Mark Shuttleworth, the CEO of Canonical (the business behind the open source Ubuntu operating system), has bootstrapped his company for 15 years, remaining doggedly committed to making the Linux-based OS a commercial success and eschewing the unicorn status-chasing of your typical VC-backed contemporary tech startup.
Having sold his first business after just four years for $575 million – at the age of 26 –and with Canonical, he admits, still not quite profitable a decade and a half in, the commitment is obvious. The strategic focus has not always been quite as clear.
Two years after a shakeup that saw his company pivot firmly away from desktop and mobile to servers/infrastructure, it has become more so. So, how is the company doing, is it profitable yet – and what of that elusive and much-promised IPO?
Mark Shuttleworth joined Computer Business Review for a chat from his Isle of Man home, to talk company strategy, the future of open source, Red Hat’s “legacy, shrink-wrapped Solaris equivalent”, why he’s been spending quality time in his garden, and more.
Mark, How’s Canonical Doing?
We have north of $100 million in run rate revenue. We are as close as we need to be to profitability. We are not particularly focused on profitability in a world where things like AI and IoT, are still essentially getting being formed; although I take cloud as pretty much done.
We’re growing nicely with mixed, quite diverse revenue base; a team of 600 people very much all over the world. And the biggest focus at the moment is enterprise infrastructure…
Why Do You See that as a Sweet Spot?
We already are the lead platform for public cloud* – so as enterprises move to public cloud it’s a very natural thing for us to do it with them. For most enterprises most of what they spend in private infrastructure is VMware and Red Hat.
Content from our partners
We know we can save companies 50 percent-plus on what they’re spending currently; make their developers more productive further up the stack, so that’s just a sweet spot for us at the moment. We made very good progress recently with headline wins in telco, finance, media and retail and I think we’ve only just scratched the surface.
From a growth point of view right now that is is where we are investing: in our ability to reach more customers and deliver more customers simultaneously.
One week this May we were busy delivering 27 new private clouds at the same time. That’s double the rate of the year before.
*The Cloud Market, which tracks the tools being used to deploy AWS EC2 compute, does indeed have Ubuntu as the most popular distro, though when it comes to images by owner, rather than platform, Bitnami – acquired this month by VMware – narrowly takes the lead…
Open Source Companies Are Struggling in Various Ways at the Moment with Threats to their Business Models. What’s the Answer to the Issue – and How Does it Effect You?
Look at the changes in attitude towards open source to start with. It’s interesting to see how the mainstream industrial players now are starting to think of themselves as both producers and consumers of open source.
Everybody’s becoming an actor in the open source ecosystem. Google’s an actor; Uber’s an actor; Airbnb’s an actor; Walmart is suddenly an actor producing interesting open source. These are all companies that specialise not in operating systems or applications, but in real world things that everybody can relate to, right?
“Ubuntu adoption is typically associated with some green field, at-scale, Github-centric change in their stack…”
So we publish a lot of open source. We publish expensive things for free.
The Ubuntu logo
But now we increasingly found ways in which the real commercial users want long term certainty that only we can provide them; we have found this new threshold of efficiency in an ecosystem for people producing and consuming open source.
Bluntly, if you’re a business and you want to be a producer and consumer of open source, you will get better value for money, sustainably, with Ubuntu than with Red Hat.
That’s really driving growth, particularly in areas of new activity. People don’t typically go back and rip out their old Oracle database on Red Hat and say they want to do that one thing on Ubuntu now.
It’s much more that they say ‘look, we’re moving to the public cloud. And as part of that we’re going to move to Ubuntu, we’re moving to machine learning and as part of that we will move to Ubuntu… so it’s typically associated with some greenfield, at-scale, Github-centric change in their stack.
Cloud Aside, What’s Driving the Uplift for You?
IT is schizophrenic at the moment. You have some people in IT who still long for the days of Solaris. So Red Hat can speak to them very compellingly. But you have other parts of IT that are saying ‘look if we don’t move first then Silicon Valley is going to disrupt our business: we need to get in to CI/CD on Kubernetes, on a multi cloud environment. We need to be doing all of these things that Silicon Valley is doing on Ubuntu. And it’s very difficult to be two different things at the same time…
Now inside any large business you’ll find both of those. You have the people taking care of the legacy. You also have people who figure that they have to lead internal disruption before they face external disruption. That’s true in banks. That’s true in retailers, it’s true all over the shop.
Lots of businesses that are longstanding VMware and Red Hat customers are now signing up with Canonical for a new projects. They’re not doing a rip and replace. They’re not throwing the other guys out. They’re just essentially preserving that at this stage, while growing a new Ubuntu estate.
What Are Your Thoughts on the Open Source Licensing Issues that have been Playing Out in Recent Years?
If you zoom out a little bit on the kerfuffle, the question is ‘where’s the give and take?’ If you look at the first real GPL (licence) generation, there is a very hard bargain at the heart of that, which is that if you were a commercial beneficiary… then you also are going to be contributing to that GPL software; period.
That was a very hard bargain. But it worked.
Then we entered a slightly loopy time where some very big companies that didn’t like the GPL generally, said ‘well someone should take a whole bunch of money and they should put it into Apache licence software that anybody can use to do anything, with no hard bargain’.
If you look at Kubernetes today, Google is spending a huge amount of money building Kubernetes and giving it away [as] they have other reasons to do that. Android the same.
There is [still] a hard bargain: if you aren’t paying for the product, you are the product.
But if a company wants to spend large amounts of its money putting it into an Apache licence thing and giving that away, well then fine. If you believe the company will keep doing it then you have a model…
We’re now into a new era of how software works economically in the cloud era.
And I don’t think anybody’s yet managed to design a balanced set of gives and gets that works for that era. The cloud response is sometimes little bit ‘let them eat cake’ and that’s not going to stand up to the test of time. But in the end I have confidence that a bargain will be struck. The consequences of a shift in the economics of open source always finds a kind of a balance otherwise it doesn’t thrive.
It’ll emerge, I’m sure, and it doesn’t overly affect us.
Why Not?
Because we are not investing in a very narrow slice of innovation.
If you think of what it’s like to be Mongo, or Redis, or Kafka, you’re hiring a team of super specialists to do one thing, and one thing only: that’s very different than essentially helping companies engage with all those open source [tools] more efficiently.
They’re totally different profiles of work, and the value propositions are totally different too.
Talking of Kubernetes, You’ve Had a Pop at Red Hat in the Past for their Approach to Containers. Talk Us Through the Issue/How You See Container Use Evolving…
I think what I said was that IBM’s primary focus was on (Red Hat’s offering) OpenShift. And the important thing there is that OpenShift is much more than container orchestration; it’s a Platform-as-a-Service (Paas): our view is that customers don’t really want PaaSs anymore. PaaSs read well with old school IT because it sounds like you never have to do any work and you get to look great.
It’s not a story that reads well for Silicon Valley. You don’t see Silicon Valley leaning on PaaSs; you see them leaning on Kubernetes.
Kubernetes is just part of Linux, so this is a really challenging story for Red Hat with OpenShift.
Mark Shuttleworth: “PaaS’s Read Well with Old School IT”
If you want to charge double the cost of Linux, or triple the cost of Linux for your PaaS, then you have to convince people that they need a PaaS. If actually all they want is Kubernetes built into Linux, well Kubernetes is free with Linux! That’s going to be a real pain point for the Red Hat numbers. They saying that they will grow dramatically on the back of OpenShift. I don’t see a huge amount of appetite.
I guess it comes back to that difference of opinion between guys who maintain a lot of legacy that essentially is a unique Unix crowd and the guys who are trying to deflect or beat an emerging Silicon Valley focus on brick and mortar industries. That old IT crowd is going to love the story of ‘OpenShift will run your databases for you’.
Read this: Canonical Fires a Shot at Red Hat with New Ubuntu Infrastructure Bundle
The flaw in that is that in 10 years, PaaS have actually never really been successful at for any length of time. The guys who know what it takes to make things work are the guys who know that they’re going to have to shoulder a certain amount of responsibility, and want the leanest stack underneath them to do that: that’s the cloud that CloudBees is referring to; that’s the crowd that I’m referring to.
If you go and talk to people who’ve had a PaaS for three, four, five years, ask them ‘are you on track with moving the number of apps to that PaaS that you thought you would move?’ and the answer is ‘no’, universally. ‘No we’re way off track’. But if the PaaS is only going to run 10 percent of your apps. Why should you be spending you know 200 percent more on infrastructure to run 10 percent of your apps better?
OK, So You Seem Really Positive About the Opportunities at this Juncture – Are You Going to Raise Some More Capital for Expansion?
That’s right. We do intend to do a growth equity round and then an IPO. I think we will go out some stage this year.
How Much Would you Like to Raise?
I think it’s quite important that we don’t do it until we don’t need the money… then that’s going to be a question of what can we usefully spend.
I think we’ve done very well operating lean, and I don’t see that we want to change that discipline, while at the same time you know getting some of the larger funds a meaningful opportunity to invest.
They’re likely to want to build a bigger position than we are likely to want to raise. We’ll see…
Final Question: What’s the Focus When You’re Not Working?
I’m building a botanical garden. It’s a lovely, lovely opportunity to celebrate the diversity of life and ecosystems. Just doing something that can’t be rushed.
A lot of what we do is figure out how to go faster on open source. But if you stick an acorn in the ground, you know there is absolutely nothing you can do to turn it into an oak tree in less than a generation.
I like that change in timescales and the thinking that’s required for for that project…
