May 10, 2026 · 39m · neon-show
How Poojan Kumar Used “Discontinuity” to Sell Two Startups | Nutanix ($12B) & Commvault ($4B)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Serial entrepreneur Poojan Kumar joins Siddharth Ahluwalia to discuss how recognizing technological discontinuities enabled him to build and sell PernixData to Nutanix and Clumio to Commvault. He shares comprehensive insights on cloud cyber resilience, AI data governance, bottom-up enterprise sales, and founder leadership.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Siddhartha as informed peer 4.3 Guest teaching 5.8 Guest disagreement 0.0 Siddhartha pushing back 0.3
speaking balance: gold is Siddhartha, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 15:43 Nuancing enterprise agent deployment skepticism
Rather than accepting a binary premise that enterprises might not be deploying agents, Poojan clarifies that adoption is heavily underway in verticalized SaaS domains like customer support and legal workflows.
Hardest push from Siddhartha ▶ 15:37 Questioning real-world enterprise agent rolloutsHost interrupts the theoretical security discussion to directly challenge whether enterprises are actively deploying autonomous agents today or holding back.
Biggest teaching moment ▶ 23:20 Clarifying eleven nines durability vs accidental deletionPoojan educates the audience and host on a common misconception, explaining that cloud providers' 11 nines of durability won't protect data against programmatic deletions or ransomware.
Siddhartha holds their own ▶ 27:28 Synthesizing multi-cloud compliance and cost inflationHost demonstrates strong market understanding by showing how attempting to optimize costs across multiple clouds usually triples risk, security overhead, and compliance burdens.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Siddhartha as informed peer | Guest teaching | Guest disagreement | Siddhartha pushing back | Why |
|---|---|---|---|---|---|---|
| Episode Preview: Identifying Sector Discontinuities and Enterprise AI Questions | 4 | 5 | 0 | 0 | Host introduces Poojan Kumar and frames his career trajectory across Oracle Exadata, PernixData, and Clumio. Poojan provides context on spotting sector discontinuities and moving from internal corporate projects to founder journeys. | |
| Founding PernixData and Capitalizing on the Flash Storage Discontinuity | 4 | 6 | 0 | 1 | Host probes PernixData's revenue numbers, capital efficiency, and timeline to acquisition by Nutanix. Poojan details the architectural shift from spinning hard disks to flash in server virtualization. | |
| The Public Cloud Discontinuity and Founding Clumio | 5 | 6 | 0 | 0 | Host inquires about the data backup landscape and market positioning. Poojan contrasts on-prem legacy backup with cloud-native cyber resilience architectures for hyperscaler workloads. | |
| Rethinking Data and Infrastructure for the AI Agent Era | 4 | 6 | 0 | 0 | Host asks how a storage and virtualization founder should build in an AI-dominated ecosystem. Poojan explains how consumption models and agentic workflows are reshaping data protection. | |
| Explaining the Cloud-Native Data Stack and Resilience | 3 | 7 | 0 | 0 | Host asks Poojan to explain the AI storage stack in layman terms. Poojan breaks down serverless architectures, S3 object storage primitives, and ephemeral Terraform deployments. | |
| AI Security Challenges and Enterprise Adoption Patterns | 5 | 6 | 0 | 2 | Host pushes on whether enterprises are actually deploying autonomous agents or still hesitating. Poojan notes widespread vertical adoption in legal and support while outlining MCP server data flow security risks. | |
| Infrastructure Shift: On-Premises to Cloud to AI-Native Development | 4 | 6 | 0 | 1 | Host compares PernixData's 25M scale with Clumio's 25M scale. Poojan explains the profound shift from enterprise field sales to consumption-oriented bottom-up product adoption. | |
| Developer-Driven Products and the Success of Tools Like Claude | 6 | 5 | 0 | 0 | Host brings up developer preferences comparing Claude against Copilot, Cursor, and Codex. Poojan agrees and elaborates on cloud backup imperatives like ransomware and programmatic accidental deletions. | |
| The Reality and Risks of Multi-Cloud Architecture | 6 | 5 | 0 | 0 | Host references Poojan's past remarks about fragmented clouds being silent killers and adds commentary on hidden compliance and security costs. Poojan validates the perspective while outlining enterprise multi-cloud realities. | |
| Mastering Capital Raising and the M&A Playbook | 4 | 6 | 0 | 0 | Host asks for tactical advice for first-time founders regarding fundraising and M&A relationship building. Poojan compares closing an M&A deal to selling a complex enterprise contract based on long-term trust. | |
| Two Different Exit Journeys: Nutanix vs. Commvault | 4 | 6 | 0 | 0 | Host asks about the different relationship dynamics behind the Nutanix and Commvault exits and closing eight-figure contracts. Poojan explains pre-existing founder connections versus building new trust channels with shared Slack rooms. | |
| From Engineering Leader to Startup Sales Execution | 3 | 6 | 0 | 0 | Host inquires how an engineering-focused technical founder mastered enterprise go-to-market. Poojan highlights the importance of humility and relying on veteran Silicon Valley advisory networks. |
Statements from this episode (21)
Poojan Kumar: Oracle Exadata reached $1B revenue as a fast-growing product
“It was one of the fastest growing products, getting to a billion dollars of revenue at Oracle while I was there.”
Poojan Kumar: Enterprise startups must be built on major sector discontinuities
“Whenever I, you know, look at starting something, at least in the enterprise, you have to look at what's the big discontinuity happening in the sector, right?”
Poojan Kumar: PernixData reached $25M in revenue before its Nutanix acquisition
“Yeah, it was about, ah, you know, twenty-five million of revenue.
Independently, yes.
Ah, when we got acquired, we're about twenty-five million of revenue.”
Poojan Kumar: PernixData had close to 900 paying customers at acquisition
“We had, you know, operations, ah, you know, in multiple countries, you know, around the globe, and, ah, close to 900, you know, paying customers when we got acquired.”
Poojan Kumar: PernixData went from founding to acquisition in 4.5 years
“Four and a half years from beginning to acquisition.”
Kumar: Public cloud backup requires fundamentally different architecture than on-prem
“And the scale and the architecture needed to go and, you know, provide backup and cyber resilience for them was very different than an architecture needed to go backup, you know, on-prem or even backup SaaS applications, right?”
Kumar: Clumio followed Snowflake's self-serve consumption-oriented platform model
“Back in the day you would essentially, you know, say, okay, how do I build a platform which is very easy to use, right? Where somebody can go, come to your website and start using it and even start consuming it, which is, you know, how Snowflake did it, which …”
Kumar: Every enterprise security workflow is being rethought around AI agents
“Every, every use case in security also is getting, you know, rethought, like, how can I build a, you know maybe an agent, right, that can do this automatically versus relying on how it was being done before. So every workflow is being, being reinvented and eve…”
Kumar: Agentic AI apps in cloud will require increased data protection
“These agentic applications in the cloud will also need, you know, protection, right? And if at all, we'll need more of it because the chances of things going wrong could be higher, right? You can imagine, right? With all the stuff happening in an agentic world…”
Kumar: Modern AI data stacks replace VMs with distributed serverless architectures
“Gone are the days that you're going to run, you know, applications sitting in big virtual machines and, you know, attaching storage to it and so on and so forth, right? You know, you're basically going to build an architecture that can, you know, scale up and …”
Kumar: No vendor has yet holistically solved AI agent data lineage
“I think that is very early days. I don't think anybody has really tackled this holistically yet, but this is I think important.”
Kumar: Enterprises actively adopt AI agents in support and legal verticals
“So, you know, for example, support is, is, is an interesting vertical, right? Which is a lot of stuff happening within support where, you know, the ticket deflection and automatic resolution of tickets, these are things that whether somebody builds it themselv…”
Poojan Kumar: Successful AI startups scale through bottom-up adoption, not top-down sales
“The ground up adoption is how you have to target it. You have to essentially think about how are you going to essentially get this thing in the hands of your end user, make it very easy and essentially make sure that it's something that is, is, is loved by the…”
Siddharth Ahluwalia claims developers ignore enterprise-purchased Copilot, Codex, and Cursor
“Cursor Codex by OpenAI and Copilot by Microsoft, they have all sold to the enterprises. But developers are not using them.”
Kumar: Claude Code's Grassroots Adoption Replicates AWS Winning Cloud Architects
“It's like, ah, it's like the equivalent of the developer for, A cloud code is exactly how, you know, you can say, you know it was a cloud architect for AWS, right? Or a DevOps person in the AWS world, right? So on and so forth. It's like, if I think there's a …”
Kumar: Cloud Durability Guarantees Do Not Protect Against Accidental Data Deletion
“The cloud vendor is making sure that I have 11 nines of durability. That's awesome. But that doesn't prevent from a deletion happening. Because if I ask a cloud vendor to delete something, guess what? The cloud vendor will delete it. Because you asked me to de…”
Kumar: Ransomware Attacks Both Primary Data and Backups
“Obviously a typical ransomware, you know, is going to essentially go and, you know, attack your primary, but they also will attack your backup. If there is a backup, because that is your resort to go and recover, right? So then, you know, that became an import…”
Poojan Kumar admits raising far more capital than his startups actually needed
“I think, you know, I, I've made my fair share of mistakes there too. When you raised, you know, way more capital than what you needed, right?”
Kumar: Startup M&A works like closing a long-term enterprise deal
“When somebody is you know, is looking to buy a company at the end of the day, you know, they're buying, you know a platform, a company, to essentially make it part of the bigger entity and really grow it, right? And so I think it's a large deal at some level, …”
Poojan Kumar: Connecting Engineers to Customers via Slack Drives Contract Expansion
“In the case of large customers that have scaled with us, even though it has started small, one thing that we have done is we've kept close. Like we would create a Slack channel between The customer and the folks on the customer side, the users on the customer …”
Kumar: PernixData advisors and VCs interviewed early sales reps and leaders
“My advisors at Pernix, at least like literally would help me even in interviewing, you know, our head of sales in, in, and before even we had a head of sales, you know, interviewing with the reps and in, in some case, you know, VCs who have essentially neither…”