In the big sense, you might even see Amazon building an appliance because they realize that not everything will be rented. After all, computing is no different than lodging or an automobile. Just because people rent a hotel doesn’t mean they don’t own their homes. They have vacation homes, they lease apartments. Similarly, people rent cars and Uber cars and they still own cars. There is a continuum of consumption that computing will follow and the existing players who are taking sides in this saying, basically, on-prem is the only way to do it are going to lose big time. It has to be a combination of on-prem and off-prem that will have to coexist.
What percentage of your sales today is software only versus appliance based?
We do a lot of work with OEMs and with Dell and Lenovo we started selling pure software as well. There are two different metrics here. There is revenue and there is billing. I think revenue-wise it is still pretty small now. Billing-wise it’s a little bit higher because we don’t recognize all the billing up front in the same quarter. In terms of revenue it’s less than 5% and in terms of billing it’s a little over 10%.
The reason I asked that is that I’m thinking through the answer you just gave me about the market transition. Do you ultimately see Nutanix being VMware-like in being primarily software focused versus the hardware sales?
Why did we have to do the appliance first? VMware needed to be acquired by EMC to [build] by brute force. That’s why they had to sell for $600 million. A company that was eventually worth $60 billion had to sell itself for one-hundredth the price to be able to brute force build the market. The appliance business helps us build a market without being at the mercy of OEMs. We took sales and marketing and logistics and break/fix and all the stuff in our own hands because we’re not just a technology company, we are a business. We are a standalone business that doesn’t have to be at the mercy of the large players who probably don’t want to see another VMware get built simply because it takes all the value out of their software.
As this opens up in the coming years, we will not take sides about appliances. If anything, we want to have a level playing field where different form factors and mechanics compete with each other and our sales force doesn’t take sides.
How do you differentiate Nutanix from Simplivity, VCE and other companies that offer HCI?
For one, we’re not a box company because we’re not just building a compute-plus-storage box. We’ve taken a full stack strategy, saying that we need to have our own hypervisor and our own management plane because the pane of glass is equally important. Everybody that you name right now is critically dependent on VMware for everything but storage. If you look at Cisco or HP or Simplivity or anybody else, they depend on VMware for the rest of the stack. Nutanix does not have to depend on VMware. That doesn’t mean that we’ll actually force our opinion down the throat of a customer who is a happy VMware customer. But at the same time, we believe that if we want to own the entire stack and do a better job of both web-scale engineering and consumer grade design, these are the ways in which we build our products now.
People talk about why, how and what. The why of the company is that we exist to make infrastructure invisible. The how, which is where we really differentiate, is web-scale engineering and consumer-grade design. You can start with three nodes and go up to 30 nodes or 300 nodes in a simple, one-click abstraction. That makes infrastructure extremely consumable. It’s very hard to touch us on both of these because these are fundamental. These are architectural principles for any company. Other players talk about doing this cheaper than Nutanix when the question is can you make it so simple that people can do this in two minutes versus two days or two weeks or three months of professional services.
I wanted to run by you a couple of comments from people that I’ve spoken to. I asked the head of VCE how he differentiates their product line from yours and Simplivity’s and he said, “What I think is fundamentally different is that we don’t think the entire market can be served with a single product. There are wide differences between customers and to be able to serve the market as a whole you have to have a portfolio.” What do you think of that claim?
It’s a very important observation. In some sense I agree with it but the difference is that I don’t agree with a portfolio of different operating systems. Linux is one operating system that’s great for all workloads. VMware is one operating system that does great for all workloads. Oracle was one database “operating system” that basically was built for all workloads. I mean, Windows is very similar. You don’t have different code bases of Windows, one for this app and one for that app. The most ubiquitous platform software systems do not have different code bases, one for each workload. That is where we are different.
We’re building one operating system code base for all workloads as opposed to one box for one workload that cannot spill over into each other, that cannot steal capacity from each other, that cannot say a word to each other. That’s still building silos actually and we don’t believe in the silo. So while we agree on the portfolio approach, the portfolio has to be different kinds of hardware like storage array, compute array, Flash array, memory array. We agree on different blocks of hardware but running a single cloud operating system software on them. The software must differentiate and the hardware should not be where you actually start to build the silos themselves.