Two converging factors are beginning to suggest that concentration of computing power into the hands of a few dominant players may actually now be taking place, at least in the realm of business computing.
Public cloud platforms
One of those factors is the growing trend for business software vendors to move their operations to public cloud platforms. I spent today in Amsterdam for the opening of Unit4's European customer conference and learned how this former client-server software vendor is becoming a cloud-first vendor, running on the Microsoft Azure platform. In doing so, it's following a similar path to Infor, which is transforming itself from a legacy client-server vendor to a cloud-first vendor running on Amazon Web Services.
A couple of months ago, I met the CEO of small business accounting software vendor Xero, who explained his view that moving his infrastructure to AWS builds in a competitive advantage over other vendors that are only now moving to their own proprietary cloud platforms.
The investment and the battle that’s going on between Amazon, Google and Microsoft is meaning now that public cloud has become compelling for any application being done ...
If you’re not in the public cloud you don’t have access to the commodity innovation that allows you to add that to your product and innovate to your end customers.
Then there was yesterday's news that enterprise quote-to-cash vendor Apttus has forsaken its exclusive relationship as a Salesforce-native ISV to launch a version of its applications running on Microsoft Azure. Although it's the first pureplay Salesforce-native vendor to make this move, it's not the first Salesforce partner to hedge its bets between multiple cloud platforms. Sage's CEO Stephen Kelly last year spoke about a multi-platform strategy that embraces Salesforce, Azure and AWS for different products in its portfolio.
The second factor is the increasing emphasis on machine learning as a crucial component of those cloud platforms. Xero's CEO emphasized that as the primary driver:
You have to have all those analytical, big data [resources]. The data center operators who are in this arms race are implementing these technologies inside their data center so you can turn them on straight away.
Unit4's latest release has an intelligent digital assistant at its heart, powered by AzureML. It was Azure's superior capabilities in machine learning that first attracted Apttus to look at Microsoft's cloud, as its CEO told me last week:
We’re not using Salesforce analytics because the level of sophistication we need is not afforded on that platform.
No wonder then that Salesforce yesterday announced the acquisition of deep learning startup MetaMind, apparently awarding its co-founder and CEO Richard Socher the title of chief scientist at Salesforce. It's not the first acquisition Salesforce has made of such technologies and is unlikely to be the last as the vendor strives to catch up with the more powerful machine learning capabilities of its cloud platform rivals.
Who'll be the top 5?
These two ingredients — a public cloud platform at scale and extensive deep learning capabilities — are the defining characteristics of a would-be computing power in the digital era. Who's in the running to make the top five?
- Alphabet, the owner of Google, can command some of the world's leading data scientists at Google DeepMind, which produced the AlphaGo AI that last month showed machines can beat humans at Go. The Google Cloud Platform is less of a leader but the two assets combine to make a strong play.
- Amazon has advanced machine learning that's been honed over many years of practical experience in its own web properties. Put that together with its public cloud leadership in AWS and it too has a commanding position.
- Microsoft has been the dark horse that suddenly makes a strong showing. Azure is developing stongly both as a cloud platform and as a machine learning resource.
- Salesforce is a strong cloud platform but needs to show it can be a player in AI. It's been acquiring a lot of data science expertise and technologies so can't be discounted but has to do more.
- IBM has both a cloud platform and the AI capabilities it's continuing to build in its Watson services. But it doesn't have a compelling story to tell yet.
Those are my top five but it's very US centric. There are substantial players in China that are also contenders. Don't forget the Japanese computing giants either — Fujitsu still harbors ambitions to be a cloud platform. Then there's Facebook, which doesn't offer a cloud computing platform but can't be ignored given its size and its work on AI. And Apple.
My list, however, notably omits some other well known names from the canon of enterprise computing. I don't see either Oracle or SAP doing much with AI. Hewlett Packard Enterprise decided it didn't want to be a cloud player. Dell and EMC are focused on providing infrastructure to others rather than running it themselves. These companies may already have missed the boat and be out of the running.
Will we be left with just this five? Or another five? Will there be a single winner one day? Personally I believe you need more players to ensure there's enough innovation going on, so I think five is much too small a number. But however few or many there are, I think those two attributes of being a cloud platform at scale that has AI at its core are the defining characteristics to look for in the winners.