Transcript
get stored in storage. As artificial intelligence accelerates, the infrastructure we usually think of running in the background is undergoing its most profound transformation yet. Now it's the other way around. Now applications are using the data to build intelligence out of it and power those applications. So things have completely flipped in more recent days with storage. In the era of intersection between infrastructure and intelligence, we need infrastructure to run all the AI applications and workloads. And then you need data to power the intelligence for applications. Virtualization really transformed compute. Earlier compute was tied to a server. You had to buy a server or a desktop or a laptop to get the compute. Now virtualization made compute portable. You could move compute around as VMs or containers. And this change really paved the way for everything software defined. So that was one big transformation that happened around 2000 to 2005 time frame. After that, the cloud transformation, that has been a big one. It has really democratized infrastructure. Like today, if you want to build a new app, you don't have to build a data center to build that app. You can directly go to the cloud and build that app. So it has created some more work, like now you have to manage the cost, you have to manage the security, you have to manage the networking. So it has added some complexity, but definitely it has made the infrastructure more accessible to everyone. And I would say the latest one is the AI, like this is a brand new one. It is going to change things significantly. Even for us, the entire software stack will change to support AI. So this is still playing out. We'll see how that evolves over time. AI needs data and it needs a data-ready platform. And that's where we see that in order to make a data-ready platform, you need to provide certain capabilities. First, being able to consolidate the data from different sources and bring it all together. The second piece is around making the data clean so that it can be consumed by the AI LLMs. And then third part is providing fine-grained permissions so that LLMs learn only from the data it is supposed to. So clearly lots of transformation from a passive data repo for files, images, audio, video, to becoming a full data platform to AI-powered applications. So that's the big change that we are seeing. We were talking about the new trends. It's different now. What was it like before? Yeah, storage was a passive data repo, where customers put their files, their pictures, their videos, their audio. So it was truly looked at as a passive infrastructure, mostly for storing data. And now it's very different because now it's an active part of an AI stack, where it needs to power all these applications with the data that it stores. So you see it more active. Absolutely. So it's an important part of the new tech stack for AI. Let's talk about customers and what you're hearing from them. What are their needs? Do they have some new needs they're struggling with? Yes. So typically I would say two categories of customers. One who are already deep into AI. For them, the AI data platform I just talked about, that's very important for them. But then there's another big set of customers who are still not there. Their problem is how to manage data at scale. Data is doubling every 18 months, and now you have to support and manage that data. So that's their number one problem. Second one that comes often is the data is now fragmented. Like it's everywhere. It's at the edge in your phone, like it is in your applications, it's on the desktop. So how do you provide a consistent operating model to manage all this data? Then security is very important, like ransomware, really many customers have hit ransomware. So they really want to know how the system, the storage solution itself can protect the data and not have to buy a separate solution. And also the cost, like the total cost of ownership is always a very important factor. Given the growth in data, they're always looking for a much better, most cost efficient way of storing their data. So these are, I would say, the four things that I hear often from the customers that we talk to. Do you ever hear them say, I wish I could do this, or I can't do this, but I needed to do that? Yes. So that we hear a lot more around the scale part. So today, generally, most customers have different solutions for running at the edge in the core, in the cloud. And now they have four different solutions that they are running with, right? And that makes it very fragmented. And they say, wish I had one console, one policy that I could apply across all of this data and that would have made their life so much simpler. Let's talk about your career a little bit. What got you into technology? What was that feeling like? Yes, I was always fascinated by technology. I always wanted to build products that could help customers solve their problems. And for the first 20 years or so, I was primarily focused on building things. And the last seven, eight years, I found it equally rewarding to take this product or products to the market and help the customer achieve the outcome that they want from these products. So the whole journey, which is building the products that customers love, and taking that products to the customer market, is what has been very rewarding and something that motivates me to continue with the technology space. And do you remember when you first started, what were some of the big challenges or exciting things that you wanted to work on? Oh, absolutely. So I still remember those days when if you had to make an international phone call, it used to cost $2 for one minute. Look at today, it's almost free. You can use WhatsApp and make that call. So the first one for me was the whole telecommunications shift that happened between 1995 to 2000. And I was actually in the middle of it working for a startup, which was building multi-protocol label switching to really get the service provider infrastructure more seamless. And over time, it's really brought the cost down and made everything more accessible to the end users like me and anyone else. And when you were working in telecom, could you imagine hybrid multi-cloud, containers, VMs living together, like all the world that we're in now? So at that time, it was all monolithic. You had one infrastructure which ran everything. We didn't have the flexibility of a VM or a container that we could move around and solve the right problems with the right form factor. I think after the virtualization, that definitely got simplified. Now we can have a very flexible, software-defined telecom infrastructure. What's the advice that you give people to motivate them? Learn the first principles of systems building, like don't focus only on the API layer. Understand how systems are built so that you can appreciate and you can make better things after that. The second one would be the ability to deal with ambiguity. Things are changing very fast and we have to be able to cope with uncertain things. And look at AI, for example, things will evolve and we have to be able to make progress with what we know and be ready to adapt to newer changes. That's very important. And third one is having endless curiosity. Things will change fast. And if you're not curious enough, we'll not dig in and learn new things. So if we want to stay relevant, we have to be endlessly curious.
