The cloud computing revolution is completely transforming the world of IT. According to research and advisory company Gartner, cloud spending was estimated to total around $260 billion in 2017, and expand quickly to $411 billion by 2020, as companies are really only at the beginning of the shift from private data centers to public and hybrid clouds.
Not only is this leading to big gains for leading cloud companies, but the transition may also upend the traditional software and platform products once run exclusively in private data centers. One notable product under threat is the traditional database, and specifically database leader Oracle (NYSE:ORCL)
It looks as if Oracle isn't taking things lying down. The company recently announced a plan to better compete as a major player in cloud infrastructure. The foray into cloud infrastructure will be costly, time-consuming, and -- since the payoff will be years in the future -- uncertain. But it looks like absolutely the right move for Oracle. Here's why:
12 new regions
The announcement was fairly specific: Oracle will build cloud infrastructure data centers in 12 new regions over the next two years: two in the U.S., two in Canada, and one each in India, Japan, the Netherlands, Singapore, South Korea, Switzerland, Saudi Arabia, and China, where, according to The Wall Street Journal, it will work with state sponsor Tencent. The additions will greatly add to Oracle's current cloud footprint, which includes only three regions, in Phoenix, Virginia, and Germany, and one opening soon in London. A cloud "region" is a geography that includes two or three "zones," with each zone encompassing a large data center.
The cloud opportunity
Oracle may be late to the infrastructure game, but it's correct to jump in. The subsector of infrastructure as a service was expected to total around $35 billion in revenue 2017, and is the fastest-growing segment of the cloud market, according to Gartner, with an estimated annual growth rate of 36.6%.
Giants Amazon, Alphabet, and Microsoft spent roughly $41.6 billion in capital expenditures and capital leases last year, according to The Wall Street Journal. While not all of that was for cloud data centers, all three have mentioned cloud infrastructure as a main strategic focus.
There are also smaller companies investing in cloud technologies, and the combined capital spending of 19 cloud service firms tracked by RBC Capital Markets reportedly reached $63.8 billion in 2017, expanding to $81 billion this year.
With its announcement, Oracle has made a statement that it wants to crack into the top few cloud operators. Its 12 new regions will bring its total to 15, which compares with Amazon's 22 regions (by the end of this year), Microsoft's 42, and Google's 15 (as of last year). IBM, a company that finds itself in a position similar to Oracle's, currently operates cloud data centers in 19 countries.
Heading off migration
As these giant infrastructure companies have gotten more clients to move to the cloud, all are trying to sell platform and software services on top of their infrastructure. Yet because databases carry so much crucial data for corporations, it's extremely expensive, time-consuming, and risky to change vendors. Still, as the database market is a huge opportunity -- $45 billion and growing -- it's not stopping these giant companies from trying breach Oracle's database moat.
For example, Amazon recently announced it has migrated 54,000 databases via its AWS database migration service since the service was introduced in 2016, accelerating over 100% in 2017. True to form, Microsoft followed soon after with its own Cosmos database migration service in May 2017.
However, moving databases, especially for large companies, is still no easy task. I recently wrote about how Amazon has been working on its own database product since the early 2000s, but is still itself not yet totally weaned off legacy Oracle databases.
That means Oracle has time to figure out a way to prevent customers from leaving for another cloud giant, and certainly having its own infrastructure service is a way to do that. Like Microsoft and IBM, Oracle already has a robust enterprise sales force, so I don't believe it's a stretch to think it can land its own infrastructure clients, especially those that already use Oracle databases.
Of course, there are significant up-front and ongoing costs to running infrastructure data centers -- one data center alone could cost hundreds of millions of dollars per year, according to analysts cited by The Wall Street Journal [subscription required]. Still, with $20 billion in net cash and generating about $15 billion in operating cash flow per year, Oracle has the means to compete in this high-growth area. Since the infrastructure move has both offensive and defensive characteristics, I'd say it's about time.