Cloud "Vendor Revenue" Projections 2015-2026

Wikibon Analysts

• David Floyer
• Peter Burris
• Ralph Finos

Premises

Three premises are tested in this research:

1. Individual cloud vendors will find it very difficult to dominate the total enterprise cloud market;
2. Vendors with reach in multiple segments of the enterprise cloud market, and have lower migration costs to the cloud will prosper best;
3. Using a Cloud “Vendor Revenue” Model allows a better understanding of the cloud market, and an improved ability to project future growth of the cloud markets and vendors.

Executive Summary

Total Enterprise Cloud Revenue Market

The traditional total enterprise cloud market is defined by Wikibon as the sum of:

• Software-as-a-Service (SaaS)
• Infrastructure-as-a-Service (IaaS)
• Platform-as-a-Service (PaaS)
• True Private Cloud (TPC)
• Business Processing Outsourcing (BPO)

Within this research, we have used both the traditional definition of each segment, and a modified definition of each segment to more closely reflect the “Total Vendor Revenue”. For example, SaaS often runs on an IaaS base; traditionally all the SaaS revenue from enterprise users would accrue to SaaS, and none to IaaS. The Wikibon “Vendor Revenue” model reallocates some of the SaaS revenue to IaaS and other reallocations. The full details are covered below. However, the total cloud market sizing, the sum of the five segments above, is the same in either view. Wikibon has reallocated revenue within the Vendor Revenue model, to help the research on the initial premises above.

Figure 1 summarizes the Wikibon research findings for the worldwide enterprise IT cloud projection using the Vendor Revenue segmentation. The overall enterprise cloud growth projection CAGR 2016-2026 is 19%. The growth from 2023-2026 is projected to be a CAGR of 8%. The details are discussed in a later section.

Enterprise cloud revenues are part of a total Enterprise IT spend, which is projected to grow in line with GDP at about CAGR 2%. Figure 2 summarizes the Wikibon cloud projections. Overall, the enterprise IT market is moving from individual traditional stovepipes (Servers, Storage, Networking, Software) to far more integrated offerings, both as public cloud services and private cloud services. This is enabling the cost of enterprise operations to reduce significantly over the forecast period at a CAGR of -7% from 2016-2026. The traditional hardware, application software and outsourcing markets are projected to reduce at a CAGR of -3% from 2016-2026.

The basis for the research are the Wikibon cloud projections published in 2016, with additional research and analysis. The publications were:

• Wikibon 2016 Public Cloud Forecast Update, which projected the 10-year growth of IaaS (Infrastructure-as-a-Service), SaaS (Software-as-a-Service) and PaaS (Platform-as-a-Service).
• True Private Cloud Projections 2015-2026, which projected the 10-year growth of TPC (True Private Cloud). TPC is converged infrastructure with automation and orchestration capabilities similar in productivity benefits as the public cloud, and with similar reductions in on-premise staff. An example of TPC is the Oracle Cloud Machine, which provides centrally managed and coordinated updates to a complete database stack, is managed by Oracle, and runs the same software as the Oracle Cloud Services.

Classic Enterprise Cloud Projections 2015-2026

The Wikibon 2016 “Classic” enterprise cloud projection, is shown in Figure 3 below. The segments combine the previous Wikibon definitions and projections for IaaS, SaaS, PaaS and TPC. In addition, Wikibon has added another segment to this cloud segmentation research, Business Process Outsourcing.

The total projected size of the enterprise cloud marketplace is projected to be $523 billion in 2022, and $743 billion by 2026. These projections use the classic definitions of IT segments, namely that the revenue is measured by revenue from the end user.

Table 1 below shows the CAGRs for each of the classic enterprise cloud segments in Figure 3. The growth of total enterprise cloud was a CAGR of 49% from 2015-2016, 26% 2016-2022, 19% CAGR 2016-2026, and 8% 2023-2026. All the segments are growing significantly in revenue. IaaS and SaaS are the largest segments in 2016, and are expected to crest by 2026. BPO and TPC are much smaller segments with faster projections over all the time periods in Table 1. PaaS grew more slowly than the other cloud segments in 2015-2016 (22%), but is expected to grow faster later this decade.

However, in looking at the cloud market as a whole, there is significant revenue that is generated for IaaS vendors from running SaaS software for SaaS vendors, and the same with BPO. In addition, much of the revenue from Paas is projected to be realized as part of IaaS or as part of SaaS.

To help look within the whole cloud market in a different way, Wikibon has developed a different and more useful segmentation of the whole cloud market by “vendor revenue”. This is defined and projected in the next segment of this research.

“Classic” to “Vendor Revenue” Enterprise Cloud Adjustments

This segment of the research details the way that the enterprise cloud Vendor Revenues are derived from the Wikibon “Classic” projections. This section, and the footnotes section at the end or this research posting, are optional for most readers, and can be returned if required later.

The Wikibon adjustments to the Classic Projections in the previous section involve how:

• All the revenue from PaaS is fully reallocated to all the other segments;
• Some of the revenue from SaaS for infrastructure services is reallocated;
• Some of the other revenues segments are allocated to True Private Cloud (TPC);
• Some of revenue from all the other enterprise cloud segments outsourcing is reallocated to Business Process Outsourcing (BPO).

Revenue from PaaS

PaaS continues to be small as a separate cloud segment.  PaaS as a percentage of the total cloud market is expected to be 2% in 2016, 6% in 2022, and 9% in in 2026. This growth is expected to trail the rest of the cloud market.

The key benefit of PaaS is to provide an integrated end-to-end development, deployment, operational management and maintenance/update platform. PaaS can make the development cycle much more efficient, especially in greenfield sites. By employing microservices, PaaS enables enterprises to realize business value much more rapidly: changes can be made more quickly and testing can be simplified. The trade-off to efficiency and rapid deployment, of course, is that for specialized requirements, this sort of generalized solution may not be appropriate. For instance, for some applications, maximizing absolute response time speed may outweigh the value of efficient delivery. In addition, the cost of migrating existing applications to a PaaS platform are high, especially for mission critical applications.

Wikibon’s analysis on PaaS shows that there are three specific sub-segments of PaaS software.

1. IaaS is including PaaS functions as part of IaaS. For example, the announcement by AWS of hundreds of functions (e.g., Lambda, the Aurora database, Chef, Puppet and many others) is making the distinction between IaaS and PaaS difficult to maintain. Similar functionality is included in Microsoft’s Azure, and VMware. Wikibon expects this trend to continue to become an increasing percentage of the PaaS. The percentage of PaaS running on traditional platforms is small.
2. Similarly, SaaS vendors such as ServiceNow and Oracle Fusion have introduced PaaS functionality on top of SaaS, again making it difficult to differentiate between SaaS and PaaS. The introduction of a common database and data dictionary across both ISV and enterprise written applications is an important element of the PaaS value proposition.
3. True Private Cloud (TPC) offerings will include PaaS functionality in response to IaaS and SaaS inclusion as part of the drive to make TPC as productive as possible to the data centers.

The bottom line: this analysis leads Wikibon to believe it is unlikely that is unlikely that standalone PaaS from a Development Environment provider (e.g., IBM’s Bluemix, HP’s Helion Development Platform, Pivotal Software’s Cloud Foundry) will constitute a significant portion of the PaaS market.

As a result of these three different PaaS segments, it become  increasing difficult to differentiate between the revenue contributions of PaaS and IaaS, PaaS and TPC, and PaaS and SaaS. Wikibon believes that for long term projections of vendor revenue it is easier to merge PaaS with IaaS, TPC and SaaS. The SaaS market is bigger market, but Wikibon believes there will be as much PaaS activity within IaaS & TPC. The allocation used by Wikibon is to add 50% into the SaaS Projections, and 50% of the PaaS projections into IaaS and TPC, allocated according to the relative market size of IaaS and TPC. These proportions may be adjusted in the future as a result of additional research.

The details of the mapping from the enterprise cloud “Classic” revenue segmentation to the enterprise cloud “Vendor Revenue” segmentation is given in the Footnotes section at the end of this research posting. This section includes the assumptions and calculations used.

Resulting Revenue Adjustments from “Classic” to “Vendor Revenue”

The percentage reallocation of revenues from the Enterprise Cloud Classic segmentation to a vendor revenue segmentation is shown in Table 2 below. A description, the underlying assumptions and detailed calculations are given in the Footnotes section at the end of this research posting. IaaS and True Private Cloud  (TPC) are the cloud segments with the biggest impact from the revenue reallocation. The impact is projected to be 6% increase for both segments in 2016, and 19% in 2026.

Enterprise Cloud “Vendor Revenue” Projections 2015-2026

Figure 4 is a reallocation of the Total Enterprise Cloud market, using a “Vendor Revenue Segmentation”. This is achieved by applying the percentage in Table 2 to Figure 3. This allows a more accurate representation of the vendor revenue potential for participants in the enterprise cloud marketplace segments. The total cloud market in Figure 4 remains the same as in Figure 3.

Table 3 below gives the CAGR figures for the Vendor Revenue Enterprise Cloud segments in Figure 4.

Table 3 shows the compound annual growth rates (CAGR) for 2015-2016, 2016-2022, 2016-2026, and 2023-2026. Overall, the total enterprise cloud grew by 49% from 2015-2016, and is expected to grow at 26% CAGR between 2016-2022. The overall shape of the total enterprise curve in Figure 2 and Figure 3 is the typical ogive curve that describes most disruptive adoption scenarios. The CAGR for total enterprise cloud is 49% for 2015-2016, 26% for 2016-2022, 19% for 2016-2026 and 8% for 2023-2026.

As would be expected with disruptive technologies, all the percentages are highest in the first column, and decrease thereafter as the size of the market segment grows, and has friction for growth increases over time. As can be seen in Figure 3, the BPO Cloud and True Private Cloud segments are in an earlier growth phase, with much smaller revenues, but with higher growth rates than the other cloud segments. The growth characteristics of the established segments of IaaS and SaaS is about the same, and are projected to crest at the end of the projection timeline.

It is important to note that in choosing a ten-year timescale in the Figures and Tables above, Wikibon is illustrating the overall market dynamics and adoption curves of disruptive technologies, and painting a likely adoption scenario. The shape of the curves, and the discussion of market forces they illustrate is more important than the exact projected numbers. In order to understand the overall market forces, it is necessary to study the growth of cloud computing in context with the overall Enterprise Projections. This is the topic of  the next section of this research.

IaaS, SaaS & True Private Cloud Revenue Projections 2015-2026

Figure 5 shows the Wikibon projection of IaaS, using the Vendor Revenue allocation within the Total Cloud revenue. The red section of Figure 5 shows the “Classic” forecast, and the blue part above is the reallocation from other classic segments.

Figure 6 shows the same vendor revenue data as Figure 5 for SaaS. The additional income from PaaS has been almost exactly offset by the loss of revenue to IaaS and True Private Cloud. Wikibon interviews with SaaS vendors indicate that SaaS companies are happy to start with AWS, but believe that they can save money by using their own systems, and avoid the business risk that IaaS vendors could learn too much about their systems and revenues. The Wikibon SaaS growth starts from a higher revenue, and is longer lasting than the IaaS growth in Figure 5.

Figure 7 shows the same vendor revenue data shown in Figure 4 and Figure 5  for True Private Cloud. There is additional income from PaaS, and additional SaaS and BPO.  The Wikibon TPC growth starts from a much lower revenue base, but has a significantly higher growth curve, and is still growing rapidly in 2016. Wikibon believes that Edge IoT systems that will be latency constrained will start to add significant growth to True Private Cloud during the next decade, as Industrial IoT, transport IoT, Energy IoT and Medical IoT start to ramp up.

The conclusions are that both IaaS and TPC will grow very fast over the next decade, but that the total available market for TPC is significantly bigger that IaaS.

Overall Enterprise Projections 2015-2026

Figure 8 above takes the same cloud data as in Figure 2, but puts it into context against spend on traditional IT Infrastructure, and operational staffing costs (purple segment in Figure 2). The purple segment illustrates an important driver of cloud computing, the reduction in enterprise operational staff costs that result from the movement of cost from the enterprise to vendor R&D. The traditional segments illustrates that although cloud computing overall is projected to account for 46% of enterprise spend by 2026, there is still significant spend on traditional infrastructure hardware and software, as well as enterprise application software.

Table 4 on the left is derived from Figure 8, and shows the percentage of Total Enterprise Vendor Revenue in 2016 by segment, including Enterprise Operation Staffing. One of the key drives for the change is the savings in enterprise operational staffing derived from deployment of public, private and hybrid clouds, which is projected to drop from 22% ($290 of $1,452 billion) in 2016 to 9% ($144 of $1,631 billion) in 2026, with a CAGR of -7%. Cloud revenues are projected to increase from 6% ($132 of $1,452 billion) in 2016 to 46% ($743 of $1,631 billion) in 2026, with a CAGR of 19%. Traditional vendor revenues are projected to fall from 72% ($1,045 of $1,452) in 2016 to 46% ($745 of $1,631) in 2026, with a CAGR of -3%.

Market Infrastructure Drivers for 2015-2026

Figure 8 above illustrates Wikibon’s projections for the next decade. The key aspects of this projection is that cloud computing brings clear improvements in computing scale by reducing the amount of proprietary hardware and software infrastructure cost from different traditional vendors. This is already seen in the market, as the revenues from traditional arrays built on proprietary software and traditional networking built on proprietary stacks are in decline. Vendors who are offering software solutions that virtualize, integrate and maintain the complete stack are doing well. Almost all traditional vendors are striving to move into providing an integrated stack (Dell, HP, IBM, Oracle, etc.), and those who are not integrating themselves are repositioning themselves as OEM providers to integrators and SaaS services (Pure, Nutanix, etc.).

Wikibon believes there a constraints on growth of cloud computing, particularly for applications where latency is important, and where the cost of data transfer is high. Data in general has high density and high gravity. It is much easier, faster and cheaper to move code to the data rather than data to the code. There are large segments of future market growth where these characteristics are important, such as the Internet of Things (IoT) and future growth of Autonomous Vehicular Grids. Wikibon has also researched the growth of mega-datacenters, independent data centers within the same physical data centers as cloud providers and the internet POPs, with the ability to directly move data within the premise and many times the speed, much lower latencies, and much lower costs. Companies such as Equinix are leading this transformation, again minimizing the high cost of moving data by minimizing the distance travelled. Assuming that hardware and software stacks used in all the above scenarios are equivalent or nearly equivalent in overall cost, and avoid the proprietary hardware and software lock-ins, there are clear business and application benefits of using true private cloud solutions in these spaces.

The overall the overall view of computing with be hybrid, with both public clouds, true private clouds and traditional computing being important segments of this hybrid landscape.

Specific Vendor Revenue Projections 2015-2026

The second objective of this research paper was to look at the potential of cloud companies such  to grow within the movement to cloud. For example, AWS, IBM, Microsoft, and Google have large vendor revenues, competing in different segments of the Cloud marketplace.

AWS competes primarily in the IaaS/PaaS market, which in Table 4 is shown as about 2% of the overall market today growing to 9% by 2026. Traditional market analysis allocates money to vendors that make the final sale to the consumer. The consumers in Figure 5 are primarily the enterprise IT budgets and the hyperscale cloud providers. Although IaaS vendors such as AWS are paid primarily at the moment through enterprise customers, AWS and others also host other cloud services, such as SaaS and BPO. The next section explores the vendor revenue opportunities for a specific vendor, AWS.

AWS Cloud Projections 2015-2022

Amazon Web Services (AWS)  currently competes mainly in the IaaS segment. AWS has introduced a very large number of infrastructure services, and has essentially developed it own PaaS system integrated into IaaS. This is in comparison with Microsoft, that competes more broadly in IaaS, SaaS and some traditional software infrastructure market segments, and future Wikibon research will investigate Microsoft revenue projections.

Figure 9 shows the percentage that AWS has of the IaaS marketplace in 2015 and 2016, and the Wikibon projection of growth until 2022 to a market share figure of 35%. It has historically been difficult for vendors to achieve greater that 35% share, given enterprise requirements for dual sourcing, and the expected strong growth worldwide (and particularly in China, the EU and Japan) of local IaaS solutions.

Wikibon has always looked at infrastructure through the lens of workload type. AWS is a a very powerful platform for lower-value applications, with a very broad set of tools and services. However, the Wikibon analysis of reference accounts shows very few companies that running high-value workloads with aggressive response time, fail-over time and recovery time SLAs. The cost of migrating existing systems to existing public IaaS clouds is extremely high and risky; there is a simpler and far less risky migration to True Private Cloud.

In addition, Wikibon believes that more advanced industrial, transport, energy and medical IoT systems will develop as edge systems processing almost all data locally, with less that 5% of the data initially generated going up to the cloud.

Wikibon believes is will be very difficult for AWS to achieve greater than a 35% of the IaaS market.

Figure 10 shows a the same AWS expected Cloud Vendor Revenue against the total cloud vendor revenue projections. Even though AWS is gaining market share in IaaS, the other cloud vendor revenue segments are growing faster, and market share against the total enterprise cloud market will decrease over the projection period.

Figure 11 shows the result of running three scenarios: upper projection, lower projection, and expected. The heavy line in Figure 11 is the expected Wikibon projection of the AWS total cloud revenue, as in the previous two charts. The dotted line above is the upper projection and dashed lines is the lower AWS cloud revenue projection. The orange dotted line and percentage figures show the expected percentage of Total Cloud Vendor Revenue for AWS.

The scenarios show there is more downside risk than upside potential, especially with the international governmental policy trends.

Conclusions

Individual cloud vendors will find it very difficult to dominate the total enterprise cloud market, or segments within it. Vendors with reach in multiple segments of the enterprise cloud market, and have lower migration costs to their cloud offerings are likely to prosper best.

This analysis leads to the conclusion that AWS will very probably  enter other cloud markets to increase its total addressable market. Wikibon believes that AWS is likely to enter the True Private Cloud market, using its own cloud systems as the basis for these distributed systems. AWS has already put a finger in this market with the announcement of a new version of Snowball in 2016, with the ability to some limited AWS Lambda function in the snowball processor.

Action Item

Enterprise IT executives should expect that the cloud market will remain diverse, with multiple vendors in each cloud segment. Executives should look at two main factors in deciding their strategic approach; workload type and data placement. 

Footnotes:

The following sections include the assumptions and calculations used to convert the “Classic” enterprise cloud revenue segmentation to the enterprise cloud “Vendor Revenue”  segmentation. The change is only to the segmentation; the total cloud projection numbers are by definition the same.

Revenue from SaaS Infrastructure to IaaS

Part of the revenue from SaaS and the PaaS uplift goes to providing IT equipment, software and services.  In previous research Wikibon has estimated that about a = 15% of revenue goes towards meeting this requirement.

This can be met by either cloud IaaS services, true private cloud, or traditional on-site or colocated services. The percentage that is projected to go from SaaS to IaaS is calculated as:

• b= (Percentage of Total IT spend that is Cloud) × (Percentage of (IaaS + TPC) that is IaaS).

The calculation of the amount of SaaS revenue that is moved to IaaS is:

• c = a × b.

Revenue from SaaS Infrastructure to TPC

Part of the revenue from SaaS and the PaaS uplift goes to providing IT equipment, software and services.  In previous research Wikibon has estimated that about a = 15% of revenue goes towards meeting this requirement.

This can be met by either cloud IaaS services, true private cloud, or traditional on-site or colocated services. The percentage that is projected to go from SaaS to IaaS is calculated as:

• d= (Percentage of Total IT spend that is Cloud) × (Percentage of (IaaS + TPC) that is TPC).

This amount is added to IaaS, and subtracted from SaaS.

The calculation of the amount of SaaS revenue that is moved to TPC is:

• e = a × d.

Revenue from BPO Infrastructure to TPC & IaaS

Part of the revenue from BPO also goes to providing IT equipment, software and services. This is estimated by Wikibon to the lower than the SaaS percentage (15%); the figure of f = 10% is used as the estimate.

As in the “Revenue SaaS Infrastructure” section above, this can be met by either cloud IaaS services, true private cloud, or traditional on-site or managed services. The percentage that is projected to go from BPO to IaaS is calculated as:

• g = (Percentage of Total IT spend that is Cloud) × (Percentage of (IaaS + TPC) that is IaaS).

The calculation of the amount of BPO revenue that is moved to IaaS is:

• h = f × g.

The percentage that is projected to go from BPO to TPC is calculated as:

• i= (Percentage of Total IT spend that is Cloud) × (Percentage of (IaaS + TPC) that is TPC).

The calculation of the amount of BPO revenue that is moved to TPC is:

• j = f × i.

This amount is added to TPC, and subtracted from BPO.