Some interesting developments are shaking up enterprises at the moment, especially when it comes to cloud technology.
Many, for instance, have already started their cloud repatriation journey, realising that maybe not all data and applications are best suited for public cloud.
Rangarajan Raghuram, Chief Executive Officer of VMware, spoke to Frontier Enterprise about his two decade-stay with the company, enterprises’ current cloud strategy, the implications of AI, and what can be expected tech-wise in the next 10 years.
You’ve been with VMware for about 20 years, and recently transitioned from COO to CEO. What was that transition like? How do you look back on those 20 years?
I started with the company when we were probably 250 people. Today we have 37,000 people. Before we had US$100 million in revenue; today we are over $13 billion in revenue. So it’s been a significant journey. It’s been an awesome experience of building a company from scratch all the way here, going from a single product to a multi-product or multi-business company, operating in so many geographies. It’s been a huge learning experience.
We’ve also been fortunate to impact so many customers, many of whom are running their mission-critical applications on our platform. So it’s an awesome responsibility as well. From that point of view, it’s been a great learning experience, an opportunity to have an impact at scale in the industry. But at the same time, VMware, its core values, and how we work together as a team inside the company, have never really changed. Even today, many people that have left VMware say this was their best professional experience. So the culture, the values of VMware remained constant. The core things have been constant, but the job has been very different.
Diving straight into cloud repatriation, do you see some sort of final equilibrium between multi-cloud, on-prem, private cloud, and public cloud?
Customers are now going from “cloud-first” to “cloud-smart.” There was a panic even before the pandemic. It was practically a herd mentality like, “I’m going to evacuate my data centre; all applications will run great on the cloud.” There was an industry midfield fueling that as well.
It turned out that unless you redo your application to take advantage of what the cloud is really good at, you’re not going to see the benefits of the cloud. So, customers don’t have enough developers, or enough money, or enough business justification to rewrite everything. They’re saying, “Okay, what runs best where, and let’s take advantage of that for what the cloud is good for,” as opposed to, “Those things that are getting repatriated are applications that probably aren’t suited for the cloud anyway.”
Given data leak challenges associated with open-source artificial intelligence, how do you see the infrastructure for enterprise AI develop moving forward?
I fundamentally believe that most medium to large enterprises will do the bulk of their enterprise applications using some sort of private AI-type architecture. It could come from VMware, or it could come from a hyperscaler, or somewhere else. If I asked you, “What is a company? What is the secret of a company?” A company, at the end of the day, is its people, the business processes they have, and the knowledge that’s encoded in their IP, however unique that IP may be. They can protect that. Having that be in a public model, regardless of what the public model vendor says, is not something that people will do.
Take VMware as an example. Our code for our vSphere, for instance, livens within our firewall. So, we would never use GitHub Copilot or some other public platform. We downloaded an open-source code completion model called SafeCoder and trained it against our data running inside our firewall. The minute we sent that to many of our customers, they were like, “Can I come to your shop and see what you’ve done?” Not because they want to use ours in a shared way, but they want us to replicate that within their own infrastructure. So, I see that as what is happening, and what will continue to happen, which is not to say that these large-scale models, like OpenAI, won’t flourish. For example, if you’re creating a marketing webpage, you’re probably creating it from publicly available content anyway, so you don’t really care, right?
But on the other hand, for the serious stuff that encapsulates that IP, let me give you another example. We, as VMware, have 300,000 customers, which probably means we have around 10,000 contracts with customers. Now, I have to take those contracts to our legal department to allow us to move those contracts into Microsoft’s public cloud. Over our dead body, right? But those contracts are going to be hugely valuable for generating the 301,000th contract, or for understanding what sort of terms we’ve given to a class of customers. That’s what I mean by those classic primary use cases.
With 5G having implications on edge infrastructure, and the AI revolution coming up, it’s a pretty exciting time for technology as a whole. How do you see the next few years panning out, and what most excites you as the CEO of VMware?
I think it’s a cliche to say that the next decade will be the most exciting we’ve ever seen. Every decade has proven this to be true, so it’s safe to assume this decade will be no different. You’ve identified some of the defining forces shaping the future. Clearly, AI will be a significant driving force in the next decade of innovation. Whether it’s robotics, more advanced cameras, or autonomous driving, this intersection of the physical and digital worlds will represent some of AI’s finest applications.
On the enterprise side, developments in AR and VR will happen more slowly due to the need for fundamental hardware breakthroughs. Not to say that AI doesn’t need breakthroughs, but we’re further along that road. Once those breakthroughs happen, things will become very real for consumers. For enterprises, developments like digital twins and new forms of collaboration will materialise. But what ties all of this together? It’s new forms of computation. AI is driving these new computational forms, like innovative processors and memory technology, which in turn enables GPUs to function effectively and leads to new networking technologies. These are the areas I’m familiar with as a systems and infrastructure guy, and thinking about all of that happening, it’s going to be truly amazing.