We thank the readers for their appreciation ofthe first articlein this series. Additionally, we would also want to highlight a few pointers based on the reader questions and commentary:
- This analysis is limited to the listed players in the public cloud space and does not extend to application companies. For example,Salesforce.com(NYSE:CRM) has apreferred partnership statuswith Amazon (NASDAQ:AMZN) Web Services (AWS) and Adobe (ADBE) runs on Microsoft (NASDAQ:MSFT) Azure. While Salesforce’s Sales Cloud and Adobe’s Experience Cloud are leading products, they are applications and the underlying cloud belongs to public cloud companies. Thus, for the purpose of this analysis, none of the application companies has been included.
- There has been some reader interest in cloud ancillary hardware companies. While some of them represent remarkable investment opportunities, some have been constrained due to a variety of factors including cloud vendors going slow on spending. Furthermore, these ‘downstream’ companies of the cloud value chain (much like application companies are the ‘upstream’) warrants a separate discussion.
- The more prominent foreign names in the public cloud market are those in China and possibly SAP (NYSE:SAP). With the Chinese companies, take for example Alibaba (NYSE:BABA). The company primarily caters to a Chinese audience. The size of Alibaba’s cloud revenue was ~$2 billion last year and for the 9 months of this fiscal, it had done ~ $2.5 billion. In addition, the company’s complex holding structure (read variable interest entities etc), the role of Jack Ma (having left the pole position, but still a key figure in the scheme of things and tending towards philanthropy), the inherent risk from the current tariff war and a need to understand the state of the Chinese economy lend it unattractive for us to look at Alibaba Cloud and other Chinese players at this point in time. Regarding SAP, the company is in an aggressive M&A mode to shore up its cloud business. Given the size relative to the players mentioned here and the risk in SAP’s inorganic strategy, we would want to see the company execute in this market for a few more quarters before digging deeper.
- Hybrid cloud management has become an important category in the cloud market, due to a mix of on-prem and cloud assets that enterprises have built up over the years. Given that a large set of enterprises still need to fully embrace the concept of cloud, it may be slightly early in the day to segregate hybrid by category.
- Interestingly only one (Microsoft) of the ‘old guard’ falls figures in the top 3 in the cloud, while the other two (Amazon and Google) as relatively new entrants in a market that is the evolution of traditional storage and compute.
We start off by taking a step back into understanding what cloud really is and the implications arising from the growing need to adopt a cloud-centric approach.
Cloud refers to the outsourcing of storage and/or computing functions by renting capacity from cloud vendors. Cloud vendors generally have a distributed presence and storage or computing is delivered over the Internet to cloud customers.
Cloud was initially considered as an inexpensive means to test the market for new products, without worrying about constraints of capacity or cost. Post the test phase, customers get better visibility of requirements and could easily move from a public cloud to a private cloud (in a public cloud the entire management of hosting is outsourced, which changes in case of a private cloud where the customer owns the responsibility of managing and hosting). Since on a longer-term basis, renting is always more expensive than owning an asset, private clouds make more sense for larger, stable businesses looking to expand margins from cost efficiencies rather than primarily relying on revenue growth.
Over the last decade, IT departments’ reluctance to adopt cloud has given way to extreme enthusiasm towards the cloud. While the initial reluctance was due to expected complexity in managing another element in the IT portfolio, the enthusiasm has been driven by the expectations of enabling innovation while drastically reducing cost. Over the last few quarters, both reasons have converged to give rise to the multi-cloud paradigm: the need to manage disparate cloud assets (acquired or in house), along with the ensuing complexity in managing them has led to the rise of hybrid cloud management solutions or system integration services for multiple clouds (private and/or public).
Another trend that has dominated the need for cloud is that of edge computing, driven by the internet of things (IoT) adoption. Edge computing refers to the processing of data closer to the edge (or at the nodes of the network, where the data is generated) as opposed to bringing the data to the cloud to process. Use cases of edge computing include real-time decision making (gaming, autonomous driving vehicles, algorithmic trading etc), where even a minuscule delay in processing (arising from sending data to the cloud and receiving processed information back) can have severe consequences.
The expected risein the number of connected devices (essentially IoT devices) and the consequent need to manage a network of on-premise and cloud assets give further impetus to hybrid cloud management.
The VMware (NYSE:VMW) management has been credited over the last many quarters for their prescience about the growth of hybrid cloud and CEO, Patrick Gelsinger notes:
And the other thing I’ll observe is that in the history of computing, now this is my 39th year in technology, hard to say that out loud but we’ve seen the swing of technology forces from centralization to decentralization over the history of the industry and cloud is going to force the centralization, edge and IoT will be a force of decentralization as we go forward and we don’t see that this world either swings one way or the other ever.
Source:Barclays TMT Conference(Dec 2018)
What corroborates his views is the slowdown setting in cloud revenues over the last three to four quarters for most of the major players. While the reasons vary, the underlying theme is that of investment having been cut back and now growth rates beginning to moderate.
Source: Company filings and author estimates.Note:LQ refers to the latest reported quarter; LQ-1 to one quarter before and LQ-4 to four quarters prior
Interestingly, Oracle is only the vendor to have seen a steady increase in capex.
Source: Company filings and author estimates
It appears that the broader market has been slowing down at least over the last couple of quarters, possibly due vendors focusing incrementally on a hybrid cloud to grab the edge opportunity. In parallel, the cloud vendors may also be preparing for an expansion of the cloud to be a more standardized abstraction of owned and rented resources. Some vendors have invested in building assets, some players have opted for collaborations and some have embarked upon the more ambitious route of M&A.
Azure has been firing all cylinders on the back of Microsoft’s wide enterprise reach and the company’s own end-user applications sitting on top of Microsoft’s cloud. In addition to Office 365, gaming has been gaining importance for the company. Microsoft did over $10 billion in gaming revenue last year, which is ~10% of the company. And ~7% of the$150 billionper year gaming market.
In March this year, Microsoft bundled all of itsgame development toolsunder the ‘Game Stack’ umbrella and will be hosted on Azure. In addition to Microsoft attempting to provide a complete toolkit for game developers, the company is also trying to address the most complex part of cloud gaming: latency issues. In early April this year, Microsoft alsoannounced a partnershipwith Akamai (AKAM), a leading content delivery network (CDN) provider.
Akamai is possibly one of the largest CDNs (or distributed networks) in the world and has been working with streaming companies for long, facilitating lower latencies in gaming and live sports streaming.
As recently as the end of April, the Akamai management commented that while managing latency is possible, the problem is that of economics at scale – a growing volume of traffic makes it a rather expensive proposition for cloud game streaming to be profitable. And yet, Microsoft has tied up with Akamai.
Microsoft had introduced subscription for its gaming products a while back while acknowledging a shift from a ‘per device to a per user’ model enabling users to consume gaming across a spectrum of devices (from consoles to mobiles). The management had also expressed willingness to invest in its gaming product to be able to capture a larger share of the $150 billion a year gaming market. In this context, the collaboration with Akamai appears to be an investment (even if the economics currently do not make sense) to further Microsoft’s and Azure’s edge capabilities.
In addition to Akamai’s edge network, Microsoft now also has formal relationships with VMware and Red Hat (NYSE:RHT) – IBM (NYSE:IBM) for hybrid capabilities. The cherry on the cake for Microsoft’s cloud stack offering is the fact the current CEO, Satya Nadella, was the executive vice president of Microsoft’s Cloud and Enterprise group before his elevation to the post of CEO.
Satya has been particularly vocal about Azure’s ‘architectural advantage’ as the key to its success. After a slight bit of wobble, a couple of quarters back, due to product mix shift, the cloud business appears to be in prime form with revenue growth and margins expansion indicating share gains.
Larry Ellison has been overseeing the company’s cloud business after the exit of Thomas Kurian, over reported differences around Oracle’s strategy of selling its own cloud versus making more of its software available over other public clouds. The importance of this executive departure can be gauged from the following:
- he was lapped up by no other than Google (NASDAQ:GOOG) to lead Google’s cloud business
- Thomas worked for over two decades at Oracle, rising through the ranks and reporting to Larry
- Oracle had started marketing its Gen 2 cloud over the last couple of years
From a market standpoint, Oracle claims that Gen 2 is a better architecture since it separates the cloud control from user code, making it a safer and technologically superior bet.
Source: 2018Oracle Analyst Day
While Oracle’s leading share in the database market has been its claim to fame, the company’s aggressive sales tactics have often been a deterrent for the adoption of its products. Also, Oracle’s cloud is fairly new in comparison to those of the existing players. Despite these limitations, the company has been trying to encourage its customers to migrate to their databases to the Gen 2 cloud.
But in terms of technology, there is no way that someone can move — a normal person would move from an Oracle database to an Amazon database. It’s just incredibly expensive and complicated and you’ve got to be willing to give up tons of reliability, tons of security, tons of performance to go ahead and do it. But we have a huge technology advantage. Again, don’t believe me read the Gartner report. We’ve never had — the Oracle autonomous database has the biggest technology lead we have ever had in the database world from a technology standpoint. The problem is we have to deliver that autonomous database on first-class cloud infrastructure to be successful in the cloud business. We need more than just a great database. We have the best database but we also need first-class infrastructure to run that database on. And we now finally have that with our generation two cloud, and I think you’ll see the combination of the Oracle Autonomous Database in the generation two cloud.
Source: Q2 2019Oracle Earnings Call
Notably, AWS still uses Oracle’s database products. Oracle also appears to have generated quite some buzz with its Autonomous Database.
The introduction of our gen 2 highly secure infrastructure featuring the Oracle Autonomous Database has been very well received. During Q3, we had nearly 1,000. Over 4,000 active trials.
20% of our customers on autonomous Data Warehouse, Autonomous Database right now are net new to Oracle. We did not have them before.
Source: Q3 2019Oracle Earnings Call
However, the decline in the company’s legacy business has not allowed the growth in cloud business to sufficiently reflect in the top line. Although, margins have improved pointing towards a larger software component in the revenue mix.
Source: Company filings and author estimates
Also notable are theslew of acquisitionsthat the company undertook in 2018 to augment its cloud feature set, especially on the AI side.
The intent of productizing more features in a cloud that can readily take care of an on-premise database appears to be the right move. The only hiccup is that the guy leading these acquisitions is not with the company anymore. Thomas Kurian’s departure also brings into question if 2019 can still be the year of inflection for Oracle’s cloud ambitions. There are two ways to think about this:
- When Larry Ellison was referring to 2019 as the year of inflection, he still was opposed to his cloud chief’s idea of making Oracle software available in other public clouds. Thus, with the departure of Thomas, whatever limited resistance the Chairman was facing is gone and now the month of the inflection should get pulled forward. Or,
- Oracle expected the skirmish to subside with both sides reaching a middle ground, ultimately benefitting of the company.
The thing that goes against Larry is the market’s long wait for Oracle to see growth, driven by the cloud business. However, the company has been repurchasing a lot of their own shares. Also, not to forget Larry Ellison has backed visionariesincluding Steve Jobsin the past.
IBM is a bit of a unique case, given the portfolio churn the company has been witnessing. IBM appears to have straddled between an incumbent and challenger, finally settling in as a major player in the cloud integration (hybrid cloud space) still wanting to sell some of its own cloud. IBM shocked the markets last October with the announcement of its Red Hat deal. The surprise was more to do with the size of the deal that IBM was undertaking and for the asset that the company was purchasing.
The Cognitive Solutions business was IBM’s beacon of light; with the Red Hat buy, IBM gained more prominence in the hybrid cloud market. Recognizing the importance of keeping technology services business separate from its core cloud business, IBM had moved its core cloud business from within technology services and bundled it with the ‘Cognitive Solutions’ business. The combination of Cognitive and cloud is likely to help IBM productize much of the company’s technological prowess into its own cloud. Factors that favor IBM’s position in the evolving cloud market are:
- The company’s long history of enterprise relationships (something extremely critical to scale up the cloud and the associated SI work), especially as the IBM helps its client wade the challenge beyond that of the ‘low hanging fruit’ of cloud migrations.
- IBM’s commitment towards the cloud (the Red Hat acquisition is by far one of the largest deals in tech) appears to be quite serious and looks in line with that of a company looking to re-invent itself.
Some aspects that are also likely to contribute to IBM’s position in the overall cloud order are:
- The Red Hat acquisition is likely to upend IBM’s business model from that of a large, legacy, enterprise software, hardware and services vendor to that of a leaner cloud and legacy solutions provider. The changes in the business model are likely to also necessitate management level changes to keep the momentum intact. As of the time ofwriting this article, Arvind Krishna, a three-decade IBM veteran, heads Cloud and Cognitive Software. Taking a cue from Microsoft’s history, IBM could significantly strengthen its position in the cloud/hybrid space, should the head of the cloud be also leading the broader portfolio of the company’s assets to better align them with the overall company philosophy. Ginni Rometty’s absence from the last couple of earnings calls may just be the first hint of an imminent change. However, clarity about management will be paramount for IBM to be able to convince customers of the company’s technology vision to be that of a marathon rather than of a sprint.
- With the Red Hat acquisition, IBM has been targeting the $1 trillion hybrid cloud market and it could make sense to not compete with the big public cloud players. However, as the world becomes increasingly multi-cloud, customers have been trying to avoid vendor lock-ins. IBM’s cloud is an excellent alternative for customers wanting to increase the diversification of their underlying base of assets while trusting IBM with the abstraction to be provided (hybrid management), to ensure standardization and smooth operations. However, to adopt this strategy, the management may need to enhance its messaging and steer clear of coming across as a competitor aiming to become one of the top three in the public cloud provider market.
Since the first half of the last year, the IBM management had been trying to build a narrative around the potential of its hybrid cloud strategy. Subsequently, while the core cloud business may have slowed a bit, the SI bookings have remained fairly strong. Expected closure of the Red Hat deal in the second half of this year and the IBM’s messaging around its cloud strategy could potentially make IBM stand out in the hybrid cloud world.
If there is a sweet spot for hybrid cloud management, VMware is in it. The company is a known leader in virtual machine software and has beenincreasingly bundling softwaretowards managing private clouds.
A few quarters back, VMware had started seeing traction in its partnership with AWS.
We continue to see customer interest and engagement with VMware Cloud on AWS. The service designed to bring VMware software-defined data center to the AWS Cloud, allows customers to run applications across operationally-consistent VMware vSphere-based private, public and hybrid cloud environments, with optimized access to AWS services.
Source: Q4 2018VMware Earnings Call
In return for allowing AWS to access VMware’s customer base, VMware’s was getting a margin advantage by not having to host.
We also see that, given it’s a subscription business, that will also delay the direct fiscal impact. As we said before, it’s a net business for us. So we’re not carrying the cost of hosting through our P&Ls. So that gives it a pure subscription license business to us.
Source: Q4 2018VMware Earnings Call
Over the course of the last year, this focus on AWS and potentially IBM’s closeness with Red Hat, resulted in AWS displacing IBM’s central role in VMware’s partner program (VCPP). VMware was aiming to progress beyond selling AWS into its own customer base and beyond to growt this business further.
We are largely selling into existing installed base VMware customers as the starting point, but we clearly expect as we go to Phase 2 and beyond
Source: Q2 2019VMware Earnings Call
And the early beginnings of Phase 2 appear to be thelaunch of AWSOutposts.
In addition to the AWS partnership opening up the possibility of healthy margin growth for VMware, VMware’s dominant position in the hypervisor space is also likely to dictate similar partnerships for other players to be able to survive and grow. In part, VMware and AWS’ relationship may have inspired the IBM- Red Hat’s deal.
Where VMware also scores is with its relationships with practically all of the large players. In a world, where multi-cloud environments are becoming the norm, VMware’s partner network of over 4,000 cloud partners does give VMware the pole position. Also, worth noting it the benefits that these relationships can potentially bring for Dell, VMware’s parent, and quitean influential one.
While there is no doubt that VMware’s role in the hybrid cloud market is likely to shape up the market to a large extent, the company’s closeness with AWS can cause some level of friction in its relationships with other cloud majors.
AWS may not have the best technology stack, but Amazon’s customer centricity has allowed for AWS to not just lead the public cloud market but also finance Amazon’s consumer business. Interestingly, the Amazon consumer business is known to be one AWS’ largest customers.
AWS’s VMware partnership is potentially an investment for Amazon with the potential for a much higher ROI. AWS does not need to educate the customers about the benefits of the cloud anymore. Instead, the company needs to manage costs to be able to grow margins while cutting costs (something that AWS has been doing with an almost clockwork-like efficiency). AWS’ announcements around deep storage, concurrency etc point that the company has reached a point where the marginal cost of incremental data processing and storage has become minimal. Thus, the fall in Amazon’s capex is well aligned with the company trying to gain entry in the private cloud space with VMware: AWS needs to host and the company has been bringing the cost of hosting down. Given that VMware claims to have hundreds of thousands of customers if AWS continues to plays nice, there is scope for both to continue milking the private and hybrid cloud markets.
While AWS’ continued investment in research and partnerships are likely to help sustain the company’s competitive moat in the cloud business, the risks are more from within.
- Amazon’s entry in many related fields has caused many of AWS’ customers to view AWS as a part of a competitor. Anecdotal evidence suggests that Microsoft may have benefited in some cases. Should Amazon continue to expand in newer business lines, competitors could benefit from displacement due to competitive threats of Amazon having access to customer data.
- AWS has become quite large and as the hybrid cloud market continues to shape up, enterprises are looking to have alternative providers in their portfolio to hedge their risks.
- While AWS’ focus on a partnership-driven model has allowed for capex to remain low, over the last few quarters, growth rates have moderated. Also, the reduction in prices has been impacting margins. More recently, the management noted sequential weakness due to lumpiness in business. Either AWS is taking on massive (by AWS standards) transformation projects that are causing lumpiness or the benefits of front-loading of capex investments in 2016-17 are beginning to wane. In either case, AWS would not want to cede much market share to either competitor given that cloud is ultimately a business of scale.
Google Cloud Platform (GCP)
Apart from the regular ramblings of technological superiority, relative lack of customer focus in its enterprise business versus other cloud players and disappointment in management over its opacity around the size of its cloud business, GCP also needs to be seen as a progressive business.
- GCP has a come a long way from selling its product to customers with the tagline ‘running like Google’ to making a concerted effort at making the Google Cloud consumable for its enterprise users. What is laudable about Google’s efforts is the fact that GCP was built for developers to develop software on it and not for enterprises to migrate legacy workloads and still the company has been winning large enterprise customers. Furthermore, GCP has been launching newer products to meet where the customer is. For example, the company launched an identity management tool, something Googlers may consider boring, but is a necessary requirement for conventional enterprises.
- Google’s hiring of Thomas Kurian was no less an event than IBM’s Red Hat deal.
- Apart from Thomas Kurian’s closeness to Larry Ellison at Oracle, Thomas is also an outsider for Google (not only from a company perspective but also culturally). Satya Nadella of Microsoft, Arvind Krishna of IBM, Patrick Gelsinger of VMware and Andy Jassy of AWS (and even Sundar Pichai of Google) are all company veterans that rose through the ranks. So did Thomas Kurian, at Oracle. Another glaring issue for the enterprise product chief will be how well can he break the stereotype of ‘homegrown CEOs doing well’ in one of the world’s most engineering-focused companies in one of the company’s fastest growing business.
- Much of the criticism faced by Google for not providing enterprise-grade support to its cloud customer stemmed from the fact that Google had repurposed its own cloud to sell outwards. Thomas Kurian was the head of the cloud at Oracle, whose Gen 2 cloud has been re-architected to go beyond the traditional cloud approach. Thus, one of the questions facing the new CEO is likely to be: While he has experience at seeing the Gen 2 cloud evolve, can he add enough bells and whistles to Google’s existing cloud product to make it readily amenable to legacy / conventional enterprise needs?
- The change of GCP’s operating focus from coders to enterprises can also pose a potential challenge for the company. AWS’ success was due to the company’s cloud was custom built to help enterprises and AWS spending significant effort in educating the audience about the benefits of the cloud versus customers’ on-premise solutions. GCP might just be behind the curve. In the early days of the cloud, enterprises wanted a hands-off approach to migration. Over the years, as people have understood cloud better, not only coder-driven startups (that were driven to the GCP due to its coder-friendly offering) but also enterprise IT managers have started valuing GCP. However, selling the product to the CIO versus the CEO are two different aspects and Thomas Kurian will be walking a tight rope balancing these two, often conflicting, requirements.
- The launch of Stadia, a cloud gaming service, is another move by Google to fuse the boundaries of its enterprise and consumer business. What augurs well for Stadia is that it has a significant integration with YouTube, where over 50 billion hours of gaming content was watched in 2018. There are potentially two challenges that Stadia could face:
- The first one pertains to the technical barriers arising from the streaming of high definition graphics with requirements of low latency, and that too at scale. As compared to movie or video streaming, where local devices can store the data, for gaming latency becomes much more important since players are waiting to react to the opponent’s next move.
- The second challenge is more subtle but likely to be equally important. Stadia is also going to be a test of how well Google has learnt from the mistakes that were made while selling its cloud. Stadia is based on Google’s cloud and CDN, which in a way is the repeat of how the company first started selling its cloud business – an internal resource that was opened up for mass consumption. Of course, people dealing with Stadia are much closer to coding than the IT executives early on at the start of public cloud sales were. While the customers of Stadia are looking forlatency and economics, something that hasbeen structurally challengingfor cloud gaming, Google’s addressal of these aspects will be a key thing to be watched out for.
The cloud market has been marked by a myriad of changes across companies and executives, aimed at growing share. Hybrid and edge appear to be the new chase worthy goals, with investments ranging from acquisitions to partnerships. However, the clamor to get a slice of the existing on-prem installed base is inevitably clear.
As we progress towards the third part of this series, Masters of Cloud: A Dark Horse,we invite readers to help us distill through the companies discussed. Building on our research and user commentary, we aim to identify stock/s that can provide the best return with the lowest risk.
Disclosure:I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.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.