickly, as we shown below (note Cloud WLAN controller and Hardware WLAN controller).
Taking all these data points together, we can piece together that their cloud business is really picking up steam, probably growing above 30% and potentially accelerating based on bookings trends (we note a large sequential decrease in deferred revenue last quarter and the 42% sequential growth in subscription bookings).
We believe the cloud business currently has roughly the following profile:
1) 30%+ growth
2) 85% gross margin
3) 90%+ customer retention
4) Emerging scale with ~$60 million in annual recurring revenues
We further believe there is ample room to grow that, in several ways. First, by getting more people who are using the free product to upgrade to a paid product. In mid-September they announced "five applications to ExtremeCloud IQ Pilot" at "no additional cost." What they are really trying to do is move those free users (or using the Navigator product) into purchasing a Pilot subscription, as well as enhance the value proposition for those already paying a subscription (improving retention). Second, to convert non-cloud managed devices to managed devices within their own infrastructure. Third, signing new customers like MLB that will take the full cloud package. And finally, adding new tiers of software products like Extreme IQ Co-pilot and Auto-pilot. If we consider those three factors, and triangulate around Omdia's comments that only 7% of access points are cloud managed, we believe Extreme is still early in its the cloud transition story.
As Extreme scales their cloud business closer to $100 million, we wonder how software investors would think about a valuation for this revenue stream. One could certainly argue that a subscription business running at 85% gross margin with 90%+ retention, emerging scale, and a long growth runway would command 6-10x sales, which would effectively cover the enterprise value currently but still only include about 10% of revenues. If you then factor in the remaining recurring revenue, which as we understand comes in at 60-70% gross margin, then you've got a growing, profitable recurring business running closing in on $400 million.
A 3x multiple on that full recurring revenue business, still only 40% of the revenue base, would put the stock at $8.00 (from $4.50 today).
We believe management is still fine tuning how they present this data and, frankly, are still a bit new in how software revenue is conveyed to the street. Looking at sell side notes, who are mostly network equipment guys, we believe they are asleep at the “switch” and are not recognizing this inflection occurring. Instead, they focus on general network equipment hardware sales (which are still important, but we believe the narrative on the stock can change over the next 12-18 months).
Gross Margin Expansion
Until COVID hit, the company was knocking on the door of 60% gross margins, and then took a one quarter lump in calendar Q1 before reverting above 59%:
In the recent pre-announcement, the company stated that they "slightly" exceeded their gross margin guidance of 59-60% in Q1, so we believe they may have finally sustainably hit 60%, or are right at the doorstep.
We believe there are additional tailwinds coming. As discussed before, the mix of higher margin Subscription/Recurring revenues will continue to rise.
Second, their upcoming "universal hardware platform", which is their attempt to move all of their acquisitions under one common hardware platform, should, according to the CFO, lower COGS because they will have "better scale with each of the products they are launching."
Finally, and perhaps slightly longer term and less impactful, their freight costs for shipping product have been elevated as they have had to use Commercial Airlines that charge "twice the rate" because of the few number of flights. As passenger traffic begins to recover, that should be another tailwind.
So, when we consider mix, efficiencies, and some potentially lower input costs for a company that was already right at 60% gross margin pre-COVID, we don't think a 61% or even 62-63% gross margin is necessarily unachievable. That wouldn't be a terrible gross margin for a software company, but for a "hardware company" it's fantastic. You can see below how the rise in subscription revenues is starting to impact the Service gross margins.
WiFi 6 as a Major Tailwind
We are sure many tech investors have looked at themes surrounding 5G. Who will benefit and lose from 5G being deployed? What are the economic use cases? Given the investor focus on and hype surrounding 5G, we think WiFi 6 has been lost in the shuffle to some degree, despite being a similar leap in technology from WiFi 5, and with perhaps more immediate and obvious use cases.
While we will not go into all the benefits of WiFi6 or the history of WiFi, you can see an excellent, down to earth pitch on its advantages here:
Suffice to say, the advances from WiFi 6 apply directly to Extreme's customer base and their desires. Thus, we believe Extreme may benefit more than other companies from WiFi 6. For instance, one of the more major advances for WiFi 6 is it significantly increases the amount of traffic that a network can process at once, which allows for a more seamless experience in high congestion areas. An analogy from the video is that if you think of data going back and forth as lanes of a highway, the old system needed to monopolize the entire highway to do even limited compute tasks, while WiFi 6 exponentially increases both the number of simultaneous processes that can occur by opening up and adding highway lanes and assigning prioritization, but can also effectively increase the “speed limit” and data packet size of each lane. For the "transient" but high concentration use cases of football stadiums (50k+ people in one small area connecting to WiFi) or airports, this is extremely important.
Our understanding is Aerohive (acquired over a year ago) was early to the game for WiFi 6, as they begin readying WiFi 6 network devices in late 2018.
It is hard to quantify exactly what the tailwind will be, but given the obvious benefits to Extreme's customer base, we believe it should be at least a moderate tailwind that does not seem to be factored into the stock in the same way 5G is embedded in "5G Plays."
In addition to the COGS benefits described above in gross margins, we believe the company's recent hardware refresh of their entire product set is more meaningful than usual, as it directly ties into their strategy of taking advantage of WiFi 6, consolidating their redundant products under one platform, and integrating with their Extreme IQ software platform. From the acquisitions the company had made they had a mishmash of operating systems and competing product lines, and a confusing market message in some cases, both to their customers and through their channel partners.
The new Universal Hardware platform is operating system agnostic and is an effort to uncouple the hardware and the operating system. A customer can now buy a piece of hardware using one operating system and then, from the cloud, change that hardware's operating system as needed. Our understanding is this is a differentiator over competitors and a “game changer” that allows for extended hardware lives and more hardware flexibility, as well as reduced need to for onsite changes.
On the Q4 call, the company commented on where they were with this refresh:
"On the product front, we had a specific program that we talked about which consisted in introducing 40 new products out of a total of 57 that was identified, and where 36 of that is largely done. But the next move for us is to look at the universal hardware platform which will be common across our pillars, from Data Center, Campus to edge, and that will drive significant savings in terms of the cost of goods sold on this product."
"We're now 90% complete with our product refresh as of the June quarter in-line with our previously announced goals. We've also began taking customer orders for universal hardware platforms that we'll be shipping in early calendar Q4."
So even though they were able to pre-announce a revenue beat in calendar Q3 (FQ1), their new hardware platform is not even shipping until Q4. We believe the strategic nature of this hardware refresh is the culmination of 5 years of acquisition integration and would expect not only gross margin improvement from inventory management, but an uptick in sales tied to the Universal Hardware concept, but also in relation to WiFi6 and ExtremeCloud IQ.
Operating Cost Discipline and FCF
In fiscal year 2019 the company did $80 million in FCF on a base of about $1 billion in sales (pre-Aerohive, pro forma around $1.1 billion). Since that time, while revenue was temporarily disrupted by COVID, they have materially cut duplicative costs from the Aerohive acquisition, lowering the run rate operating cost base from $93 million to $55 million and their overall cost case from $145 million a quarter to $120-$125 million. This $120-$125 million cost base is allegedly stable, although if sales came in too hot there would be additional sales commissions:
"So going forward, you should be expecting our operating expenses to be between $120 million to $125 million. And the only way it's going to go up is if travel goes back to about $3 million to $4 million a quarter, and I don't see that happening in Q1 nor in Q2, and if the bookings are so strong that we end up paying more commission to our salespeople, which I'd be delighted to do, by the way. But there's no reason why operating expenses should be going up outside of these two things. And therefore, we're going to get operating leverage. Every time the revenue goes up by $10 million sequentially, about 60% of that goes into gross profit and the majority of that 60% goes into the operating profit."- CFO, Citi Global Technology Conference, September 9th
At the same conference, the CFO commented on FCF generation:
"So, the next couple quarters, it's probably going to be $15 million to $20 million. As we hit Q3 and Q4, it will probably be $20 million to $25 million."
This would imply around $80 million in FCF generation, similar to 2019. Does this make sense given the other disclosures?
If we assume $25 million in cash interest, $5 million in taxes (low tax payer with $287 million in US Federal NOLs), and an operating expense base of $122.5 million a quarter (while assuming depreciation roughly matches capex), we get the following FCF sensitivity table (ex-change in working capital):
You can see that if the company averages $1 billion in sales on 60% gross margin, they will do around $80 million in FCF, all things equal, so that seems like a fairly reasonable bogey.
In the most recent pre-announcement, while the company did not exactly specify FCF generation, they hinted at it.
"The combination of improving business trends and continued cost and expense control drove solid operating leverage, and in turn strong cash flow. This allowed us to repay $20 million of our $55 million revolving credit facility during the quarter and reduce our net debt by approximately $24 million sequentially – all while maintaining ample liquidity as cash and cash equivalents remained relatively stable compared to the prior quarter."
We believe this suggests FCF generation somewhere between $20 and $24 million, which would be above the CFO's original forecast of $15 to $20. It's certainly possible there was and will be working capital tailwinds, as the CFO has talked about the $60 million inventory balance being "too high" and they should continue to build their deferred revenue balance.
To sum, we believe the combination of having a stable operating cost base, rising gross margins, and high confidence in the upcoming quarters' results should result in numbers moving up, particularly FCF, which is currently estimated to be $41 million compared to the $70-$100 million we think they can achieve over the next 12 months.
Given at least 60% of the business is still non-recurring, we won't pretend the stock can't be driven by quarterly fluctuations, however we do think forecasting visibility is increasing with the rise in recurring revenue and other initiatives the new CRO has put in place. The increased visibility is something both the CEO and newly hired CRO (who has a strong track record of software transitions). We would suggest anyone interested in the name listen to their investor session at the User Group Meeting:
The company has $396 million in gross debt, vs. $193 million in cash, so net debt of $203 million. If they can generate the $80 million of cash they are forecasting next 12 months, that net debt level can materially decline to a much more manageable level, and FCF/share and EPS can also be helped, albeit somewhat modestly given fairly low interest rates.
Further, the CFO has made it very clear this is where all of the cash is going in the short term. From the Citi Conference:
Jim Suva Analyst, Citigroup Global Markets, Inc. Q : Still sticking to your financials, can you talk a little bit about priorities for cash, cash flow, uses for cash? What's your management's priorities for using cash?
CFO: So we've got three top priorities. The first one is pay down debt. Second one is pay down debt. And the third one is to pay down debt.
And from the BMO Conference:
Our focus is really to pay down debt. You're all aware of that at BMO since you're part of the bank syndicate that supports us. But right now we really want to generate as much free cash flow as possible and take out overall debt. If you recall we've got roughly $370 million on the term loan A, plus $55 million roughly of revolver that's outstanding, and so we're going to generate free cash flow this quarter. Hopefully we're going to generate more free cash flow in our December quarter and the focus will be to pay down debt. So we don't have any plan to do share buybacks and we'll continue to be opportunistic in M&A but obviously our balance sheet currently really restrains us from doing anything significant. So paying down debt is my top priority. And that will really come from the recovery in our operating profitability combined with the reduction in inventory. We're currently north of $60 million worth of inventory on our books. I consider that it's too high and so there's a big focus for us on reducing that inventory.
We think leverage has absolutely been an overhang on the stock but it may be the market is underestimating how much debt they can pay down, and how that will change the complexion of the company in the next 12-18 months. For instance, if they were able to pay down an additional $80 million in debt from Q2, 2021 to Q1, 2022 (12 months), suddenly their net debt is $120 million on an EBITDA base moving to $100 million, so really pretty modest leverage from where they are now.
Extreme is either cheapest or second cheapest on every relevant metric versus a group of network equipment comps:
A few points of attention on this table:
1) Extreme would have to rise 33% just to get to the second cheapest comp (RBBN) on EV/Gross Profit
2) We believe Extreme will generate materially more FCF than consensus, thus their FCF discount may be even more extreme than represented here
2) While Extreme is growing sales slightly slower than the comp group, because of their business transition they are growing gross profit and EBITDA materially higher than the comp group
3) While we believe revisions will move higher, Extreme has still dominated relative revisions versus the peer group
4) Despite growing faster, being cheaper, and having better recent revisions, the stock has been a huge laggard and by far the worst performing group YTD, 1 year, and 3 years.
5) Despite having a higher gross margin than the comp group, the EBITDA margin is 11% lower, which suggests to us how it could be attractive to an acquirer
Given where the market is for Tech Stocks and given that Extreme's gross margins are shooting up with a much higher % of revenue now recurring, we believe the stock at .75x sales does not make sense. Although it has certainly been lower, margins were also lower, and we would argue revisions are now riper to move higher