Episode 4
Courage, Context, and the Next AI Business Model
Key Takeaways
· Entrepreneurship requires courage, often derived from studying history and understanding how disruption unfolds over time.
· Early-stage companies frequently face existential technical and market challenges before finding product-market fit.
· Leadership during adversity requires focus, composure, and solving problems rather than assigning blame.
· The SaaS model that disrupted perpetual software may itself now be facing disruption in the AI era.
· AI-native companies are likely to move toward consumption-based pricing and outcome-driven value models.
· Generative AI has the potential to transform every role in a company, not just software development.
· Leaders must balance peacetime leadership with wartime thinking to anticipate disruption before it arrives.
· The ultimate challenge of AI is ensuring it serves humanity, not just corporations.
Dheeraj’s Watershed Moment
Dheeraj’s defining watershed moment came during the early months of the COVID-19 pandemic in 2020.
After more than a decade building Nutanix, he realized the market was shifting toward streaming infrastructure rather than owning it. With global data centers shutting down and cloud adoption accelerating, he faced a profound decision: remain solely an operator of the company he had built or step back and rethink the future.
That moment forced him to separate his roles as founder, operator, and investor. It ultimately led him to start again and build DevRev around a new problem space shaped by AI and changing enterprise workflows.
Full Transcript:
Declan Waters (00:05.842)
Hi everyone, I'm Declan Waters and welcome to the Watershed podcast. Today I'm absolutely thrilled to be joined by Dheeraj Pandey, founder and CEO of DevRev. Dheeraj co-founded Nutanix in 2009 and he helped create an entirely new category in enterprise infrastructure. And he scaled it past a billion dollars in revenue, took it public in 2016 in one of the defining IPOs of that cycle.
Then at the point where most founders transition into board seats and legacy mode, he reset the clock and started it again from scratch. I'm so pleased you did that, Diraj, and I'm so happy to have you on the podcast. Welcome.
Dheeraj Pandey (00:48.088)
Thank you. Thank you, Declan. I love the watershed word because the very fact that we are working again is also equally watershed as anything else.
Declan Waters (00:56.498)
It is. We've come back full circle after almost over 10 years, which is wonderful. So thank you for joining us. Okay. Let's get straight into it. And I wanted to just touch on your origin story and maybe your first watershed moment. So just to set the scene for our audience here, as I mentioned in my intro, you built one of the most defining enterprise companies of the last decade.
So, I'd love to know, just to kick us off, what is the through line from that young engineer to the incredibly successful founder and entrepreneur that I'm talking to today.
Dheeraj Pandey (01:36.162)
You know, a lot of it has to do with how business and life weave together. You know, your understanding of history and politics and geopolitics and tech history and even, you know, things like, you know, the last 500 years of the way empires were, you know, even thousand years of how empires were, you know, built and lost. I think a lot of that is useful because when you talk to people, you need to bring out history.
So, I'm a big fan of storytelling and history is a great way that I actually learn to not just manage, but also in my own ways inspire and motivate. And that was the thing that I always had throughout my life. I mean, even when I was very young, I used to think hard about would I be getting into economics and would I be getting into physics and finally computer science was it when the internet was born, I was about to get into college.
And software became a means to an end. It was never an end in itself for me. It was an expression of the modern, it was an expression of the disruptive and the rebellious. And that's what basically made me an entrepreneur over time.
Declan Waters (02:51.986)
The history side of it is intriguing for me because I've personally seen you talk to media influencers and others over the years and you have this wonderful way of connecting the dots, right, with history and technology and the intersection of everything. And it just makes the story much more interesting. Was history important to you before you started your entrepreneurial journey or did it come during that process?
Dheeraj Pandey (03:27.994)
It's about courage. At the end of the day, entrepreneurship is about courage. It's about defying conventional wisdom because if you were to ask people to do something, they're like, well, why would you? Because there's already N different versions of it. So you look at history to derive courage. And that's what I've always done. Now it doesn't repeat, history doesn't repeat, but it rhymes. And there ways in which you can actually go and you know, even talk to yourself and, you know, riff with yourself before you go and riff with others about why this could work. It's also about imagery, you know, at the end of the day, even when you're on Zoom and your video is off, you need to create an imagery for others to get convinced.
Like today I was in a call and we were talking about how user interface is changing dramatically from, you know, just graphical user interface to conversational. And, you know, you have to invoke things like even from 1980s, mid 80s about when Intel was pummeled by the Japanese on the memory business and it took them like four or five years to actually figure that out.
And they went to their customers and like, hey, we're getting out of the memory business and they chuckled and they said, what took you so long? You know, so the market knows more than and is ahead of you because you're in this innovators dilemma. And similarly, conversational UI versus graphical UI. You know, could be in this side of this where the world has not changed and AI is not accurate enough. But if you got on the other side, you realize how you could double down on the new and the modern to really change everybody's lives. So it's always been about using the past to basically catapult to the future.
Declan Waters (05:03.922)
Thanks.
Declan Waters (05:16.306)
I want to stick to that, the courage that you talked about and just maybe use that as an opportunity to pivot to Nutanix. Most of us won't ever experience what it's like to build a company like Nutanix. You've done it and the courage that you must have had must have been extraordinary throughout different stages of the company's journey. Could you give us a sense of some of the hardest moments that you felt during those days. Maybe there was a point where you genuinely thought, you know what, this might go the other way, right? I'm sure there were some of those moments.
Dheeraj Pandey (05:54.53)
Yeah. I mean, it happened in the first three, four years, probably before you joined. Because for the audience, know, Declan and I used to work together. We were fellow newtons, as we would call ourselves, know, mutants with an N. You know, we had three near death experiences and a lot of had to do with the not the technology, but the product and also the market. I mean,
We had made some assumptions about our technology and the partner that we were dependent on was VMware and they were not ready for our kind of a workload. How serious a workload we were trying to run, we were trying to make software defined when most of the things that we were doing before us was being done in hardware, that to proprietary hardware. So there was a couple of times in those first three years we were like, man, I don't think this architecture could work. And there was a real near-death experiences. And then finally,
When VMware turned against us, the only startup that they turned against was us because we were so good at what we were doing, including hiring people from them. It felt like we were cornered and how we got out of it.
And, you know, they provided an escape hatch to us because when we had no other option, we embraced open source and made it our own to even compete with them over time, you know, to create a long lasting company that was not just about you know, data management and storage virtualization, but also compute virtualization and network virtualization to really go and work the whole body on software defined data center, not just software defined storage. I think a lot of those first four or five years were extremely instrumental in building the future of the business.
Declan Waters (07:41.362)
And what changed you after that? How did that change you in terms of your perspective on things, your views on leadership? How would you characterize that?
Dheeraj Pandey (07:52.718)
You know, we never blamed anybody. We never pointed fingers. I mean, the biggest ways we reduce stress was to be focused on fixing the problem. And I think when you get out through all of that sort of mess, you realize that nothing can be too sort of impossible or difficult. And there's nothing in life that cannot be captured. know, you can basically go through the worst of moments.
So I think it's normalizing of the adversity that I really took away from there. And, you know, how do you behave when you have turbulence and you've lost an engine because the passengers look to a lot of that stuff too. If you're the pilot of the plane, you can be panicking and you can't be pointing fingers. You know, what you got to figure out is how do you land?
I mean, in fact, after we went public, we had another such moment of, what the public cloud is here. How do we change the wings of the plane at 35,000 feet, which is business model change, going from appliance to software to subscription to annual contract value selling. A lot of that stuff happened at the public company. In fact, VMware had to go private to make those changes. So when you normalize adversity and you go through all that stuff, I think...
That to me is what defines leadership.
Declan Waters (09:14.81)
Yeah. I've been looking forward to asking you this question and it's a simple question, but for some reason I've been really looking forward to asking you this. Did success feel the way you expected it to during that time? Because there's so much success happening around you.
Did it feel the way you thought it would?
Dheeraj Pandey (09:37.837)
It's a very profound question by the way.
Dheeraj Pandey (09:44.312)
I mean, the humbling part of that success was that it's never a straight line. Internally, you had to keep thinking about what else will go wrong and how do you see around the corner.
I think the best piece of that success was employees and customers. We made so many employees go and pay for their cars and their homes and mortgage and for customers to go buy stock in the company because they were such believers in our customer service and our ability to never fail them. That was not just dopamine, but that was endorphin and that was oxytocin.
Declan Waters (10:04.56)
Yeah, Hands acrossed, by the way.
Dheeraj Pandey (10:31.073)
Serotonin, everything that you could expect from purpose you got from customers and employees.
Declan Waters (10:35.568)
Yeah, absolutely. Yeah, well said to rush. OK, let's zoom out a little bit. We're recording this on February 24th, 2026. Tonight is the State of the Union address. So let me ask you this. If you had to give a State of the Union address for the AI market right now. What would you say?
Dheeraj Pandey (11:04.845)
I think it will start with the status quo, the last 20 years of tech. The SaaS market is beginning to realize that the market is pushing back and not just Wall Street, but also Main Street. And this happened 15 years ago when people had box software and perpetual license models of licensing, the market pushed back and they said, look, you got to do more for me because this is not going to work. And open source was there and e-commerce was there with B2C side, we're consuming and streaming a lot of our commerce through these big giants.
So what we saw was, you know, companies coming in and saying, we'll do more for you. You know, this old model of buying things for five years and, risk all being yours and all that stuff is actually the past. And therefore subscription came about because people said now you can bite size it, can bite by the month, by the year, by three years as opposed to perpetually. And that's what's happening right now. Subscription itself looks like box software. There's too much wastage, there's too much complexity. Partners come in, then hardly partners, they're going and really being the messiahs of complexity.
They throw arms and legs, time and material at the problem. And the market is like, what do we do differently? And AI companies are coming in and saying, look, we'll do more for you. And by the way, when we do this, we'll share the risk with consumption. If you don't consume, you don't pay. until we show you outcomes, you also don't pay.
So there's a new business model that's upon the incumbent SaaS companies, who only 15 years ago went and disrupted the old business model of perpetual software. And in this AI era, there'll be winners and losers, you know, which is a cliché, but what's also upon us is companies that are basically saying we have the context, we have the data, we have the eyeballs right now. SaaS has all that stuff. Therefore you trust our AI. Well, in reality, they don't have the context.
I mean, even Microsoft doesn't have the context. At the end of the day, they have OK context with the data they own, which is their emails and their calendar and their Microsoft Teams and maybe a little bit of CRM and ERP that they have, but they don't have the context of everything that they never owned. Salesforce is the same problem. They probably have context of the data they have, the people, the users, the permissions and all that stuff, but they barely know anything about non-workforce environments and so on and so forth.
So the big opportunity for AI companies is they will do more because Salesforce alone doesn't define your customer. It's too much that happens outside of Salesforce for you to really understand your customers. And AI has to go build that context and to go and help them not just search through this knowledge, but also answer, which is a bigger problem than search and finally make it actionable.
No different than the way developers are actually becoming generative. I mean, we want everybody in the company to be generative. And we've seen that half of the tokens are being used for code, which is a universal language for writing software now. But text is another one, which is natural language.
And how do you do this for product management, project management, cost customer support, software, sales and marketing and everything else, we need to think about what's the cloud code for all these other people, not just for developers, which I think they're feeling pretty good about, but for everybody else, it's still a missing thing. And that we believe is what's going to make the whole industry generative. like all things bubble, we probably over-investing in hardware and data centers.
It's only question of time. It's not about if, it's when we realize that man, we have over capacity and it will probably take a little longer to consume all this capacity.
Declan Waters (15:31.602)
Thank you, Dheeraj. That was a really beautiful framing of everything. How do you, how does Dev Rev fit within side that landscape that you've just described?
Dheeraj Pandey (15:45.134)
Well, on one hand, we look at the new sort of user interface modality of conversational surface. So we look at what Cloud Code is doing, or Cloud Cowork is doing, if it's trying to do this on a laptop and a mobile app with a real conversational design thinking. It's in fact more about design than anything else. I mean, all the model companies that didn't have a great consumer app, they've failed.
They've kind of lost their way. And the ones that have really thought about how are you consuming it are the ones who winning. So business software is gonna be very similar. You've got to think about a new design modality, which is conversational. And then the boring stuff comes in, which is data and permissions and users and workflows and process and all that other stuff. So we have chosen, you know, two big sort of ecosystem, Salesforce and ServiceNow.
And by Corollary, you can say Zendesk and Atlassian and all this other stuff, but a lot of work management software, whether it's for sales or for operations people or support people, I think it's a big opportunity to take a lot of that mundane-ness of SaaS, that data and its permissions and all that stuff, and bring it to a cloud-like interface and make every employee you know, understand context of the customer, context of customer success, context of product and how product needs to serve the customer. Do it really, really well and make every support engineer, every sales engineer, every sales representative get to become generative.
Declan Waters (17:28.722)
You're on a journey to build what I think is going to be another iconic company, Dheeraj, right? And it's a very exciting time for you and the whole team. It's a very different company from Nutanix. So could you maybe touch on what are you doing differently this time around from when you did it the first time with Nutanix?
Dheeraj Pandey (17:49.07)
Well, let me start with what we're doing similar, an immense focus on design, like infrastructure was boring, VMware hadn't done as good of a job. I mean, the hardware guys had done nothing about design. So it was a great opportunity for us to bring it all together. And because we brought a lot of components together, changing that surface was actually a great opportunity. And it's very similar here as well.
You know, there's lots of silos. Every department thinks they have the context, but none of them truly understand the customer and how they serve the customer with their products and services. So we bringing a lot of that together into the knowledge graph of a company. So that thing is very similar to the way we used to talk about converged infrastructure back in the day with data centers, which is basically what even AWS was doing back in the day, you know. Now what's different, you know, we are having to think really hard about the human in the loop.
And while 15 years ago, 10 years ago, we're thinking about Apple-like thinking, which is like, how do you make humans delighted with this thing? I think we are at a point where we have to say, why do we even need a human in the loop? Now, whether it's a customer experience, which is how we help our customers, customers self-serve everything, or it's even employee experience, like whether it's going from search to answers to actions.
How do you make the human really focus on bigger, better things and let the machine take care of everything? So design has changed in that way where you're constantly asking the question, what will be different and how do you avoid the human in the loop who's not creating any value probably except for moving hot potatoes, tickets for example. That's a whole business that ServiceNow made is to manage tickets as if they're pets, know, forget about cattle or bees.
And AI is going to come and say, look, in the world of self-serve, you don't need tickets. You know, so I think we're having to think very differently about what to get rid of, like the way we got rid of storage back at Nutanix. And we're trying to get rid of that. But the business model piece is entirely new. I mean, it's not even subscription, which really was a version two of Nutanix.
In version one of DevRev, we have to think hard about consumption, we have to think hard about outcomes. And we have to think a lot about change management, is way bigger a problem. Because Nutanix was a technology risk, but not a market risk. The real market risk there was AWS saying, look, we're going to stream infrastructure. That was a market risk. We were taking a technology risk where the puck was, the puck was where data centers were, where VMware was, where EMC and NetApp and Cisco and HP and IBM were.
But the puck was headed towards streaming, not even owning infrastructure. This time around, we have to basically think about where the puck is headed, which is an aging population, immigration not being solved for, globalization actually shrinking. So you've got to really do more things with fewer people and fewer immigrants or whatever.
The only way you can really still improve productivity and get more efficiency is through AI. So we have to think about some problem statements that I never had to think about.
Declan Waters (21:26.354)
There's a phrase I've heard you use recently and that's work softer. What do you mean by that?
Dheeraj Pandey (21:34.35)
You know, everybody worth this salt is basically saying, we will make AI help you accelerate everything. Which means you're adding to the entropy. You know, all you're saying is I'll help you get, achieve even more entropy. and, that acceleration is not what we expect this thing to be. If anything, we want, the soft power of humans to be around slow down pause, you know, instruct AI to do bigger, better things.
And that soft power of humans is what we want to really go and tap into. It's not about losing jobs. It's about doing your job so well that now you don't have to work 80 hours. Now you can do in 60 and 50 hours, know, maybe 40 hours. Now this happened during the industrial revolution. I mean, in 1872, the legal number of hours you could work every week was a hundred hours, and the legal age of work was 12.
And then we reaped the advantages of industrial revolution, where we said, look, we got to be better as humankind to probably give a backstop at 40 hours and really 18 being that age of work. And I think we are back to 80 hours in the last 15 years because of globalization, because of the world being flatter and all that stuff, know, COVID and remote work. And so we got to really use this principle of what machines could and should do to improve human judgment and also human lives.
We feel like we also have to cancel a few things that became an end in themselves. Like, for example, service desk. You've got to cancel service desk and say, work software with answers and actions. You've got to cancel this thing called enterprise search because it's really built with a lot of wastage, you people don't even come and search for things. So when you cancel that, what do you deliver? You know, consumption based pricing model, which helps you not just search, but answers, know, which is precise answers is about structured data.
And finally actions, which is about becoming generative. So we've got to cancel a few things. We feel that that's what it means to work software. mean, even to get rid of SaaS is being too bold. We have to be respectful. I would rather we say we Work software being next to ServiceNow, being next to Salesforce.
How do we take out all the data and the context and the permissions and the workflows and run it in a new stack while being respectful of the history of SaaS in the last 20 years? That to us is working software. It's this idea of team intelligence as opposed to I'm going to get rid of my team because of AI.
Declan Waters (24:27.986)
Yeah, there's so much in there, Dheeraj, we could do another podcast to unpack a lot of what you've just said. But I want to switch gears a little bit to leadership and legacy. There are so many budding entrepreneurs.
When leaders get watershed moments wrong, is it because they're moving too fast, you think, or are they optimizing for the wrong thing? How would you kind of characterize that?
Dheeraj Pandey (25:01.943)
I mean, there's two kinds of leaders, the peacetime leaders and wartime leaders. And peacetime leaders are just really good with relationships and keeping the status quo and, you know, generating profits and squeezing more from profits. And then there's wartime leaders who are paranoid about things and they are willing to look around the corner.
The ones who actually get it wrong, and it's on both sides, by the way, there's the innovators dilemma for the peacetime folks. It's a real problem because they're extremely optimistic about the present as opposed to paranoid about what's going to happen in the near future. And then those who are actually on the war side, wartime leaders who are actually not thinking hard about the bird in hand and they constantly worried about the two in the bush.
And the bird in hand is what's going to be the infrastructure in which they can really go build bigger, better future as well. So when I look at a lot of entrepreneurs who are constantly selling and moving on, selling and moving on, as opposed to selling and staying and sticking and being able to really go through adversity together and then find the upsell opportunities. That's where the real mistakes happen.
And people who think they're wartime leaders, but they're not paranoid enough are the other mistakes that I see out there. That they're extremely optimistic. Now, if you're optimistic, you're better being on the peace side, you know.
Declan Waters (26:43.506)
What's been, would you say, the most defining watershed moment in your career? I'm thinking about it in the sense of the one that's changed how you see everything after it, right? And it can be something you've never said on a podcast or a TV show before. That's even better.
Dheeraj Pandey (27:06.541)
I think it was COVID, February, March of 2020. I had been an investor and operator, a founder, a CEO for 12 years, almost 12 years. And then I realized that whatever I had built is now a stable sort of market, but it is now peacetime sort of market, you know, where you can maintain it and milk it.
But now people want to stream everything that we were selling to our customers. They don't want to own that thing. They want to go stream that thing. So how I needed to think about that moment where people wanted to stream infrastructure going forward. Because COVID had happened, all data centers were shut down. There's no point owning data centers anymore. There was no point owning all that. And nobody could come to a physical data center, except for the ones who were building hyperscale infrastructure, know.
That to me was a watershed moment. Because now I had to figure out, there's a part of me that's an investor in this business, there's a part of me that's an operator in this business. How do I decouple my left brain and my right brain?
And think about this very differently and become an investor while I think about, you know, really diversifying out into a new problem statement, which happened to be similar in some ways about integration, about conversions, about the missing context between different departments, but at the same time, it tears you apart to make that decision. The fact that what you had built over the last 12 years is now something that you can just call yourself an investor for.
Declan Waters (28:50.576)
Yeah. What do you care about less now than you did 10 years ago? What do you say? Yeah. What do you care less about now than you did 10, 12 years ago? When you think back to those days.
Dheeraj Pandey (29:07.073)
Being liked more than being respected.
Declan Waters (29:10.386)
That's interesting. Yeah.
Dheeraj Pandey (29:12.609)
I liked being liked 10 years ago, 15 years ago. Now I like being respected. Even if it means because I think a lot of founders entrepreneurs, they're very good with, you know, design thinking. So they relate to people, their end users. And that strength is also a weakness when they have to make tough decisions.
And I think in the last 10 years, I've had to really be sort of schizophrenic about both these that, know, it's great to be liked by end users, but it's better to be respected by employees.
Declan Waters (29:54.995)
Well said. I've got one final question for you. Who are you building for now at this stage of your life?
Dheeraj Pandey (30:06.669)
You know, obviously a big part of it is what legacy do you leave for your children? Because they're all so confused about AI and in many ways they want to co-opt AI in their homework and their programming classes and things like that. In fact, they feel like they don't need to do what we did, just like we didn't do what our parents did. Like for example, log tables to actually do things rather than calculators and computers over time.
So I think we need to actually leave this world where machines are really complimentary to people because, you know, my children and my children's children, they also need that kind of a future.
We talk a lot about climate change and world peace, but there is an equally existential problem that we are creating right now with AI that we need to really reconcile with and say, you know what, it is not a zero-sum game because we're making everything to be a zero sum game as opposed to it is a creative. So I want to really work towards that future where AI is for people rather than AI is for employers or AI is for corporations.
Because if AI is for corporations, exactly what's happening right now with globalization and the sort of pushback against all the factories being in one country and all that stuff is gonna happen. So we need to really think about AI for humankind rather than AI for corporations.
Declan Waters (31:43.218)
They have an AI for humankind. What a profound way to finish the pod with Diraj. I really appreciate your time. know, having seen the first chapter of your journey up close, it was a real privilege. And now watching the second act unfold, it's so clear to me and so many others that you're still building at this really, really high I want to thank you for being open about the inflection points, not just the milestones, but also the moments that really changed you. So it's been an absolute privilege talking to you, Dheeraj. Thank you.
Dheeraj Pandey (32:19.351)
Likewise, sir, and I'm really thankful, grateful for your questions that actually created this opportunity to reveal. But I look forward to doing another one of these as and when we make work softer public.
Declan Waters (32:35.602)
Well, if you enjoyed this conversation, this is the Watershed podcast where we explore the moments that change how leaders build. Thank you so much for listening.
Dheeraj Pandey (32:45.037)
Thank you.
Declan Waters (32:50.394)
Thank you.