It sounds like you're saying the SAS Apocalypse might be real, but it's not real for Octa.
What people miss is that the pie is getting much much larger, because everyone's trying to sell something. So when you hear big company CEOs say AI is writing 90% of our code, they're trying to sell something.
We're doing introspection here. I told you to be emotional.
This is Decoder. Decoder is just therapy for me personally. At this point you can tell what my problems are, right?
That's right. You're casting them out amongst the...
Hello and welcome to Decoder. I'm Neil Patel, editor in chief of The Verge. Today I'm talking with Todd McKinnon, co-founder and CEO of Okta. Okta is a platform for managing security and identity across apps. It's a $14 billion market cap company. But SaaS companies like Okta are under pressure in the age of AI. This is the so-called SaaS apocalypse. Todd recently said he was paranoid about it on Okta's earnings call. Okta CEO Todd McKinnon, welcome to Decoder.
Thank you for having me, Neil. I feel like a real theme of Decoder lately is me being emotional about the nature of software in 2026. I can't think of anyone better to do it with than you.
Would you like me to soothe your emotions or not? I'm going to start with your emotions. A few weeks ago on Okta's earnings call, you said you were paranoid about the SaaS apocalypse. Talk about where you are on that.
I'm really good at talking about my feelings for massive groups of people.
Well, you did. Here we go. You said, 'We are paranoid and we're making sure we're using all the latest technologies to have something resilient and secure.' That's a big thing to say. Talk about where you are in the SaaS apocalypse.
Let's start with me personality-wise. I'm challenge-driven. What I see is a huge challenge and opportunity. The pie for technology is expanding greatly with AI. This is bigger than cloud computing. Okta has had success: $3 billion in revenue, growing over 10%, 20,000 customers. The opportunity is massive. It's daunting because it would be great if things didn't change, but the prize is huge.
You've talked about the total addressable market growing. I want to come back to the SaaS apocalypse. For a run-of-the-mill productivity tool, I understand the argument. But Okta seems more insulated because of identity and security. What specifically has you paranoid about agentic software and customers building their own tools?
I want to be run-of-the-mill.
Right, but they're all fine. You can replace one project tracking tool with another. But Okta is different. What is the mechanism that might be a threat to Okta?
If you look at what these tools can do, it's amazing. I grew up as a software engineer, and that world is being revolutionized. If you don't question how you've built your company, you're naive. But Okta is well positioned. You can build the features, but you have to connect it to everything – thousands of applications. That integration and mission-critical reliability takes years. Also, there's a reputational aspect: what cyber company do you trust? People buying cyber tools have to answer to their boards. So security and infrastructure software is different from other app categories.
Making sure it works 100% of the time takes years and years. Is there also a reputational thing?
What are you going to trust? The proven solution or something your team just cooked up? Infrastructure and cyber software are well insulated from people vibe coding it themselves because of brand and trust.
There's a bit of 'no one ever got fired for picking IBM' in there. And more cynically, you want a vendor rich enough to sue if something goes wrong.
Or the glass half full would be can support me.
Yeah. Your job is to have the glass half full. I'm trying to connect the dots between insulation and healthy paranoia. What is the mechanism that leads you to say we have to be vigilant? Is it cheaper competitors or internal teams building their own identity solutions?
I compartmentalize it in two areas. First, as CEO, the most important job is strategy: which markets to be in and how to win. There's a big new emerging market: AI agents need to log into stuff and need a system to keep track of them. That's a massive new market. Second, how we execute: the technical shift is significant. The organization has to turn up the change quotient from 20% to at least 60/40. Give teams freedom to experiment, but also be skeptical of what you see online – everyone's trying to sell something. When big company CEOs say AI writes 90% of their code, they're selling something. You have to take that with a grain of salt.
Tell me about the change. It sounds like you think there will be a universe of agents that need to be permissioned, and Okta should focus on that. You're not so worried about people vibe coding their own tools. Is that right?
We have the opportunity to win the battle to be the identity layer for AI agents. That could be the biggest category in cyber. Cyber is about $280 billion a year; identity management is about 10% of that. This new agent layer could be the biggest category by far. Winning that is job number one.
Tell me your calibration on how much it's acceptable to lose the identity piece to vibe coding in order to win the bigger agent control market. Why would anyone keep paying you annually if they can build cheaper?
They're not mutually exclusive. The attributes of reliability, trust, integration, and vendor support are foundational in both markets. Every organization is aware of the agentic enterprise potential. They want to automate and divert labor budget to technology. But they're getting mixed signals. They need foundational building blocks to wire it all together. That's where the big opportunity is.
Yeah. And so that's where the big opportunity is to take the first steps on what could be the biggest category of cyber.
When you look at things like OpenClaw, everyone is buying Mac minis to air gap it and giving it all their login and passwords. Do you see that? You've accomplished nothing. But that's where the excitement is. When you put rails on that, it feels like you're foreclosing opportunities. How did you look at OpenClaw and how is it informing your thinking?
It's the ChatGPT moment for agents. It opened everyone's eyes to the art of the possible. My son's soccer game parents were talking about OpenClaw. These are non-tech people. They see that agents need data to be interesting. The rails we're putting in place are giving enterprises a list of agents they have, a system of record, and controlling what agents are connected to. The tension is between more data and more connections. Companies like Palantir, Snowflake, and Databricks are doing well by pulling data into one warehouse. But the longer-term model is having the right permissions and access tokens for agents to access data directly. You have to facilitate secure connections that can be monitored and pulled back.
When I said this was going to be an emotional conversation on software development, the nature of our relationship to databases is at the heart of that existential crisis. Let me get your answer directly. It sounds like you're saying the SaaS apocalypse might be real, but it's not real for Okta in the way most people think.
What people miss is that the pie is getting much larger. Software spending is about $1.2 trillion, IT services about $1.8 trillion. The markets are getting bigger. Every piece of technology in the stack will get agentic features. There's a new layer: the digital worker layer. It's hard for legacy companies to build digital workers because they are siloed. There will be losers and disruptors, but that fires me up.
So it's very hard to break through and go broad. You've brilliantly opened the door to the Decoder questions by talking about org charts. Tell me about Okta's org chart. How are you changing it?
The guiding principle is to give great people an area they can be great in. It's people-driven. Second, cluster things to minimize communication paths. But it's hard. Beyond that, org charts are not the solution to culture; it's management and leadership. Get the right managers.
So it sounds like you're pretty functionally structured. How is Okta actually structured? By business line? Do you have a crack AI team off in the corner?
On the go-to-market side, it's functional. On G&A, functional. On R&D, it's by platform. We have two platforms: Okta and Ozero.
The other question I ask everybody on Decoder is about decisions. How do you make decisions? What's your framework and how has it changed?
We're doing introspection here. When I started Okta, I was very decisive. But then I slowed down because I had no safety net. As the company grew, I realized I needed more input, but then I learned to trust my instincts more. The key is deciding which decisions to make. Focus on the important ones and get the details right.
I told you to be emotional.
This is Decoder. Decoder is just therapy for me personally. At this point you can tell what my problems are.
You're casting them out amongst the gods. But it's interesting. So when I started Okta, I worked at Salesforce and felt decisive. But then I slowed down because I had no boss. Then as we went public, I needed more input, but I realized my instincts were good. Now I lean into my instincts and focus on which decisions to make.
Put this into practice for me. The big decision we've been talking about is Okta chasing the agent framework. How did you make that decision? Did you stare at the ocean?
The high-order bit is recognizing that everything in the stack is going to change. It's similar to when I started Okta in 2009, seeing the cloud version of everything. Now everything will be revisited in an agentic world. The demand for connections between agents will be massive. We're good at building something that scales and connects to many systems. That's the bet we're making.
Take me inside the moment. Did you write an email? Open a Google doc? How did it actually work?
Last year I was in the process of meeting all our hundred largest customers in person. I wanted to tell them about our vision of a unified identity platform. At the same time, the teams were working on agent identity.
These meetings, I would pitch what I was talking about, and then there'd be interest in, you know, 'Oh, we should look at this, we should... we didn't know how far along you were.' And then I started throwing in this agentic stuff at the end of the meeting. And whenever I would get to that, the people in the meeting would just stop and they'd be like, 'Wait, talk about that some more.' And then that kept happening until, you know, 25, 30 meetings, 40 meetings in. I would flip it around and we would start with the agents and the new identity type and what customers were thinking about doing with agents and how they're seeing the potential of the digital worker and agents and all the confusion, and we wouldn't get to the other stuff. We had our big conference in the fall, and you know, it was like the last vestiges of the old pitch followed by the agents. And after that conference, I just said, 'Listen, we got to flip this around. People want to hear about the agents. That's the direction they're going, and that's what we need to pivot and totally focus on.'
All right, so let me ask you my crash-out questions about all of this. Here's my first one, and you're a great person to ask this question to because you build a lot of software. You've built a company around building software, like very bespoke, very complicated software, and you're trying to sell a lot of software to people who, like you said, would like to replace labor with technology, right? And that is... there's a lot there. I'm looking at the state of the art in AI right now and I see some cool stuff happening, and I find myself constantly wondering: can the LLM technology we have today, that is a foundation of all of these AI systems, can it bear the weight of our expectations? Can it actually, on any reasonable timeline, do all of the things that people think it can do? Because I can see it doing some things, and then I see it just hit walls over and over again. And I say, well, if it's brittle, people are not going to adopt it because that brittleness is exactly where you want a human being to just be available to overcome whatever boundary the AI is going to find for itself. And I can give you examples, but I'm curious if you see that broadly and if you think the technology can actually develop to the point where the market becomes as big as what you're describing.
Absolutely, the technology can develop. I think there's a lot of wild extrapolations going on right now, which, you know, but I think that even if you don't meet the wild extrapolations people are talking about, the market is still massive. And I think it's going to take a lot of innovation and good product work and good engineering work and good process work to make sure that we can achieve these benefits, even though it's not some wild extrapolation of some magic LLM that can do everything in the world. So I see one example, right? Every software developer I know is like, especially the senior ones, are like, 'I'm now just describing software, like I'm just writing.'
That's a great example. That's a great example. So now I believe that is very real and very powerful. But I also believe that there's going to be more software engineers in 5 years than there are now. And the reason I believe that is not because I think those people are wrong, but I think that what's going to happen is, first of all, there's just way more software that we need to build that can be built. And two, what's going to happen is the software engineers are going to be figuring out how to make it work at scale. How to make sure that systems can be maintained. How to make sure we understand how the agentic workflows what they actually built, and we need to modify them for the next way. No one's ever maintained an agentic developed system for 5 years. No one's ever figured out how to make it scale. That's where all the work is. And when you combine that, you know, we're going to build 10 times more software, that adds up to more people being required to do it. So I think both can be true.
Where are those people going to learn how to do it, right? You've already described this, right? The traditional career path, the traditional org chart is breaking down. I think Meta announced that like one manager will now oversee 50 ICs. When I say we're on the cusp of some wild org charts, that's what I mean. Like some very strange corporate structures are going to blossom here. If the problem is, okay, no one's ever maintained an agentic system for five years and we need more developers to do it. Where are all those developers going to learn the skills to evaluate the code that agents are writing and deploying and saying, 'Okay, you got it wrong. Here's how you need to maintain it.'
