This article was highlighted for PRO+ Tech subscribers - find out more about PRO+ here.
In the days of big-headline partnerships within the cloud industry and minimal visibility into possible financial benefit, there lies one company, VMware (VMW), retaining the throne of the hybrid cloud and being the center of the migration process from on-premises to public cloud.
For those who are not tech-savvy, understanding VMW's products and opportunity can be a challenge. That's where I fit in. After years working in the sell-side covering "IT-Hardware" companies and, most importantly, talking to industry contacts, it became easier for me to understand this company's technology and what value it gives to its customers. And, that's part of the reason why I believe (or "we" for dramatic purposes, see bio) the shares of VMW are undervalued; Investors don't actually understand the market opportunity of VMW and, thus, are valuing it as a hardware company that is subject to the "booms and busts" of corporate IT budgets. We believe otherwise; VMW is highly immune from the cyclicality of the IT Hardware space, and its products, which are basically uncharted with no clear TAM, are underappreciated by the market.
Our price target of $195 per share is based on applying a 20x multiple of next year's estimated free cash flow of $3.9 billion, per FactSet. A 20x multiple is still cheap compared to the 33x multiple IBM (NYSE:IBM) paid to VMware's cousin, Red Hat.
We believe that the market is not giving full credit to the promising non-vSphere portfolio and that Dell EMC (DELL) may acquire the 12% stake it does not currently own of VMware.
Important notes before reading: VMware occasionally breaks down its bookings stream by product. The math behind estimating each product's revenues is based on various assumptions, which is why we used VMware's top sell-side analyst, in our view, Karl Keirstead, estimates for the revenue breakdown of each segment.
Also, it's worth noting that valuing vSAN and NSX is done to give some color on the sizes of these segments and is not used in any way for our valuation for VMW as a whole. After all, buying a stock based on SOTP is the most naive form of investing, we believe.
VMware is a software company that deals with virtualization of the data center. Its main product is vSphere, the server virtualizer, which contributes to 60-70% of revenue mix by DB's estimates. The company has introduced vSAN, which virtualizes storage within the server, or direct attached storage, and NSX which virtualizes networks. The company is also a respected player in the containers market. We will deep dive into each product and what it means to VMW.
To understand VMW, readers must understand what's virtualization. In its simplest form, it's a software stack placed on a server through a physical layer called a "hypervisor" that enables multiple operating systems to be placed on one server and thus sharing the server's resources across multiple end-users. The technology increases data center's utilization meaningfully, which explains the rapid adoption of VMW's technology. However, despite its usefulness, the product is negatively affected by the migration to the public cloud as it thrives on-premises (within the private data center, aka traditional data center).
For VMW, vSphere is a mature technology that has been the core of VMW since its introduction in 2009. Since that time, the company has made several sub-one billion dollar acquisitions that may turn out to be disruptive. Below, we listed VMW's most important acquisitions.
We believe that the two most promising acquisitions are VeloCloud and Heptio. We won't go into details about these acquisitions as they're still small to move the needle for a $60 billion company. However, it's worth noting that VeloCloud is disrupting the WLAN market by "software defining" it, while Heptio is embracing the Kubernetes management tools for VMware's containers. We will talk more about containers in this report.
VMware is 82% owned by Dell EMC as it was part of EMC which Dell acquired in 2016. However, VMware has been publicly listed since 2007.
VMW Products Explained
As said, the reason we are long VMW is that its products are uncharted (with minimal quantitative data), and it does not have a clear competitor to "grease the wheels" of the valuation process. We believe that VMW's emerging products, vSAN, NSX, and containers are not fully appreciated by investors and sell-side analysts and that a higher valuation should be granted to such portfolio. In this report, we try to go into details about VMW's non-vSphere portfolio and quantify the market opportunity of these technologies.
Before explaining VMW's vSAN, it's worth explaining the storage market briefly. There are two types of storage in data centers; direct attached storage (DAS) and external storage.
DAS is a group of HDDs and SSDs that are attached to the server. It's fast, simple, and cheap. DAS is only suited for light workloads, and it's not software-based.
On the other hand, external storage is shared, rich in software, expensive, and complex. NetApp (NTAP), HP Enterprise (HPE), Pure Storage (PSTG), Dell, and IBM are the main vendors for enterprise external storage.
VMW generates revenues from vSAN by selling it as a license and by selling maintenance contracts over a 2-3 year period, subject to renewal (revenues for maintenance are recognized over that time frame), same as vSphere. The mix of license/maintenance is roughly 60/40% for the announced $1 billion vSAN bookings, per VMW. vSAN is priced per-CPU, and its price can range from $2,500 to $5,500, without including discounts that can reach 50% in some cases.
vSAN's primarily market is Dell's VxRail, a hypercoverged infrastructure product which combines vSphere, vSAN, and NSX for networking virtualization (picture below).
In 2018, Dell has materially cranked up its sales effort around VxRail in 2018, directly compensating an army of sales reps for its HCI product which means that the true ramp-up of vSAN sales has only been taking place for two years despite the product being existent for more than 5 years.
A chart speaks a thousand words, so it's worth illustrating the HCI market and VxRail/Rack's market share per IDC (subscription required).
VMW has disclosed that roughly 25% of the price of VxRail can be contributed to vSAN. Last year, Dell has started pushing its sales force to sell more of VxRail instead of Dell's XC Series, which is based on Nutanix software. Per IDC, Dell sold $45 million of XC Series HCI in 1Q17 and $80 million in 1Q19. During the same time, Dell sold $65 million and $294 million of VxRail. This can show the effect that Dell can have through its massive sales force in pushing out a product.
It's worth noting that vSAN's true number of bookings is a matter of subjectivity as VMW estimates vSAN's value within the sold VxRail, and these estimates are based on unknown assumptions that can be skewed to a degree. What makes the matter more complicated is that the industry is known for deep discounts that exceed 50% for enterprises, so knowing the exact contribution of vSAN within VxRail is a matter of more art than science. However, we believe that if VMW would overestimate vSAN's potential, it would have to underestimate vSphere's revenue growth, given that they are sold in one bundle. We believe it would be naive for VMW to do that, and thus would rule out such scenario as investors care more about vSphere, which represents 60-70% of VMW's revenue stream and is considered the core product, than care about vSAN's growth numbers that are coming of a small base.
As we said, vSAN is operating in an uncharted industry with very few public players. To us, Nutanix (NTNX), a software-only company, is the closest competitor to vSAN given that it sells a whole HCI product which includes a free native server virtualizer or VMW's vSphere (depending on the customer needs), the Nutanix Acropolis Distributed Storage Fabric storage virtualizer, and "Nutanix Flow", which is a competitor to VMW's NSX.
At first glance, Nutanix would look more of a competitor to Dell's VxRail than vSAN. However, the fact that 60% of NTNX's HCIs are based on VMW's vSphere, given the partnership between both companies, makes vSAN and NSX the only products that NTNX can differentiate from.
The main use of vSAN is for VDI, or virtualized desktop infrastructure. VMW has "VMware Horizon", while Nutanix has a partnership with Citrix (CTXS), the market leader with a 43% market share per IDC. VDI is virtualization of the desktop where the operating system, server, and storage are located in the data center and not on the end-device. It's basically using a monitor connected to the data center through the internet instead of having a whole PC. The issue with VDI is that computing and storage at the data center require huge amount of servers and storage, which made adoption of VDI devices expensive and complicated. vSAN, NTNX, and Atlantis (private) solved the problem by virtualizing servers and storage, thus increasing server/storage utilization and reducing costs.
Note: We used "may" because Gartner and IDC disagree about who's the market leader between the vSAN and Nutanix. After all, both companies report their own estimates using their own channel checks and value recognition assumptions.
Valuing vSAN is not a simple task. Amateurs would point to Nutanix P/S multiple of ~4x and apply it to vSAN's estimated revenues of $600 million for the current fiscal year. We would argue that VMW is in much stronger position to push vSAN growth further given several reasons:
1) Dell's 30,000 sales force and vSphere's 200,000 installed base, which reduces customer acquisition cost.
2) Nutanix burns cash ($76 million in its latest fiscal year) and is transitioning to a subscription-based model, which pressures the valuation, while VMW is significantly profitable (33% operating margin vs. -50% for NTNX).
3) Nutanix Enterprise Licensing Agreements (ELAs) are of $1 million per quarter, a rounding error, while that number is more than $400 million a quarter for VMW. It's worth noting that ELAs make it easier to sell a product, given close extensive relationship with the enterprise (read here about VMware's types of agreements).
4) The fact that Nutanix does not have a strong partnership with any cloud hyperscaler makes NTNX more vulnerable to the migration off on-premises (both hybrid and public cloud). While the company has announced the Xi Cloud Services with Google Cloud (NASDAQ:GOOG) (NASDAQ:GOOGL) and AWS (NASDAQ:AMZN) well before VMW made any hyperscale partnership, the Xi did not seem to gain traction, and there has been limited discussion by NTNX management about it. On the other hand, VMW has a strong partnership with AWS, which makes it a more solid bet on hybrid cloud.
As such, we believe that vSAN fits more into the high-growth software class than the deeply challenged Nutanix. We used the "Rule of 40" to screen competitors (rule of 40 says that attractive companies are those that have an FCF margin plus revenue growth of 40%).
Given that vSAN is increasing its bookings by 60% a year per management, it easily fits into this category without even trying to figure out what the FCF margin is. The average and the median sales multiple is around 12x, which gives vSAN a value of $10.2 billion using the estimated $850 million in next year vSAN revenues that Deutsche Bank analyst Karl Keirstead estimates.
When asked to prioritize VMW's new product initiatives, VMware's CEO listed NSX, hybrid cloud (of which vSAN is one component) and containers, in that order. He followed, "obviously I love vSAN, but it's not as important as NSX".
To explain what's NSX there are two ways; a simple way, and a more sophisticated way. The easy way is that VMWare is doing to networking what it did to servers, virtualizing the network so it becomes software-defined, which makes it smarter with a smaller appliance footprint.
The first thing I think about when trying to explain NSX is to think of it as the GPS. Previously, we used to print out a map to find the best way to reach our destination. If we faced an unexpected congestion, much time would be wasted trying to figure out our way using another route. Today, we just plug in our destination on our smartphone, and on our way there, Google Maps (I'm no fan of Apple Maps) would suggest to us the best available route to avoid a possible congestion. Network virtualization works the same way, instead of going from node-to-node in a straight line, the software would choose the safest and the shortest route.
But the simple definition does not address the market opportunity of NSX. The other way, unfortunately, does.
We often hear a lot of major security breaches incidents. Think of Marriott (MAR), Capital One (COF), and Facebook (FB). We're not talking about the biggest of "all-time" breaches. We are talking at 2019 breaches within companies that spend hundreds of millions of dollars on security.
There's one thing in common with all security breaches; hackers tend to get inside the system through tricking some high-end user, and once they're in, they install the "C2 Infrastructure", which is used to command the controlled device, on each device while moving laterally within the data center. The process would continue, and the attacker would plant an "idle malware" on each device it passes through. Once an "active malware" is discovered by the security software, the attacker would wake up an idle malware and continue the attack. For now, we have just explained how they get in, but not how attackers steal information.
For them to do so without being detected, the attacker parcels the data into small, encrypted payloads during the exfiltration process. This segmentation of data, since it's happening at the very low level, is hard to detect with segmentation security software that operates at the macro level. Here comes in the role of the technology behind network virtualization, "micro-segmentation". While micro-segmentation has been known for a long time, the adoption of this technology was operationally and financially infeasible before engineers at VMware and Cisco (NASDAQ:CSCO) changed the status quo few years ago by successfully applying virtualization technology to security.
For a long time, the security industry has focused on protecting the network by building firewalls surrounding the data center parameter while giving minimal importance to securing the data center from within. This is now changing with virtualization technology that makes "micro-segmentation" a much-less costly process. Whether it's Cisco's ACI or VMware's NSX, the core thesis of protection is the same.
Just as vSphere didn't seriously affect HPE/Dell, and vSAN didn't do so to NetApp/Pure Storage, we don't believe that NSX poses any serious threat to Cisco or Arista (ANET). On the contrary, we believe that network virtualization which secures the data center from "within" is complementary to networking appliances that secure the DC "parameter". Thus, we do not believe that NSX is facing any deep-pocketed competitor. It's worth noting that, in theory, software-defined products should commoditize IT hardware, but it's hard to prove the practicality of this point for now.
While Cisco does have its own software-defined networking product, ACI, and one might think its strong presence in the enterprise networking appliances market (59% market share) would crush any initiative by VMW to get into this market, we would argue that what matters is who controls the hypervisor, not the appliance.
To elaborate more, the hypervisor, which is the software-based glass layer in the server, is the brain of the virtual machine. vSphere, vSAN, and NSX (which all make an HCI if found in one appliance) are all installed on the hypervisor which by itself acts as the backbone for the micro-segmentation process as all information passes through the hypervisor. With VMW's 200,000 customers using its ESX hypervisor, VMW has the same access at the same level that Cisco has through its networking appliances (routers, switches, hubs, and modems). It's worth reminding that VMW would not be taking share from Cisco in the hardware market, instead, both players would penetrate the existing hardware networking appliances with software.
As vSAN, applying a multiple on NSX is more of an art than science. We believe that Palo Alto Networks (PANW) may be the best comparable to NSX as one of NSX's main missions besides speed of data transfer is network security. Going back to the work of DB's Karl Keirstead, he estimates that NSX revenue would increase by 37% next year. Using our regression model for PANW, which takes into account the activity of its multiple related to growth estimates since the company went public, we believe that NSX would be valued at 6.8x NTM revenues. Keep in mind that we're being conservative here as the operating margin for NSX should be much higher than PANW, given that NSX is part of a huge company, and thus, the incremental operating margin is much higher. Deutsche Bank believes that next year's NSX revenues would be around $1.3 billion. Applying 6.8x multiple on that would give us a conservative valuation of $9 billion for NSX.
Source: Westlake Research, FactSet
Before digging deeper into this section, it's worth explaining the difference between containers and virtual machines. As said, VMs virtualize the hardware, where an application would think that it solely operates in its own server and storage while in fact it's part of tens of VMs within one server. Enterprises benefit from virtual machines as it increases utilization rates for hardware, which means less capital spending on such expensive items. On the other hand, containers virtualize the operating system itself where an operating system can work within any hardware. Containers mainly benefit the developer as he/she would write a code without worrying about its compatibility with a certain hardware. Containers also increase the utilization rate of hardware as they would allow multiple operating systems to be installed on one hardware appliance.
During my time in the sell-side, I asked my analyst what's the biggest threat facing VMware, he responded, "containers". And that's what most who are not familiar with VMware believe (P.S.: we do not cover VMware).
After digging deep enough, we concluded that containers and virtual machines are complements, they do not supplant each other except in rare circumstances. I would like to share the following from Gartner (subscription required).
A majority of organizations today run containers in a VM due to the strong degree of hardware isolation that VMs provide. We have seen the following developments in the past few years to enable easier coexistence of VMs and containers:
Easier provisioning of containers on VMs: Hypervisor vendors such as VMware and Microsoft have introduced technologies such as vSphere Integrated Containers (VIC) and Hyper-V containers respectively that make it possible to wrap containers in a VM, often using lightweight operating systems, with a common control plane for deploying and managing them.
Managing VMs with container orchestration: Most organizations deploying containers use Kubernetes as the scheduling and orchestration layer. KubeVirt is an open-source project that potentially allows VMs to be managed using Kubernetes tools. KubeVirt is a Kubernetes add-on consisting of custom resource definitions (CRDS), controllers and an operator leveraging a range of Kubernetes extension mechanisms. Using Kubernetes and KubeVirt allows organizations to launch containers and virtual machines on the same cluster with a similar underlying infrastructure.
Micro VMs: While there are many benefits to running containers in a VM, there are two key disadvantages: (1) resource and performance overhead and (2) high acquisition cost. Vendors are working on container-optimized, lightweight VM technologies (“micro VMs”) as well as on sandbox container runtime, which limits kernel functionalities to enable isolation without VMs. Examples of micro VMs include Kata Containers, spearheaded by Intel (under the OpenStack Foundation) and Firecracker, which is an open-source project created by AWS.
VMware embraced containers with its Cloud Foundry initiative, which was transferred to Pivotal Software (part of Dell at the time and now part of VMW following the recent acquisition). The container market is uncharted too with players such as Red Hat's OpenShift, Docker (private), and AppXXXXXX (private).
PVTL's subscription revenues are more than $1 billion run rate per its latest filling before it went private. It's hard to quantify the market opportunity for containers, but moving legacy, and new, applications to the cloud should be a long-time tailwind for VMW. To quantify the growth ahead, it's worth noting that PVTL increased its subscription revenues by 40% Y/Y in the first half of 2019. While PVTL was trading at an average of 4x sales, the market was discounting the fear that Dell may raise more capital and increase PVTL's float. Under VMW, we believe that PVTL can be valued at 8-10x sales, which is in line with what other subscription-based software companies are trading at. Also, the level of integration with VMW should not be underestimated.
PVTL's main competitors is Red Hat's OpenShift. The containers market is so early that it's still in a "land grab" stage of development. However, the fact that VMW embraces both, VMs and containers, two uprising technologies, makes its standing unique.
Only amateur investors would think that VMW's SOTP is clearly worth more than its enterprise value. The market is not dumb to give such company with such promising portfolio a crazy-cheap multiple. Even if the market does not deeply understand this non-vSphere product, the company's "low double-digit" growth and its 30%+ FCF margin warrant a high multiple.
The stock is currently trading at 16x NTM FCF per FactSet, 15% higher than its 4-year average. While we prefer buying at (or below) average valuation, the stock looks now attractive to where it was trading at six months ago.
That makes us more comfortable buying at these levels is the valuation relative to its mature peers. VMW stock is trading nearly in line with its peer group. We believe that the stock should trade at a premium. Why at a premium? Because the company has an above-par score. The "Rule of 40" proves VMW's operational excellence. Needless to say, a superior operational performance requires a superior valuation.
Proof behind our main investment thesis
As stated previously, we believe that the market does not fully understand VMW's products. And we believe investors have total right to ignore this black hole. The company's products are uncharted with no clear competitors, and it's extremely hard to understand the market opportunity without expert consultation. I personally have spent more than 25 hours trying to deeply understand VMW's portfolio and yet there's still a gray area for me. The following charts prove why we believe that investors think more of VMW as a company that is tied to the hardware refreshment cycle, and not a software company that enterprises would spend on to reduce their expensive hardware budgets.
Since March 2019, the time when the market recognized that we are entering an IT-spending slowdown, the number of downward estimates revisions by sell-side analysts was excessive for most IT-related companies with the exception of VMW, Oracle (ORCL), and CSCO.
Yet, VMW's multiple has dropped as if it had peer-like estimates revisions and not only a couple.
We believe that VMware should trade closer to Microsoft's (MSFT) multiple of 24x NTM FCF per FactSet. Before the latest selloff of IT Hardware stocks, VMware was getting closer to trade at MSFT's multiple. We believe that "at-par" valuation with MSFT is not a fluke.
The reason? Wall Street only cares about numbers, eventually. And numbers are on VMware's side. Microsoft is expected to grow revenues by 15% next year with a 30% FCF margin. On the other hand, VMW is expected to grow revenues by 14% while having a 36% FCF margin. Both are exposed to the IT budget in an indirect way and both are software-based cloud-exposed companies.
Would Dell buy the rest of VMware
We also believe that it's possible that Dell might buy the rest of VMware (12%). The reason? Its access to VMW FCF. As found in our table below, VMW contributed 68% of the parent company's FCF since 2015. While Dell does own the majority of VMW, management has stated on numerous occasions its commitment to not use VMW cash to pay back any of the parent company's debt.
And as it relates to VMware, I mean look, nothing has really changed in terms of how we operate with VMware, and we want to make sure that, they're growing their business appropriately, and so there is no intent or plans or I'm not modeling getting any cash from VMware. (November 29, 2018 FQ3 earnings call, FactSet)
While the reason for Dell not using VMW's cash seems voluntary at first, it's worth noting that, if Dell does use part of VMware's cash, it would signal to the market that the company is willing to juice this company that is highly dependable on R&D and is operating in a fast changing industry, which would negatively affect VMW's valuation. Thus, Dell's best option would be to buy the rest of VMW, which would take off Wall Street's pressure and enable Dell to use all of VMW's FCF as it sees fit. The beauty of such a deal is that it would be slightly dilutive if even when excluding any cost synergies (like investor relations and some duplicate positions) while assuming a 35% buyout premium.
If our math is correct, vSAN and NSX have a combined valued of ~$20 billion, or nearly one-third of VMW's current market cap. However, we do put a higher valuation on all VMW's product portfolio when combined, given the huge potential of cross-selling the products. It's now well known that the new data center would be software-defined and that hardware would eventually be commoditized. While the process won't happen overnight, it gives companies like VMW a long-term tailwind. We believe that VMW's current margins are near trough levels as products like vSAN, NSX, and containers are still at their relatively early stages, and as these products ramp up, incremental margins should be huge, given the pure software nature of the product.
Based on our personal estimates, we believe that 20x NTM FCF would be an appropriate multiple for VMW, which makes $195 per share our price target. Thus, we rate VMW as a "buy" with a two-year time-frame.
Disclosure: I am/we are long VMW, PSTG. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.