After recently completing a fairly comprehensive review of Microsoft (MSFT) and its valuation (I recommend you read that piece first if you haven’t already), I wanted to take a bit of a deeper dive into Azure. The cloud computing division of Microsoft has been responsible for a significant portion of the company’s recent growth and the future seems rather bright. Though, with Amazon (AMZN) still dominating the space with Amazon Web Services (“AWS”) and Alphabet (GOOGL) threatening to become a more significant competitor with Google Cloud, this bright future isn’t some foregone conclusion. There is still much to be seen in the cloud war and this article aims to determine how Microsoft looks to fare.
What is Azure
Before we get too far, it’s important to understand what exactly Azure is so that we can understand how its market and how it competes in that market. If you believe you already have a pretty solid understanding of Azure, or just don’t care, feel free to skip this section. Azure is a cloud computing service, offered by Microsoft. Across Azure, Microsoft offers a range of different products for consumers to choose from. I’d like to take a look at some of the service’s core offerings, providing a brief overview of each, so that investors can better understand the technology they’re dealing with here. Before we get to that, however, let me introduce you to the cheekily-named aaS acronyms. The main acronyms to know are Software-as-a-Service (“SaaS”), Platform-as-a-Service (“PaaS”), and Infrastructure-as-a-Service (“IaaS”).
SaaS is rapidly emerging as one of the most important components of the digital age. Proprietary algorithms and massive data sets, created by and constantly improved upon by large tech giants, go far beyond what others are capable of creating on their own. While this is largely true even for most companies in the S&P 500, this is especially true for small businesses or individuals. One of the most critical components of SaaS is machine learning, or artificial intelligence (“AI”).
Let’s be clear, AI is not at the level where computers are gaining sentience. Nor is that the goal here. AI’s applications in businesses are numerous and, as consumers, you’ve likely seen it in the form of online chat support or product recommendations. In 2020, Oracle (ORCL) estimated that 80% of companies used chatbots, or planned to use them by the end of the year, to streamline the customer experience. Yet this only touches the surface of how AI hopes to impact the business world and global marketplaces.
Intelligent data analytics is a primary focus for machine learning developers and users moving forward, given how prevalent data is becoming in this digital age. As data sets continue to grow in size and complexity, it no longer seems feasible for humans to be the ones responsible for analyzing such data. I may be alone in this, but this calls to mind a scene from the 2015 film, The Big Short, in which Michael Burry orders a recent hire to compile a list of the top-twenty best-selling mortgage bonds, with all of the mortgages within each one. Understandably, the associate is a bit taken aback by this request, as the manpower he’ll need to dedicate towards manually completing the task is quite severe. After receiving the list, and examining its contents himself, Dr. Burry comes to his famous conclusion that the country is in a housing bubble.
It’s mundane operations like these, and many others, that should be automated so that employee time and resources can be used more efficiently. Note, it’s still a human directing what the desired outcome of the analysis is, algorithms still need to know what they’re looking for to be effective, so we’re not getting rid of the human here. It’s all about using resources, and employees, more efficiently. However, AI may soon begin augmenting business decisions as well.
In large businesses, the decision-making process is often filled with bureaucracy, meaning even insignificant decisions could take months to be finalized. Not to harp on popular culture too much, this reminds me of a scene from Ford vs Ferrari. After a crushing defeat at Le Mans, Henry Ford II asks Carol Shelby why he shouldn’t be fired on the spot and the racing program terminated. Shelby responds simply, “All due respect, sir, you can’t win a race by committee.” The point being, Ford’s corporate bureaucracy will never be able to overcome the nimbleness of Ferrari and, in order to succeed, it is imperative to act swiftly and correctly. With the guidance of AI, decisions could be made faster and more logically, with some estimates citing an improvement in an organization’s profitability of 5%.
To build a lead on this front, Microsoft has been developing Project Brainwave. The goal of Project Brainwave is to provide ultra-low latency real-time AI solutions without increasing complexity and overhead. Microsoft aims to achieve this by allowing customers to use its neural processing units (“NPUs”) based upon its field-programmable gate arrays (“FPGAs”). I recognize that I just threw out some nonsensical jargon, so allow me to explain exactly what this means.
Think of an FPGA as an alternative to graphic processing units (“GPUs”). GPUs were originally designed to improve the loading time and quality of graphics, but have now found an alternative use in machine learning systems due to their capability to process multiple pieces of data simultaneously. However, GPUs aren’t perfect. Relying upon transistors for their power input, the processing units run at high temperatures when in use and require significant cooling. This makes them rather costly to operate and maintain. This high strain also causes GPUs to wear out, meaning they may only last around three years. So this is where the FPGA comes in.
As the name (specifically the ‘field-programmable part) suggests, FPGAs are able to be customized for different uses after they’ve been manufactured. FPGAs typically last 2 – 5 times longer than GPUs and require significantly less power to operate due to lower cooling requirements. Obviously, there must be some sort of catch, otherwise, they would be more utilized. Right? To the individual consumer, certainly. While the customization of FPGAs is one of its greatest assets, it is also one of its greatest hurdles. FPGAs are incredibly complex to program, requiring a highly experienced team to get it right. However, for a company like Microsoft, that’s not an issue. The programmability of FPGAs will allow the company to continue to iterate on its setup, enabling the company to provide the optimal solution for every situation and take advantage of future industry improvements. Where a GPU is the chainsaw, the FPGA is the scalpel.
So we know why the company uses FPGAs, and what they are, but what about NPUs? NPUs are essentially another step to optimize FPGAs for machine learning capabilities. Modeled after theories of how the brain works, with millions of artificial neurons, think of NPUs as GPUs, but designed only to process machine learning applications. They are far more efficient then standalone GPUs when completing tasks, in some cases even in the realm of 10,000x so. Using an FPGA as the base for an NPU, rather than a GPU, allows Microsoft to create incredibly effective processing units for its machine learning with a focus on continuous improvement.
With the hardware out of the way, we can begin to understand exactly what is meant by the goal of Project Brainwave. These immensely powerful NPUs, which are effectively artificial brains, can make incredibly complex computations, analyzing absurd amounts of data, in fractions of a second. Project Brainwave makes these incredibly complex and powerful machines available to users of Azure. Microsoft says that Brainwave enables the “trifecta of high-performance computing” with low latency, high throughput, and high efficiency.
The applications of such a technology are numerous and the company partnered early on with Jabil (JBL) to examine its use case in manufacturing. The manufacturing services giant’s interest in the technology lies with the screening of products on assembly lines. With a live video feed, Project Brainwave has the capability to detect defects and divert them away from the downstream assembly process. The service aims to improve upon manufacturing efficiencies, and ultimately eliminate the need for manual inspections.
Microsoft’s Doug Burger was the pioneer behind the idea to use FPGAs for NPUs, making the company a first-mover in the sector. While GPUs offer an alright backbone for NPUs, it is becoming increasingly evident that FPGAs are the better option for sustained development in a rapidly progressing field. Being the leader in this space, as Microsoft appears to be, should serve the company incredibly well down the line as AI and machine learning only become more prevalent.
Even beyond the realm of AI, Project Brainwave stands to contribute to greater workload efficiencies. While software continues to push closer to optimization, engineers are struggling to find ways to improve performance. Search engines, for example, need to provide results to their users as fast as they can. However, the quality of results also matter. As engineers work to refine the quality of their results, they can’t sacrifice on response speed. Additionally, more data becomes available every day, meaning that these algorithms have more data to sift through than ever before. With highly-optimized programs, implementing these changes without compromising on speed is incredibly difficult. However, with the transition to NPUs based on FPGAs, engineers at Microsoft were able to dramatically speed up the search times of Bing and improve its search accuracy. This is more of a PaaS application, though I thought it was fitting to include here due to the conversation of Project Brainwave.
Data analytics also look to be dramatically improved through the utilization of data mining. Put simply, data mining is an automated process of collecting large sets of data. For example, algorithms can scan the internet to find how the public seems to be receiving a new product or ad campaign. With social media now providing an open source to the opinions of millions of people, data has never been so widely available. However, it would be impossible for a person to collect that data all themselves. Using AI to sift through and organize it all means that these companies can actually put this massive databank to use. AI also plays a major role in determining where to target certain ads, improving the effectiveness of campaigns and allowing advertisers to get more value for their dollars. Currently, Microsoft is previewing its Azure OpenAI platform, which is aimed at pairing advanced language models with Azure’s other enterprise capabilities. This looks to be the building blocks of Microsoft’s future AI applications.
Considering the immense potential of these various applications, it shouldn’t be of much surprise that SaaS is expected to be the largest section of cloud computing moving forward. By the end of 2022, 78% of companies plan to rely almost exclusively on SaaS for their applications. As the technological world continues to progress, SaaS will only get more important.
IaaS is probably the most intuitive of all the “aaS” applications. Utilizing a pay-as-you-go system is often an effective way for companies to mitigate overhead and costs when dealing with their data. It’s also likely what most people most commonly associate with the cloud. Microsoft Azure’s storage services enable companies, or individuals, to remotely store their data. Think Gmail, or Apple’s (AAPL) iCloud.
Cloud storage solutions may be attractive for a number of reasons, though it can really just be boiled down to convenience and cost. Cloud storage enables users to access their data wherever they are, so long as they can connect to the internet, with multiple users accessing the data at once. Expanding storage is also made far easier, as users need only to upload the files in question to the Azure cloud and it will be taken care of. While, oftentimes, you’ll end up paying a premium for convenience and reduced overhead, the fact that these cloud data centers hold the data for tens of thousands of enterprises means that there can be a bit more efficiency of scale. Additionally, the big data sets that companies are analyzing need to be stored somewhere. As more and more data gets collected every day, it may begin to make more sense to store data in the cloud due to its scalability and accessibility. There are plenty of reasons to switch to a cloud-based storage solution and, as a core offering for most cloud-based products, Microsoft’s Azure is more than an acceptable option.
For those that are largely uninitiated to the complexities of cloud computing, it may be somewhat surprising to see storage listed as just one small part of a larger product. To most people, their experience with the cloud is limited to only storage. Think iCloud or your Gmail account’s maximum capacity. However, when it comes to a fully-fledged cloud computing platform like Azure, storage is just a small part of a bigger picture. It’s for the best too. Logically speaking, any company with money can create data storage centers. What’s significantly harder is developing all of the proprietary software and platforms that organizations run on. In this sense, the barriers to entry are quite high as very few companies have the resources to even try to compete.
One of Microsoft’s greatest assets in this new fight for cloud computing, may also be the hold that Windows has on the private server sector. In 2019, Windows held 72.1% of the global server market. While AWS claims to be “the proven, reliable, secure cloud for Windows,” I think it’s probably fair to say that Azure offers better integration here. For businesses looking to migrate from their Windows servers into the cloud, Azure likely offers the easiest transition.
Another simple, yet popular, IaaS application that Azure offers is virtual machines “VMs”. These are far less complicated than they may sound and simply enable users to run a multitude of operating systems off of one device. Think of it like this, you have your Apple computers that run Mac OS and you’ve got the rest of the world which, primarily, runs Windows but can also run something like Linux. Applications are tuned for each OS, though some applications only support one OS. For this reason, many organizations ultimately run across more than one OS and, as such, need to switch between machines to do so. A virtual machine eliminates that need, reducing costs and complexity of operations. VMs have also found a new use case in this age of work-from-home and consistently disrupted work schedules, as they allow for far greater versatility without forcing companies to sacrifice efficiency. Because the hardware is now secondary to software, it’s much easier to continue work operations even if employees are out of the office. This may also contribute to Windows sales growth, at least at the enterprise level, as users without Windows devices may now choose to run the software.
Azure also offers options for companies to offload their servers into the cloud. Similar to most other IaaS applications, this is ideal for reducing company overhead and operating costs. Additionally, it will aid in ensuring growth isn’t throttled by IT concerns. Quite often, you’ll see a combination of these multiple elements, such as an SQL database deployed to a VM running off of an SQL server, all in the cloud. To make sure all of this works well, Azure has developed its own private network. Azure Networking ensures that companies don’t need to compromise on loading speeds as a result of their cloud reliance, as it aims to deliver low latency to all of its customers. With AT&T (T) moving its own 5G mobile network to Azure and providing Microsoft access to its IP, it’s evident that the company’s infrastructure is pretty top-notch. Bringing down latency and improving capacity are incredibly important aspects of a quality cloud service, which Microsoft now seems to be leading in as well.
Quite often, PaaS can be lumped-in with SaaS, and even with IaaS in some scenarios, which can make it go underappreciated at times. However, PaaS is an incredibly important component of the cloud equation and is really where AWS built its cloud lead. But more on that later. PaaS essentially provides a framework for developers to create cloud or web-based applications, improving scalability and lowering the coding work, and hardware maintenance, that goes into the creation of different products.
One of the most popular PaaS services, offered by most cloud providers, is a Content Delivery Network (“CDN”). CDNs are essentially networks of different servers that aim to reduce data transfer speeds by sourcing information from whatever server is closest to the user. As e-commerce and video streaming become more popular, the need for CDNs has continued to grow in recent years. Azure CDN is Microsoft’s solution to this issue, with Azure Media Services aiming to build upon that with a focus on video streaming.
Another area of strong growth is the Internet of Things (“IoT”). Coined in 1999, IoT is a bit of a buzzword itself. IoT essentially is just a term to cover objects with sensors, processing ability, software, and other “smart” systems that can exchange data over the internet. As you may imagine, the IoT is a fairly popular market segment and cloud computing aims to play a role in managing it. Most often, these devices are actually running their software in the cloud, not on the device itself, and using the cloud to send data to other devices. Azure offers a dedicated IoT platform, Azure IoT Central, where devices and their data can be monitored.
This is also where Microsoft’s innovations in machine learning and AI begin to come back into the fold. Most IoT applications utilize a tremendous network of devices to collect data. Microsoft’s incredible neural networks can rapidly analyze this data, making adjustments to different devices in real-time as needed. Azure’s software is also able to generate visuals that can allow data engineers to easily digest the information recorded by the IoT devices. As IoT continues to grow, from $100 billion in 2017 to around $1.6 trillion in 2025, the revenue available to cloud providers is staggering. Only about 30% of the IoT market value is expected to come from hardware, which is expected to drop to 20% by the end of the decade, with the lion’s share of revenue going to software providers. 25% of a $1.6 trillion market is $1.36 trillion and that insane growth is something that Azure is certainly looking to get a piece of.
Emerging trends, such as Blockchain, also stand to benefit Azure adoption. With Azure Blockchain Workbench, Azure provides a platform for blockchain development and integration. Developers don’t need to worry about creating their own complex parameters or bases to build off of, Azure provides all of that. The Insight Partners project that the market for Blockchain applications will grow at a CAGR of 72.9% through 2028, rising from a $4.935 billion market in 2021 to $227.996 billion in 2028.
The metaverse is another trend that looks to be a major component of the future tech world. While I have my doubts regarding some of the grand plans that many seem to have for the metaverse, I do believe that there is little argument against a future where the digital space is immensely important. I like the work that Microsoft is doing with Teams to establish a workplace metaverse, but that’s not exactly my focus here. Instead, as a PaaS provider, Microsoft stands to be one of the major enablers of such an ecosystem.
Looking back at the IoT, Microsoft has developed a product it calls Azure Digital Twins. As Microsoft explains, “Azure Digital Twins is an Internet of Things (“IoT”) platform that enables you to create a digital representation of real-world things, places, business processes, and people.” The market for digital twins is expected to grow from $3.21 billion in 2020 to $184.5 billion by 2030, representing the greater expected adoption of digital workspaces and environments. With applications such as these, Microsoft is incredibly well-positioned to adapt to the changing technological landscape.
Shifting Market Dynamics
Most people that use the internet are probably aware, at least on some level, of the, somewhat, recent AWS outages. While I don’t anticipate that Amazon will really lose any customers as a result of these mishaps, I do believe that it has exposed a critical weakness in the current cloud-based computing system: failures like these are almost unavoidable. Sure, companies can work their hardest to make them exceedingly rare, and exceedingly rare is pretty good, but it’s not perfect. Amazon isn’t some tech giant of old trying desperately to hang onto its laurels. No, this is one of the modern giants that has invested billions upon billions of dollars to create a reliable, best-in-class, network. Yet, still, it fails.
This is where the emergence of a new cloud solution seems to present itself. As it becomes clear that individual outages will not be eradicated, it seems that the answer is for companies to diversify their contracts. Instead of contracting just one cloud provider, companies will likely shift towards a multicloud approach. Recent IPO HashiCorp (HCP) is building its business on this very idea as its open-source software, among other things, aims to facilitate the creation of such systems.
The idea of a multicloud network isn’t exactly anything new. In fact, around this time last year, about 75% of AWS customers also utilized another cloud provider for secondary workloads. According to a Gartner report, 81% of multicloud customers have found this approach beneficial. This inevitability, the creation of multi-cloud networks, is something that Amazon has been a bit reluctant to embrace. It’s understandable, as the leader of the sector it’s not exactly in your interest to cede ground by encouraging customers to combine two or more services with your own. However, for Microsoft, multicloud provides a great opportunity. This is exactly why the company has created services like Azure Arc, which is specifically designed to support multicloud networks, and Microsoft Sentinel, which is a cloud security system that is able to work across multiple cloud networks.
That being said, Amazon is now slowly starting to adopt the idea of multicloud networks. This begrudging admittance from the industry leader seems to indicate that this inevitable change is now upon us. The future of cloud computing will see customers utilizing a collection of cloud providers, the task now shifting towards becoming the most popular cloud provider within these bundles.
As the market leader, Amazon does still have the advantage on this front, though its competition is certainly starting to catch up. Back in 2018, Bezos said that “AWS had the unusual advantage of a seven-year head start before facing like-minded competition.” Now, over three years later, that advantage has shrunk as its powerful competitors look to make up for lost time with substantial investments.
Google, Oracle, and Microsoft have been essentially buying customers, promising large investments under the condition that the company uses its cloud network. This was part of Oracle’s bid for TikTok and, while this strategy is expensive, aggressive customer acquisition is critical in loosening Amazon’s grip on the market. Perhaps the single largest expenditure in the field was Microsoft’s acquisition of Nuance last year for $19.7 billion. Microsoft’s Cloud for Healthcare, which launched in the fall of 2020, is the clear beneficiary of this acquisition and Microsoft claims that the acquisition has nearly doubled the total addressable market (“TAM”) of Microsoft’s healthcare services to nearly $500 billion. Nuance solutions, as of April 2021, were used in 77% of US hospitals. Nuance’s CEO, Mark Benjamin, commented on the merger saying that it would now allow the company to go more global and increase its scale, indicating Microsoft’s own ambitions for the acquisition.
This move is part of Microsoft’s broader cloud strategy to offer industry-specific applications to its customers. Essentially, the company is approaching each sector as its own unique challenge, developing a set of products that most effectively serve the given sector. The acquisition of Nuance proves that the company is more than willing to dig into its massive war chest to make the investments needed to successfully implement this. By providing a dominant service in multiple niches, such as healthcare, Microsoft will create a universally dominant and capable service. Acquisitions of companies like Nuance is an incredibly effective way of achieving that goal.
Though, it seems that Nuance might not be the last of Microsoft’s major investments in Azure. In January, Bloomberg reported that the tech giant is planning on manufacturing its own chips for Azure, hiring Mike Filippo, a former Apple and ARM semiconductor engineer, to bolster its efforts on that front. It’s unclear exactly what Microsoft aims to do with the proprietary chips it is developing, perhaps aiming to replace Intel’s (INTC) FPGAs, though it will be interesting to see how the company moves forward on this front. Keep in mind, even after Microsoft’s $68.7 billion acquisition of Activision Blizzard (ATVI), the company will still have $61.92 billion in cash (based on its latest earnings report), giving it the power to make necessary investments.
This is all part of a major cultural shift at Microsoft, which was spearheaded by the 2014 appointment of Satya Nadella to CEO. Windows was, and still is, Microsoft’s legacy product. It’s also not as if the operating system is decrepit and past its prime, it still dominates the computing landscape. However, Nadella was cognizant of the fact that Azure was an area where Microsoft could grow. Yet, while he wished to invest in the product, he faced internal resistance as the powerful Windows division vied to retain its corporate power. Oftentimes it would thwart attempts to move top talent to the much smaller Azure division.
Upon becoming CEO, Nadella made a few decisions that Microsoft had avoided in previous years. But his removal of the Windows name from Azure’s product title was perhaps the clearest statement of intent. Azure deserved to stand on its own, moving out of Windows’ shadow. This was even further cemented when Nadella split the former Windows development group in half to focus on future mobility and continuing to lessen the company’s reliance on its legacy product.
Nadella’s promotion to CEO was, in of itself, a powerful statement towards the future direction of the company given his previous role leading the Azure division. His actions since becoming CEO have only solidified this stance and Azure is now the company’s top priority. In this conversation, hosted by the Stanford Business School, Nadella discusses how he had to look back to the company’s origins to decide its future.
“When I [Nadella] joined Microsoft in ’92, we used to talk about ‘Our mission is putting a PC in every home and every desk.’ It was pretty inspiring… except by the late ’90s, we had more or less achieved that – at least in the developed world – and since then, we had this struggle of ‘what’s next?’ And I felt that I needed to sort of go back, in fact, to the very origin of the company. I mean, Microsoft got started when we built, or when Paul [Allen] and Bill [Gates] built, the basic interpreter for the Altair . And I believe that everything that needs to be known about Microsoft in 2019 [time of recording] can be traced back to our origin which is, ‘We build technology so that others can build more technology.'”
While this idea of creating a technological platform once meant bringing PCs into the home of every person in the world, cloud computing presented a new opportunity. Azure, and other cloud computing services, now have the ability for businesses and individuals of all sizes to have access to some of the most advanced computing platforms and algorithms in the world. This shift, which will enable Microsoft to reshape entire industries, brought about through Nadella’s vision, is undoubtedly the future of the company.
To embrace the growth and the risk of investing in cloud computing, Microsoft had to change how it valued performance. Instead of focusing on just profit and revenue, the company began evaluating managers based on customer satisfaction and usage. Changing the previous scope of leadership from transactional to transformational was a major step in accomplishing this. Nadella has credited this decision with enabling the company’s “harsh transition” to a company focused on cloud computing, evolving from a company of “know-it-alls to [a company of] learn-it-alls”.
If it’s not clear already, I am a tremendous fan of Satya Nadella and believe his leadership is amongst the strongest in the world. The growth of Azure remains a priority for the company as it continues to make big strides in the sector. Microsoft isn’t used to coming from behind, or being second-best, but this uncomfortable reality seems to have bred some of the company’s greatest innovations in recent memory.
Market Share Acquisition
AWS remains the top service for the time being, yet this may no longer be due to any actual product advantage. I am confident in saying that Microsoft has an equal, or perhaps even better, product at this time and the company’s leadership has demonstrated a willingness to invest heavily in improving the product’s capabilities. Continued progress, aided by a drive for customers to diversify their cloud contracts, will enable Microsoft to close the gap to Amazon.
If you look at the data, it already is. According to data from Synergy Research, Microsoft is far and away the fastest grower in the cloud computing space. Maintaining this level of growth in market share, while the market itself is exploding, is a rather impressive feat. Even as the market share of other cloud providers, as well as IBM (IBM), seems to be falling rather quickly, Synergy noted that even they have experienced revenue growth. What does seem to be apparent, however, is that the industry is consolidating.
With this consolidation, it appears that Microsoft is capturing most of the freed market space as it edges closer to AWS. According to a report from Flexera, Microsoft is now leading in the adoption rates for enterprise cloud applications. This is the first time Microsoft has surpassed AWS on this metric in the eleven years that Flexera has been analyzing this data, which I believe marks a significant turning point in the overall cloud race. To me, it’s clear that AWS is far from the dominant service it once was.
As detailed above, Azure has grown into an incredibly mature cloud service capable of fulfilling most business applications. Its neural network looks to be one of its greatest differentiators as, when looking at something like storage, customer quality really only comes down to cost. By developing an effective neural network, Microsoft is well-positioned to capitalize on the rapidly expanding IoT network and analysis of massive datasets. As such, I believe Microsoft is best-positioned to capitalize on this next wave of cloud adoption.
Bob Evans, a leading analyst of the cloud market, describes how one of Microsoft’s greatest assets in the cloud space goes beyond its impressive computing capabilities and infrastructure. He explains that “the magic behind Microsoft’s extraordinary success in the cloud has been its ability to help corporate customers modernize not only their ‘digital estates,’ as CEO Satya Nadella calls them, but also their business models, their competitive postures, and their strategic visions.” The company’s ability to turn customers into long-term partners is something that not other cloud operators do and is something that unlocks tremendous long-term value. This, as most things at Microsoft do nowadays, comes from the culture spearheaded by Satya Nadella.
The emergence of greater multicloud utilization will also likely work to increase overall cloud spending, contributing to the market’s growth, as some companies may feel it’s worth it to spend the money to back up their cloud products with multiple providers. While AWS maintains many of its legacy customers, I wouldn’t be surprised to see some of these legacy customers begin diversifying with Azure. This leveling of the playing field is where Microsoft may stand to benefit the most.
Looking at all of these trends and the current state of the cloud market, it’s important to contextualize this from the perspective of Microsoft. Market share acquisition has been outstanding for Azure over the past few years, though this rate of growth is almost assuredly unsustainable. Even if it lasts a few more years, there will come a point when the company’s focus shifts from market share growth, to market share retention.
It seems that Amazon has already reached this cap, its market share staying largely the same, at around a third, for the past few years. So the question then becomes, what’s Microsoft’s cap? I believe that the answer there lies in the fate of the smaller cloud providers. The ‘others’. The market share acquisition from companies like Google and Azure has almost exclusively come at the expense of these smaller companies’ share, which currently make up ~26% of the market. I see this dropping to as low as 6% by the end of the decade.
As SaaS becomes a larger part of the cloud computing equation, smaller cloud providers will simply be unable to keep up with the innovation that behemoths like Amazon, Microsoft, and Google are able to produce. It’s not just SaaS that’s widening the gap, it’s already present with PaaS and IaaS applications, I just expect SaaS to further contribute. To be clear, while these smaller providers will likely see their market share drop, it is likely that their revenues will continue to climb or, at least, remain stable as the market itself expands rapidly.
So, with 19% of the market supposedly up for grabs, I anticipate that Google and Microsoft will be the greatest beneficiaries. Over the past four and a half years, Microsoft has doubled its market share from 11% to 22%, posting the highest growth of all. Currently, Google resides just below the 10% mark. Looking at the above charts, momentum is clearly on Microsoft’s side. Yet, the larger it becomes, the harder it is to maintain such aggressive growth. As such, as it breaks the 25% barrier, I expect growth to begin progressively slowing.
Now, I cannot provide my projections for market share. The horizon is simply too far with too many variables for me to feel comfortable doing so. That being said, I can give some of my baseline expectations. By the end of the decade, I expect Microsoft to surpass 32% market share, giving Amazon a run for its money and perhaps even topping it. My intuition is that Azure will top AWS before the end of the decade, though I recognize that this is a more speculative comment and it should be regarded as such. I also expect Google to surpass 15% of the market by then. Again, speculative assertions. Additionally, I don’t mean to imply that either company can’t hit their respective targets before 2030, just that I would expect this at the very least.
Growth In Cloud Computing
While Microsoft’s growing competitiveness in cloud computing is good to see, the real boon lies with the growth of the cloud computing market itself. About 80% of companies that adopt cloud computing report operation improvements within the first few months of use, signaling quite a surge in cloud usage. In fact, while only around 5% of organizations currently have all of their data infrastructure in cloud networks, 64% expect to operate exclusively in the cloud within five years. Gartner projects that more than 45% of all corporate IT spending will go towards cloud services by 2024, up from 4% in 2017 and 10% in 2021.
This shift has been accelerated by COVID, as “cloud-first” ideology spreads throughout the business world. This acceleration is expected to enable 21% CAGR in global spending on public clouds through 2025, to create an $809 billion market. Though, it is worth noting that estimates have been historically low on the growth of the cloud. Early last year, in April, Gartner revised its 2021 outlook to 23% growth from 2020 to 2021 after projecting 18% growth for the same period in a November 2020 report. Actual growth, according to Synergy Research, was 37%, indicating that cloud growth continues to surpass expectations. Canalys reported slightly slower growth of 34%, still far ahead of projections made by Gartner.
According to a report by Allied Market Research, the cloud computing market is expected to reach a size of $1.62 trillion market by 2030 with a CAGR of 15.8%. All things considered, this may be yet another underestimation of cloud growth. Assuming the report is a fair ballpark estimate, however, it’s worth examining what is expected to be the major drivers behind this growth.
According to Canalys, the shared virtual or augmented reality space, commonly referred to as the metaverse, will be a major driver of cloud growth. This shouldn’t come as much of a surprise, given the immense potential of the space, and Canalys explains “Cloud services are well positioned for individual developers and organizations looking to enter the metaverse. Compute will be in high demand in virtual and augmented reality environments, while storage, machine learning, IoT and data analytics will be essential to support operations such as digital twinning, modeling and interactivity in the metaverse. Continued investment in the metaverse by developers will result in a massive opportunity for cloud service providers, especially the hyperscalers.”
While PaaS is currently the smallest area of cloud computing, it looks to be the fastest grower through the next decade. Microsoft has been heavily investing in building out its PaaS capabilities, making them the strongest on the market. We are currently in the midst of a dramatic restructuring of the global cloud networks and what happens in the next few years will likely determine the next few decades. With Microsoft seemingly best-positioned to handle the next generation of cloud applications, I expect it to capitalize the most on the changing market.
Last quarter, according to TechCrunch, Microsoft Azure generated ~$10 billion in sales. As Microsoft doesn’t disclose Azure revenue, we need to rely on estimates, which often come from market share breakdowns, to assume the division’s revenue. Annualized, that’s $40 million in sales. But that doesn’t exactly take into account the extreme rate that Azure is growing at. After the company’s 46% growth in Azure revenue last quarter, analysts expect growth of 39% next quarter, 37% the following quarter, and 41.9% growth in the quarter after that. Using those estimates with the latest quarter’s results, Azure is operating with annual sales of $69.97 billion. Yet, according to data from FactSet, Microsoft has beaten analyst estimates for Azure’s growth rate every single time they’ve been recorded. At the risk of being a broken record, it seems more than likely that Microsoft will beat these estimates.
During the company’s most recent earnings call, it guided for continued growth in Azure revenue, higher than what was originally expected. At a certain point, it almost seems as if it’s more reasonable to expect Azure to beat estimates than meet them. So, as Azure continues to grow rapidly and more market opportunities present themselves, Microsoft stands to benefit tremendously.
Allied Market Research expects the global cloud market to be worth $1.62 trillion by 2030 which, given the theme with most cloud predictions, is set up to be a conservative estimate. Regardless, assuming Microsoft captures 32% of that market, the company stands to generate $518 billion per year from Azure by then. However, given how dominant I expect Microsoft to be in this next wave of cloud computing, I do feel that my 32% estimate may ultimately be considered conservative. With this, the company will likely continue to exceed analyst expectations, acting as significant catalysts during earnings season.
Given Microsoft’s Intelligent Cloud sector’s operating margin of 42.5%, I would estimate that Azure will be able to generate about $220.15 billion per year in profit off of Azure alone by 2030. At a P/E of 20, this would yield a valuation of $4.403 trillion. Just for the Azure business. This gives a glimpse into exactly what the potential is for Microsoft in the cloud segment. As I discussed in my previous article, Microsoft is a company that is poised for tremendous growth and even its current valuation of $2.12 trillion doesn’t do it justice. I think that this in-depth analysis on Azure more than exemplifies this, putting a number to the company’s greatest growth opportunity.
While Windows is still a top product, it doesn’t provide the opportunity for growth that it once did. As Nadella noted, there is already pretty significant market saturation. This is the crux of the Azure investment thesis. Azure is able to transform the money printer that is Windows, into further growth in a new business that is highly profitable in its own right. Even when looking at something as successful as Windows, companies that aren’t moving forward are moving backwards and, with Azure, Microsoft is moving forwards.
I’ve talked a lot about Satya Nadella in this article, yet I made no mention of the company’s leader in my previous piece. That was a mistake. Someone like Satya Nadella is an incredible asset to their company. At times throughout this piece, I’ve almost used Nadella and culture interchangeably because of how much of the company’s current environment comes from him. That’s an incredibly valuable position to be in. Having a positive work environment is becoming increasingly more important to firms, as employee satisfaction may be a key driver of value growth.
I think this is probably best exemplified by the outperformance of companies that rank highly among Glassdoor’s annual “Best Places To Work” list. Fellow Seeking Alpha author, App Economy Insights, recently put out an article covering this topic and I would highly recommend reading it. While the results may be a bit skewed, due to tech’s general outperformance, I believe it still offers solid evidence to support the idea that strong leadership and culture are catalysts of growth. It’s hard to quantify exactly what the impact can be, though I believe Microsoft is set to benefit immensely from its leadership through the foreseeable future. Down over 15% YTD, I hesitate to say now is the right time to buy given the current period of uncertainty we’re in, but it certainly looks rather attractive. Considering everything I’ve examined regarding Microsoft, both in this article and my last, I’d like to reiterate my prior claim that Microsoft is the ultimate buy and hold.