Turning Data from a Liability Into an Asset

Daniel Newman:
Hey everyone. Welcome back to the Six Five Summit 2024. I’m Daniel Newman, CEO of the Futurum Group. Excited for this day one data observability track opener. It’s someone that we have had more than a few times on the Six Five here at the summit. Also had conversations throughout the year.

Someone I consider a friend and someone that I believe is doing incredible things there in the valley. Sanjay Poonen, CEO of Cohesity. Welcome back to the summit. So happy to have you here. Appreciate you opening up this track.

Sanjay Poonen:
Daniel, thank you. Congratulations on all the success you’re having on the Six Five, and an honor to be with you on this track of the summit.

Daniel Newman:
I want to get you on the record saying to Pat and I, one of these days that you guys are killing it because then I can use the meme just like we did with Michael when we got him to say that. We didn’t even ask him. We didn’t pay him to. No, really appreciate it. And same to you. Over the last several months, you have been making quite a splash.

You have announced what I believe and I think you can confirm this is the largest deal in the data protection space when you complete the tie-up between Cohesity and Veritas. Very exciting. We’re going to have a great discussion today about data, turning it into an asset.

But hey, before that, I just teased it out for everyone out there that maybe hasn’t heard, Sanjay and his team announced a really exciting deal. Give us the quick one, two minute update, kind of where things are at. I know not everything can be public, but I’m sure people are excited. This was a really big deal in your space.

Sanjay Poonen:
Yeah, no, listen, it is the biggest deal. It’s a $7 billion transaction that combines number eight and number three. Technically, the way it’s working is we’re acquiring the data protection business from Veritas. They’re spinning out what’s left over. The total transaction size is $7 billion. It will create a pro-former 1.6, but going into next year about a $2 billion company, 27% free cash flow margins, about teens growth, maybe even higher.

So rule of 40 type of company. They’re a leader number one, in market share. So certainly a very profitable rule of 40 company and hopefully we do well. We can take the company public after we’ve had some time to execute. The integration planning is going well. We’re getting many of the regulatory approvals we’re expecting on time and we’ve said we expect it to close later this year, hopefully by the fall of this year. We’re meeting some of the people on the other side and planning out ways in which we’re going to organize ourselves.

Getting the playbooks already, it’s a little bit like the preseason for whatever your favorite sport is, football, let’s say, and in the summer months you prepare for the fall. I think it’s similar to that. We’re getting very prepared, we’re very excited. And then of course the most exciting thing for me is planning out the innovation that we have planned with an engineering team that’s going to be two exercise of any other modern competitor.

And then also talking to a lot of customers. I’ve had a chance to talk to many of our largest customers, many of their largest customers, just to get a sense and share with them our perspective thoughts of roadmaps and the feedback, Daniel has been delightful. So super excited, cautiously optimistic, and we want to create an iconic, obviously we’ll be in a good path to two billion for sure, but we want to create an iconic fiber revenue company.

And when you do that, you’re creating a company like, I don’t know, Snowflake meets Palo Alto or a CrowdStrike meets Databricks. These are special data and security companies at the junction of three tech vectors, multi-cloud, security and AI.

Daniel Newman:
Well, listen, it would be foolish for anyone to bet against your ability to do that. I would also tell you that when you hit five billion, you’re going to want to be the next 10 billion company. And I know that leaders like you rarely ever settle on a result, but at the size you’re coming in creatively between the combination of the companies, it’s a very ambitious growth path.

But let’s talk today about some of the innovations, some of the technologies and some of the trends that are driving us in that direction. I mean, look, AI is the trend. Off the back of that, everything from your data hygiene to your resiliency to data management, and of course security are all big focus.

And then we’ve got CIOs, CISOs with flat budgets, new pressures. How do they actually execute against the potential of AI? You’ve got CFOs raining down on them saying you can’t do all the things. I don’t have the money to do it all. You’ve got the bite between CapEx, OpEx, all of these things. Let’s start there really quickly, like data, AI, security, these are big opportunities, big challenges for a company. What are you seeing as the biggest challenges of all of these converging technologies and innovations and what’s the impact that you expect this to have?

Sanjay Poonen:
Yeah, as I talked to, I mean one of the things I’ve been very blessed to work for some great companies over the last 20, 30 years, but most of that has been in the data slash analytics space. We didn’t call it AI and certainly not generative AI, but many of the things we were trying to solve in analytics was predictive analytics and optimization. My computer science undergrad work and my master’s work was all in that sort of data.

So I’m a data junkie. I love data. I love analytics, I love algorithms that optimize. And so all of these, but in the early days of the ’90s when we were doing some of this work, much of the work on AI was expert systems. We didn’t have the compute resources that a GPU would afford. I mean, I didn’t even know who NVIDIA was in ’90, they started in 1993.

But at the end of the day we were still thinking about this as business intelligence, data warehousing, and most of my years at SAP working for phenomenal leaders like Bill McDermott and creating the analytics business of SAP, I long for this type of day, which is now the world of AI at generative AI and here we are.

I think through my years at VM, also working for phenomenal, Pat Gelsinger building the security business there. Security is very much about being able to operate like doctors do on when you think about disease research, the way in which we’re going to solve cancer, Alzheimer’s is large amounts of data and genome sequencing and being able to run AI algorithms to better design life sciences drugs.

Security is the same way. The more data you have and the more you can analyze threat vectors, the more likely you’re to protect people. So it comes back to a data problem ultimately. So having learned a lot in the last 15 years about security, having been a lifetime lover of data, I look at now this sort of junction of where Cohesity stands and I described it as sort of a Snowflake meets Palo Alto type of opportunity.

If you were to tell me there was three vectors coming together, multi-cloud, security and AI, and at the core of it, large amounts of data to let you do more special things. One of the things we’ve set ourselves a mission at the company to secure, manage and provide insights into the world’s data and to help our customers with business resilience.

So the second statement is kind of an outcome of the first, but coming together with Cohesity, we will have almost 90% of the Fortune 100, almost 70% of the global 500 and hundreds of exabytes on our platform that we manage, I mean almost a 100X bigger than almost any of our competitors.

That puts us in a very strong position to say, what are we going to help our customers do with that data? So think of that as their goal. There’s two things you can do. One is play defense to protect that data from bad guys. That’s paramount. You can’t play offense and you don’t start with defense. There’s no point trying to mine that data if it’s getting corrupted and exfiltrated and stolen.

But once you protected that data, which is the security element of everything thing we’re doing from ransomware and so on and so forth, let’s also help our customers play offense which is search, discover, summarize, that’s what RAG and some of the things we’re doing, retrieval of the generation and generate AI with NVIDIA’s investment in us. And I think a combination of this defense offense play on data, many of our pure competitors we respect all of them, are playing defense like us very well, but no one’s really been able to play offense on that data goldmine.

And the breakthrough came for us, which we’ll discuss I’m sure in our discussion today when we started working with Microsoft and NVIDIA and really became the first company to build out this RAG capability on top of backup data, patented the idea, released Gaia, had Jensen talk about it in his keynote, you were there covering the event. So lots of exciting things, but it all comes back to that confluence of multi-cloud data protection, security and AI.

Daniel Newman:
And if time allots, I want to come back and maybe ask you a little bit about the convergence of these different architectures in the tech stack, storage and data protection and of course traditional sort of ISP software and how kind of different are approaching all this data is available and can be used because I’m hearing obviously storage is being sort of reinvented right now.

Data protection is being reinvented a little bit right now. Of course networking is being completely … and this is opening doors to you, the fact that you could patent something, put RAG on top of something that was historically merely defense and start to actually offer something different, it’s significant Sanjay. And it opens up a new TAM to you.

It also makes you in some cases a competitor in new spaces because of what people can do with your data, which every company is doing now. I like your offense defense analogy. I think that’s really true. It’s kind of the overall gen AI state. Maybe you could think about how you’d compare my thought to yours is I keep calling it prune to growth.

So what we’re seeing right now is most companies are doing the where do I find efficiencies with AI? How do I get cost out? But as we all know, the four trillion of expected economic growth is going to come from the grow parts of the equation. But before we can grow, we have to protect, you said defense, offense. So we prune, we get efficient, we get lean and we get secure.

One of the hottest topics in the industry right now is keeping customer data safe. It’s keeping your proprietary data safe. That’s the data that’s going to be useful in gen AI. So you’ve got to secure, you’ve got to manage, and then you’ve got to deliver insights. But the industry’s evolving.

I’d love to kind of understand how cyber threats are playing a role in all this because I feel like the defense, you said people were doing it well, but I don’t feel we talk about it a lot. I feel like cyber people are like cyber, I don’t know, just sign me up and give me the gen AI. So where are we at?

Sanjay Poonen:
Yeah, I think it’s a true point. I think one other way of thinking about the defense, offense analogy is I often lose, this is the brake and the gas pedal of a car. Security is like the brake of the car. It keeps you safe, it allows you to stop in time. It protects you from an accident. You absolutely need a brake in your car. I would hate to be driving a car with no brake at all in it.

But a car with only a brake never leaves the parking lot. Every now and then you want to let loose and be able to get that gas pedal, gen AI and AI capabilities and mine and allow you to get insights is that accelerator. So that’s another way of thinking about it. I think, listen, we study every cyber, there’s a number of very seasoned security leaders inside the company, including yours truly.

And we have Kevin Mandia on our board. We talk a lot in our board meetings to every meeting about every cyber attack. I mean I’ve been following this. Good friends with Jenny Easterly who was the head of CISA in the past. I did a lot of this including in front of Congress at VMware. We study every attack, SolarWinds to growing your pipeline, to what happened recently at Prudential. I mean there’s a variety of these, MGM and we detect because the entire industry is learning every time this happens. It’s a lot like Covid-19.

Any attack of that kind becomes like a major disease. You want to study how it happened, what happened, and protect yourself the next time. And the good news is in this industry, many of us, including some of our competitors, we talk a lot about how do we protect collectively the industry. It’s a village.

And quite frankly, when we get calls from our customers or prospects to help us, we don’t go and their house is on fire, I mean often when they’ve just had a situation, we don’t go in trying to peddle our products. We go in like a doctor trying to ensure that they can recover very quickly, even if it’s not with our products, but give them advice from what we’ve seen.

The more that I think the industry becomes consultative in helping people protect themselves from the next time that happens or saying, listen, the moment MGM happened, everybody in the hospitality industry called us. Why? Because they’re worried about this happening hitting every other hotels. Oil and gas, after Colonial pipeline.

We had a surge of oil and gas customers interested in how do we protect themselves from the next thing not happening to. Prudential, protection services. So we study every one of those and I think no offense can start without a very strong. So when we think about data, we think about it like an iceberg.

And the top of the iceberg is primary data, your hot data that’s in Oracle or in Snowflake and that you query actively or in OneDrive, whatever, structured or unstructured data. And then as it ages backup, snapshot, vaulted archive data goes to the bottom of that iceberg. It’s pointless being able to ask questions about the bottom of the iceberg if it’s melted away or stolen. Okay.

That’s in essence. And every ransomware attack right now is on secondary data, is on backup data because it’s basically an index time series of everything you’ve had in the past. And it’s much easier for the bad guys to corrupt your secondary data, exfiltrate it, maybe do double or triple extortion ransomware scenarios. But once you’ve protected yourself, and that’s not difficult to do in a way where your recovery is fast and efficient, you’ve got to move to playing offense.

And the challenge in this industry, I mean the data protection is the traditional industry has been thinking about it as a storage problem first, not as a security problem and not as an AI problem. And the AI opportunity now opens this up. And when I talk to people who’ve been in this industry, the data protection industry longer than me, they acknowledge it. They just give excuses for why. Well, it’s because it was too hard to do it.

You had to uncompress the data, rehydrate it, but no one can disagree that generative AI and some of the techniques that GPUs and RAG bring you don’t create a mind-blowing opportunity for us. And that’s what we want to focus on. So time will tell, but we go in now with some very good scenarios that we talk to customers. Almost every one of our customers has tens of thousands, hundreds of thousands or a million PDF documents sitting somewhere either in their primary stores or their secondary backup stores.

And I tell people, listen, what would you pay an intern to summarize a thousand of those video documents to something you’re trying to find in that proverbial needle in the haystack? They’d say, listen, easily a few hundred thousand. Well what if you added robot that could do that for you? That’s Gaia. So when we show them those use cases, get them to see the demo it’s being, and then what we typically do is we’re bringing video with us.

I love, I mean, NVIDIA people really think about AI by vertical industries very easily. And the moment Jensen not just talked about us on stage for the 20 seconds of our fame, but then also invest in our company, we’re able to bring their people into our sales calls with customers and it just makes us look a lot better.

I mean, I’ve always believed that Isaac Newton statement, right? You see clearly because you stand on the shoulders of giants and we want to do as many sales calls with the biggest in security, Palo Alto, CrowdStrike, Microsoft, Zscaler and the biggest in AI example, NVIDIA, Microsoft.

Daniel Newman:
So Sanjay, we’ve got about five minutes and I’ve got two big questions for you. So I don’t know how well you can do these. So I’m going to give them to you. I’m going to avoid my commentary, which is hard for me. And I’m going to ask you these because I want to get them both in. One, is you talked about this RAG capability, I really want you to talk about what you are doing there because secondary data traditionally hasn’t had intelligence. So why do it and what are you doing?

Sanjay Poonen:
Yeah, I’ll make these very quick so you can get both your questions. It was compressed and it was very hard for you to do because you had to rehydrate it, uncompressed. Think of it by searching inside a zip file. That’s probably the best example everyone could understand. You unzip the file and then get insights into it. Well, with what we’ve been able to solve, we had an index ready data platform.

You attach the vector database and a LLM to it, that’s the essence of what a rag solution is. And the architecture is beautiful. It’s optimized by some of our best engineers. That’s what we patented. So the beauty of it is you can then type a question, that goes then and searches all those millions of PDF documents and finds a summary.

So good example, maybe all of your policy inside HR, one of the things that we’re piloting inside our company is people want to search what’s their maternity, paternity leave policy. And it’s all in a bunch of documents. It could be either in primary stores or in secondary backup stores.

You type that query and it comes back with a nice beautiful summary of everything. The summarization is done by an LLM, but the building of that vector database is done inside the elements of the RAG solution. So it’s a beautiful architecture.

Daniel Newman:
Yeah, it’s very, very interesting. And it obviously is a data gold mine. And so RAG is one very well understood and very financially achievable way to do AI with your data and of course secondary data. All right, last question and Sanjay, I appreciate you so much taking this opening spot for our data observability track, but you like to talk about the five S’s that customers should consider when building a strategy to secure, protect and provide insights on their most valuable asset, data. What are the five S’s? What do they mean? Take us home.

Sanjay Poonen:
Yes. And thankfully one of the S’s is not Sanjay. The five S’s are the following, we’ll focus on the five that are important and broad. Number one, speed. You want something, it’s all about speed of cyber recovery. Number two, scale. Can it handle tens of petabytes, hundreds of petabytes, exabytes eventually because the data’s big. Number three, security. How secure is this not just with, I mean reactive needs to be zero trust all across the board.

Simplicity, it has to be a beautiful user experience. We call it Google’s, sorry, consumer simple enterprise secure. And number five, it needs to be smart, AI. So those are the five is speed, scale, security, simplicity, smart. And I think when you put those together, those are often the recipe of what we tell our customers you should think through.

Those are often the reasons customers pick us at Cohesity because of those five S’s that we thrive in. And it’s a good framework that we educate our go-to-market and product people as they talk to customer.

Daniel Newman:
That was fast though too. You hit all of them and you hit them right on the nail, Sanjay. So listen, I want to thank you so much for taking the time. I want to congratulate you. I can’t wait to see the coming together of Cohesity and Veritas and what you do with that because of course the proof will be in the pudding.

So after you get this deal put together and you deliver on these innovations and you take it to market, I can’t wait to get you back here on the Six Five talking to me, talking to Pat, talking about the vision, talk about the future. Sanjay Poonen, CEO of Cohesity, thanks for opening up this track at the Six Five Summit.

Sanjay Poonen:
Thank you Daniel. Appreciate it and always great to talk to you on this side. I wish you all the best.

Daniel Newman:
Everyone out there, we appreciate you tuning in. Stay with us. We’ve got so much here in the data observability track. Come back, stick with us, more to come. Studio, sending it back to you.