Can Oracle realize its ambition to compete as a hyperscale cloud?
- Summary:
- For Oracle to realize its ambition to compete with the hyperscale cloud leaders, it has to step up infrastructure spend by billions
One very telling indicator is to follow the CAPEX — how much they're spending on building out their infrastructure — as former Microsoft veteran Charles Fitzgerald pointed out in a posting on his Platformonomics blog back in May. His conclusions were scathing, as I'll explain in a moment. But it may be too early to write Oracle out of contention just yet. Even though the vendor has an enormously steep hill to climb, it has an opportunity to emulate Microsoft's success in parlaying dominance in an existing market to become a cloud heavyweight.
But the Oracle cloud is not there yet, and the window of opportunity won't be open for long. Fitzgerald doesn't rate its chances in his commentary on a chart that compares Oracle and IBM infrastructure spend to the 'big three' of Amazon, Google and Microsoft:
That red line you may mistake for the x-axis is Oracle’s CAPEX spending. While they have finally separated themselves from the axis in the last couple years, they’re still by far the smallest spender of the bunch. Amazon, Google, and Microsoft each spent more on CAPEX in 2017 than Oracle has in its entire history.
While the big three have converged to spending about 12% of revenue on CAPEX, IBM and Oracle are eerily similar at 4.8%. It is interesting to compare Oracle to Microsoft, as both were asset-light software companies at the dawn of the century. They had similar CAPEX/Revenue ratios until 2005, when Microsoft started putting its software in data center-sized boxes.
But there's another interesting feature in that chart. As Fitzgerald notes, Oracle has at least started to take its cloud CAPEX seriously:
It is up-and-to-the-right as we expect from cloud providers and there is a discernable, if late for the overall cloud market, inflection point in 2014.
And if you look closely at the chart, while Microsoft did pull away in 2005, its CAPEX ratio started falling back a few years later. It wasn't until 2012 that it finally took off. A year later, Oracle's spending, albeit from a lower base, began a similar trajectory. If it is to continue that pattern, we should expect to see a big leap in CAPEX spend this year. Just as Microsoft reached a certain point in its evolution as a cloud platform and then began doubling down on its investment, Oracle may now have reached a similar point.
Think back to the history. Microsoft had launched Azure because it already had experience of running cloud infrastructure at scale for its Bing search engine and other projects. In the beginning, no one went to Azure unless they were already a Microsoft customer. The Azure customer base grew as companies migrated key pieces of their Microsoft server infrastructure to the cloud, such as Exchange. To make sure the Azure cloud was able to serve all of their cloud infrastructure needs, Microsoft added support for other platforms and introduced new services, most notably machine learning. Over time, Azure became competitive as a cloud platform even if you weren't already using Microsoft kit.
Oracle is now following a similar strategy. Having built up its cloud infrastructure platform over several years, the roll-out of its autonomous database offering this year now gives customers a strong reason for moving key parts of their on-premise infrastructure to the cloud. At the same time, the Oracle Cloud is adding support for other platforms and building out services such as integration and machine learning. Just like Microsoft with Office 365 and Dynamics, Oracle also has a significant SaaS estate that connects into its cloud platform, providing another source of customer demand.
(By the way, in case you were wondering, Fitzgerald's take on IBM is brutal: "We see IBM’s CAPEX slowly trailing off, like the company itself." I don't see any cause to dispute that analysis).
The question now is, can Oracle make the shift from a cloud provider serving the needs of its existing customer base to become a general purpose hyperscale cloud? Fitzgerald doesn't see it happening:
Oracle finally understands the threat cloud poses, and have aggressively responded, but don’t yet seem to have a realistic view of their prospects ... they’re way behind, need to spend tens of billions to have a competitive global cloud infrastructure, and have a much more severe customer problem than IBM: their customers hate them.
Yet even the customer relationship may be changing as Oracle recognizes the importance of keeping customers happy in the everything-as-a-service environment.
Oracle certainly seems to be setting its pricing to be competitive. A new study by Redmonk analyst Rachel Stephens has looked at comparative hourly price points of the main cloud players and while I do have some caveats about the assumptions on which the study is based, it suggests that Oracle is aggressively undercutting the market leaders on memory and processing. The picture's less rosy once storage and other factors are taken into account, so that the total cost of an instance ends up more expensive than its rivals — but it's in the ballpark.
In addition, the purpose of the study is to look at trends and it doesn't claim to directly compare prices. For one thing, the figures are based on list prices, so Oracle clearly has wiggle room to offer competitive rates to secure strategic deals.
The next two years are going to be pivotal for Oracle. With its autonomous database offering in place, it has the opportunity to bring significant cohorts of customers into its cloud platform. From a pricing perspective, the vendor seems ready to do the necessary. The big question is whether it has the stomach for the scale of spend it will need to build out its infrastructure to stay in the game. According to Fitzgerald's calculations, it will need to at least double that spend this year — that's $2 billion more — just to stay in the game. It will soon become evident whether it has stepped up to the plate when it publishes its financials. As Fitzgerald points out, "unlike marketing, CAPEX doesn’t lie."