Cloud Computing: Who Owns the Castles in the Sky? (I)

With PaaS, the stack is a little bit more controlled by the vendor, who usually dictates which database technology and middleware components are used. It's a predefined set of tools that the customer uses to build a software application. An example is Twilio. They've developed their own programming language on top of a set of tools that allows a customer to build a communications application or build communications function into an existing application.

You found that industry estimates for the shift to the cloud in PaaS and IaaS are probably too conservative.

Nelson: One of the more challenging things to predict in any industry is a secular trend or a new theme. I don't think many people predicted that Apple would have such a rapid rise in the smartphone market and that the demise of BlackBerry would be so quick. That's the perfect example of how difficult it is to project exponential growth in certain technologies.

But with the cloud, there's an increasing understanding that running on-premises infrastructure is a wholly inefficient way to conduct business. Within an on-premises data center, you're lucky to extract utilization rates above 10% to 15%. Within a cloud data center, utilization rates are up to 95%, so they can spread those costs over massive customer bases. That's a compelling reason for enterprises to move into a cloud environment. They don't necessarily want to spend less on technology; they want to spend smarter.

Our belief is that over the next five or 10 years, enterprises will migrate legacy technologies that run inefficiently, meaning they're only utilized part of the time or they're built on a very dated set of technologies. With the savings generated from that gain in efficiency, they can invest in more modern technologies like artificial intelligence, machine learning, the "Internet of Things," and edge computing. And all of those services are available in the public cloud. The public cloud vendors stand to gain a substantial share of the IT budget and tap into a portion where they never had a presence before.

How long do you anticipate it will be before the migration is largely complete?

Nelson: It takes a very long time to migrate to the cloud in totality. Netflix was a very early adopter, and it took them seven years to migrate their entire business from an on-premises mode to a public cloud model. That speaks volumes because Netflix is a business that hasn't been around very long and was already pretty modern, while some other enterprises have legacy technologies that date back 30 years. It's reasonable to assume that most enterprises will take between five and 10 years to get to a steady state where they are happy with the amount of assets they have in the public cloud.

Is there anything that might disrupt this process?

Nelson: A few years ago, the most significant concerns were data integrity and security. When you have one customer in one partition and another customer in another partition on the same server, you could have data breaches if that ecosystem isn't managed carefully. Those risks have been mitigated, and the perception has changed as we've seen more companies undergo their own data breaches. Enterprises aren't the best managers of their own IT, and they can't spend at nearly the same rate as a public cloud vendor can on security. Moreover, public cloud vendors are managing a homogenous ecosystem, and they know it inside and out.

Another challenge for multinational organizations is the constraints surrounding data sovereignty internationally. In Europe, every country seemingly has its own set of regulations about how data needs to be handled and where it needs to be stored. It may need to live in a physical data center within the boundaries of that country.

For public cloud vendors that want to handle the workloads of multinational customers, it is crucial to have assets distributed globally as well. It's not enough to just build out thousands of data centers in the United States because if you get a customer that has interests outside of the U.S., you wouldn't be able to serve that customer from a performance perspective and/or a regulatory perspective.

That's why we've established Amazon.com and Microsoft as the big winners in this market today and going forward. They've cleared the hurdle of making really large ramp-ups in capital intensity and capital investment in these businesses, and they have the most distributed network of public cloud services globally. That in and of itself is a huge stepping stone to winning deals. Beyond that, the premium services they already provide around themes like Internet of Things, machine learning, and artificial intelligence deepen their expertise, and ultimately allow them to win more deals. It's sort of a virtuous cycle. They've gotten past that initial capital outlay, and now they can focus on those services that are going to attract customers that want to drive their own businesses forward.

It seems like size and scale are essential to developing and maintaining a competitive edge.

Nelson: Absolutely. By our estimation, this market could be worth hundreds of billions of dollars in very short order in terms of addressable market opportunity. But to address that market, you must outlay tens of billions of dollars to build out that global infrastructure. If you're already behind the eight ball, if you're an Oracle or an International Business Machines (IBM) and you haven't made the requisite investments, you're already very far behind.

The leaders aren't stopping their rate of investment in these businesses. Data-center buildouts are beginning to slow at Amazon and Microsoft, but we're still seeing heavy investment on the R&D side, leading to even deeper product portfolios that are increasingly difficult to replicate. And as more data flows through these services, they become smarter and faster. It's kind of an arms race.

First and foremost, that level of capital intensity and that global scale affords the biggest players a cost advantage, because they can lower prices as they attract more customers and pass on those economies of scale without degrading their own margins. Then, there's an intangible asset component to these businesses as well that comes from having experience managing hyperscale cloud data centers, which we would certainly attribute to Amazon and Microsoft and increasingly to Alphabet and Alibaba.

Then it ties back into that premium services portfolio, offerings around the Internet of Things, artificial intelligence, machine learning--the things that are driving incremental investments at enterprises. Those two things are really the building blocks for a competitive advantage and ultimately an economic moat in this market.

Longer term, as the cloud market matures, there's going to be some level of switching costs for customers. It's going to be very difficult to replicate an ecosystem at a different vendor. Amazon and Microsoft each have their own areas of expertise. For Microsoft, it's a deep and inherent knowledge of Windows-based applications, of which there are millions out there. That creates a natural landing place for a lot of those workloads in Microsoft Azure. Amazon has the deepest and broadest product portfolio of any of the cloud vendors.

The information contained within is for educational and informational purposes ONLY. It is not intended nor should it be considered an invitation or inducement to buy or sell a security or securities noted within nor should it be viewed as a communication intended to persuade or incite you to buy or sell security or securities noted within. Any commentary provided is the opinion of the author and should not be considered a personalised recommendation.

©2017晨星有限公司。版權所有。晨星提供的資料：(1)為晨星及(或)其內容供應商的獨有資產；(2)未經許可不得複製或轉載；(3)純屬研究性質而非任何投資建議；及(4)晨星未就所載資料的完整性、準確性及即時性作出任何保證。晨星及其內容供應商對於因使用相關資料而作出的交易決定均不承擔任何責任。過往績效紀錄不能保證未來投資結果。本報告僅供參考之用，並不涉及協助推廣銷售任何投資產品。