The $600 Billion Wake-Up Call

Hi, everyone. " I've mentioned it recently, but if you don't know, I'll say it again. I had a great time at Cisco Live.

I guess it was three weeks ago now. Something like that, yeah. It was out in Vegas.

I interviewed some amazing people, but this next gentleman somehow evaded us. Well, he was a little busy, to tell you the truth. But he was nice and kind enough to come follow up here with a "Techstrong TV" interview.

Let me introduce you to Kamal Hathi. Kamal is the SVP and GM for Splunk, which of course, is a Cisco company these days, or has been now for a couple of years. " It's great to have you on here.

My pleasure. Fantastic to be here. Looking forward to our conversation.

Thank you so much. Kamal, I always like to give people a sense of who they're listening, watching. Yeah.

I said you were the SVP and GM for Splunk, but give us a sense of your career- Yeah ... path, your journey. Yeah, absolutely.

Look, I'm at heart, a data guy, a technologist, and I've been in my role at Cisco and Splunk a little bit more than a year. I run the Splunk business all up. And we're doing a lot of great things.

Splunk is reinventing in a fantastic way with lots of impact on relevance to AI and keeping companies up and running in this crazy world we live in right now. And then before that, I was the chief technology officer at DocuSign. And there, I also helped the transformation from a signature company to an AI-first platform, driving business value with AI in documents and contracts, and all that.

And before that, I spent 22 years of my life at Microsoft. Really? And there I ran a lot of the data products.

So if you ever use something called Power BI, I helped found it, and I was the GM for the business. Worked a lot on things like Excel pivot tables, SQL Server, the database. There's a thing called Analysis Services, which does OLAP and those kinds of things.

So very much a high-scale enterprise data and value via that to customers. And based out of Redmond, Washington, 98052. Literally, I can throw a stone at the Microsoft campus.

I lived most of my career there. But that's my background. Love it.

If you need data- Love it ... then Splunk is data. For someone who's the SVP GM at Splunk, a data background is the perfect thing to have.

I remember, I guess it had to be 2005, 2006, the first time I encountered Splunk. I came back, I had co-founded a venture-backed company, and I told our board, "I saw this company. They've got the hot stuff," because they could capture just about anything.

Right. The question's going to be what do I want to capture and what do I just not want to? Because it was so revolutionary back then.

You couldn't do that. Yeah. And of course, Splunk built on that basic ability to capture everything and anything, and now provide analysis and actionable intelligence on top of it.

Right. And look, that's why Cisco paid what they paid, right? And that's why it's such an important part of the Cisco go-to-market these days.

Yeah. But Kamal, we were talking offline before we started, and I said that ability lends itself to so many different silos. A lot of people think of Splunk as a security tool.

" "Splunk's a security tool. It's a modern-day SIEM. " Probably one of the best observability tools, if not the best there is out there.

Yeah. Other people say, "Well, it's data. " And from where you sit, how do you explain it to people?

Yeah. Look, Splunk is a platform for unifying visibility across all your machine data, your telemetry, your metrics, your traces, your events. And by doing that, it enables security, it enables observability, it enables trust in emerging technologies like AI.

And the way it does it is it breaks down these silos. Because sometimes you don't know whether it's an observability problem, it's a security problem, it's an AI problem. You don't know.

And so you need something that ties it all together, and Splunk is a universal glue in an enterprise that ties all relevant data together to unlock visibility and let you know what exactly is happening. If there's an incident, Splunk enables you to know if the incident is a technical observability problem with your services, your servers, or is it a security exploit that's about to happen? Or it can help you understand whether or not this just happens to be a problem with your data pipeline.

But Splunk is uniquely positioned, is really combine end-to-end correlated, cohesive data across the entire enterprise, and via that, let you get ahead of downtime. So we deliver what we call digital resilience, and Splunk is a digital resilience company. I love that.

It's the first time I've heard that applied to it. I'm going to steal that, if you don't mind. All right.

Kamal, I wanted to pivot in a little bit here. You guys just recently released the Splunk 2026 Cost of Downtime Report. Right.

And look, I've got 25-plus years in cyber myself, venture-backed startups and so forth. We've all known that downtime is expensive. Yeah.

Quantifying it, though, is usually a very eye-opening experience when you see just how expensive it is. Let's examine, if you will, the report and some of the findings. And let's start with, I always like to ask, what was the idea behind the report?

Did we have sort of a preconceived notion we were looking to explore? What drove this year's report? We've been doing this, as you know, for a while.

Oh, yeah. And the one thing that is important is for our customers to understand that when they talk about downtime, it's not a single-dimensional problem. They think it's an IT problem half the times.

Well, many times it's actually not an IT, it's a security problem. And the implications, the impact of that is exponentially higher than they sometimes imagine. And so I think really surfacing that was critical to help our customers understand that you cannot look at downtime as a monochromatic, single-dimensional issue.

And really understand that it is across multiple facets. And that cost that you mentioned actually is way more than anybody even tries to understand. So really surfacing is important, and we had that driver call to action, especially in this world where now everything is changing with AI.

We have Mythos and all these things going on. In that world, this awareness is critical. Agreed.

You mentioned Mythos. I had a couple conversations again today about it, your insecurity and just- Yeah ... what's going on.

But we're just at the tip of the iceberg of what Mythos might wreak here. And when we say Mythos, I hate to put Anthropic under the spotlight. Yeah.

Because let's face it, the latest version of ChatGPT Security, whatever it's called, has similar type of capability. " Yeah. So I try not to make Anthropic the boogeyman.

But the important piece of it is vulnerability discovery is going to become like water, commodity. What you do about it is where the money lies, right? Yeah.

And to me, man, that's right in your sweet spot, right in your wheelhouse. Because now that I can take all that vulnerability data now, if I'm Splunk, I could use it to create- Yeah ... actionable intelligence, what we need to do.

I'd like to go turn back to the report, though, if it's okay, Kumar. Sure. Of course.

There's always three headlines in these reports, right? What are the three key findings that you think people should know about? Yeah.

So I think first of all, just awareness on the fact that the numbers are staggering. I don't know if you remember from the report off top of your head, but the aggregate cost to the industry is about $600 billion. It's 50% more than just two years ago.

And to give it another way of looking at it, it's $15,000 a minute. And a single incident sometimes can have multiple percentage point declines in shareholder value. Oh, absolutely.

It's a big deal. That's the one thing that really, to me, it strikes me directly. The other things that are really important is just around the fact that besides the financial toll, there's a direct connection to security.

People think about downtime as an IT problem, and that's not true. Things like ransomware have just gone up tremendously, nearly tripled since 2024. And people keep misclassifying these things as IT issues.

And when they're actually a security problem, and if you don't get ahead of these things, it can really become a really big impact on your brand and your company. Sure. Regulatory pressure.

81% of tech leaders have talked about regulatory fines. Average about $51 million per organization just by the fact that you had these kinds of incidents and outages. And the last thing I mentioned already is around brand health.

20% of marketing professionals that we talked to said that it takes over a quarter to recover for the brand. I think it's more than that. So these are some things.

There are other interesting, important things that are really important is around the need to have an end-to-end view. Break down the silos between various types of data, various types of incidents, observability, security, et cetera, and the need to automate. Automation has become critical, especially in the age of AI.

You mentioned the threats that are coming in that are causing these downtimes. That is machine speed acting at you. You have to respond at machine speed.

You have to fight fire with fire. It's automation is critical to being successful at this. And so there are some things we found we can talk about more.

Mm-hmm. But from a very high-level point of view, these are some of the high-level items. Good.

At $15,000 a minute, every minute counts, right? Yeah. So when we talk about automation, because one of the things we're seeing with all this AI is I call it AI scale.

Yeah. Right? It's velocity, so speed, and scale.

That's AI scale. And so when we're talking things like $15,000 every minute, the speed of your response kind of outstrips the speed of humans that we- Right ... there's just no ways around it.

And, it gets up there in things. And the thing to remember is, well, look, when you're talking about something like $600 billion annually, there's multiple outages, there's multiple downtimes going on simultaneously. Yes.

So it's 15 grand here, it's 15 grand there every minute, right? That meter's running like crazy. Of course, $600 billion.

It's just another AI data center or two, right? In today's world, the way we're building them, but it's real money. It's real money and it's real brand awareness.

" Right? "And within a quarter, our stock will bounce back. " Well, that changed a couple of years ago, right?

With the Target breach and then Equifax and- Yeah ... take your pick of some of these other high-profile ones. Look, C-level people have lost their jobs, right?

Stocks have taken significantly longer to rebound. Absolutely. And that black eye of, didn't they have a breach, that big breach a couple of years ago?

It sticks around longer than it ever did. So these are all real issues. Kamal, let me ask you another question.

" Anything in the report here that sort of wasn't on your bingo card? I think the one thing that, not necessarily being on the bingo card, but the importance and the impact of AI is something that's evolving. So from a point of view of the most successful companies who are able to handle these kinds of problems are, I would say, AI experts in some senses.

They are organizations that we call them as AI workflow and triage experts. So they're 74% more likely to avoid these public data breaches and disclosures. And it's not just about breaches and disclosures.

It's also about how you're handling the exposure of sort of vibe coding and other kinds of things in inside your organization. If you don't have control, if you don't have trust in AI, it really can get, be a problem. So you mentioned machine speed and scale.

The third one is trust. Yeah. You must have trust in AI if AI is going to be driving the machinery.

So that's the one thing that came up a lot. And because agents are here to stay, and you need a very- Absolutely ... clear strategy on how you're going to go and manage these systems.

Because the flip of it is, if you don't, you can start actually having impact from a mismanaged AI strategy. And so this is really important. And the last piece there is just the whole AI budgets, they are sort of out of control.

We talk about this notion of tokenomics. Yeah. And when you think about your AI trust and governance overall, this notion of knowing what is going on, what is the agent doing?

Are they doing what they're supposed to do? And do you have a very clear control over their cost? So these things are bubbling up more and more now, and I think, at least for me, when I talk to customers, almost on a daily basis, the importance of having trust in AI is becoming a critical problem.

Agreed. There's that old adage, though, right? Trust has to be earned.

That's right. And I think that for a lot of organizations, we're still in that earning trust- Yeah ... stage.

But interestingly, I saw a study, I think we've talked about it on Techstrong Gang today. For all the talk about the token maxing and the token economics, and is it working? Does it affect the bottom line?

Organizations that are embracing AI and are using it saw something like 160, I forget it was 120 or 160%, a large percentage productivity gain. Yes. And it reminds me of sort of the DORA, I don't know if you're familiar with DevOps and the DORA guidelines, right?

Right. Where high-performing IT teams. Look, the company's stock did better, the business- Yeah ...

did better because their IT teams did better. I think we're going to see a similar kind of relationship. Companies that embrace AI and trust it and use it well, and again, that trust was earned because they've gotten familiar with it and comfortable.

I think they're going to outperform, and it's going to create a widening gap of haves and have-nots, right? Companies that embrace it, companies that don't. And, that may well determine the winners and losers.

Absolutely. There's no question. I think AI is no longer an interesting new thing.

Right. AI is a key part of our daily lives, and we talked about automation. We talked about the ability to stay ahead of threats.

We talked about controlling and governing these tools. And you need to have a very intentional strategy for it. And that allows us, as users of AI, to get the most out of it.

And that's what our report shows, is that 74% of organizations who are able to really get value out of this are these AI experts, who we use. When I say experts, the word is more loosely used, but these are folks who have intentionally looked at AI. They have a human to agent strategy, identity, how does it all fit in, observability of the problem, token economics, infrastructure.

All of these things for AI are new, and you really need to make sure that you are extremely focused on maximizing the value, and obviously leveraging the productivity and speed that you get out of it. Saw another report on AI jobs today. Even if you just use AI a couple of times a week, not for everything you do- Yeah ...

but a couple of times a week, you're so many more times likely to have a job today- Yep ... than people who don't use it at all. And it's in there.

Kamal, we're over time. I apologize that- Yeah, no ... we went so long.

But for people who want to get more information or take a deeper dive into this year's Cost of Downtime Report, what's the best place they should go? com. Be right there off the front page.

Yep. It's the site for the report. It's the site for what we do.

All the things I talked about, trust in AI, agentic solutions for the SOC, security operation, agentic SRE solutions. com. Love it.

" I hope you won't be a stranger. Come by any time and chat. Of course.

Love to have you. It was a pleasure. Thank you so much.

Thank you. Thank you. " We're going to take a break.

We'll be back with more. Stay tuned.