This is one of my primary investment themes for '26. I've started nibbling on enterprise software names. I'm early, but I find it easier to monitor them when I have small (.2%) positions.
I'm in the IT industry, use tools like CRM and NOW, and I can say with confidence that those thinking AI's going to replace imbedded software really doesn't understand much about - well, about enterprise software. The private data housed inside the software can only be accessed and interpreted by the software - or accessed via APIs but that's limited compared to what's needed for inference. So the enterprise software co's control the ability of AI agents to access the data. That's game over - they keep their moat and build on it. Unmaintainable AI-generated code isn't going to displace these systems that companies run on - not anytime soon.
$NOW I can't bring myself to buy - because IMO it's just terrible, terrible software that lowers the productivity of everyone using it. Maybe that means the agents will be even more valuable, but I'll have to see it first.
This is the kind of 'boots on the ground' insight I love.
You nailed the technical reality that the 'AI Bears' are missing: Data Gravity. The idea that an external AI agent can just 'read' proprietary enterprise data without the native application’s logic/governance is a fantasy. As you said, the software companies control the gateway (APIs), which means they control the AI utility.
I also love the 'nibbling' strategy. It’s a great psychological hack. Nothing focuses the mind on a ticker like having a stake in it.
Thank you for this very well written and reflected article. I like how put AI and the software leaders back into perspective. For example I looked at Salesforce from a fundamentals perspective and found a pattern that signals pure strength rather than weakness. The company is becoming an economic powerhouse with strong cash generation potential. This definitely adds to your findings.
Part of the appeal of enterprise software is the product does what it says on the tin or you can sue someone over it. Vibe coded software doesn't provide a litigation target. That's a bit of a headwind to AI products trying to compete with enterprise software.
The 'Laughter Indicator' is honestly one of my favorite buy signals. When a pitch for a cash-gushing monopoly gets laughed at, you know the sentiment is washed out.
You nailed the metric mismatch: The market is currently obsessed with the speed of the engine (tokens per second), while completely ignoring the weight of the fuel (data gravity).
Let them chase the speed. We'll stick to the gravity. Great to have you reading.
This is one of my primary investment themes for '26. I've started nibbling on enterprise software names. I'm early, but I find it easier to monitor them when I have small (.2%) positions.
I'm in the IT industry, use tools like CRM and NOW, and I can say with confidence that those thinking AI's going to replace imbedded software really doesn't understand much about - well, about enterprise software. The private data housed inside the software can only be accessed and interpreted by the software - or accessed via APIs but that's limited compared to what's needed for inference. So the enterprise software co's control the ability of AI agents to access the data. That's game over - they keep their moat and build on it. Unmaintainable AI-generated code isn't going to displace these systems that companies run on - not anytime soon.
$NOW I can't bring myself to buy - because IMO it's just terrible, terrible software that lowers the productivity of everyone using it. Maybe that means the agents will be even more valuable, but I'll have to see it first.
This is the kind of 'boots on the ground' insight I love.
You nailed the technical reality that the 'AI Bears' are missing: Data Gravity. The idea that an external AI agent can just 'read' proprietary enterprise data without the native application’s logic/governance is a fantasy. As you said, the software companies control the gateway (APIs), which means they control the AI utility.
I also love the 'nibbling' strategy. It’s a great psychological hack. Nothing focuses the mind on a ticker like having a stake in it.
Great to have you along for the ride in '26!
Thank you for this very well written and reflected article. I like how put AI and the software leaders back into perspective. For example I looked at Salesforce from a fundamentals perspective and found a pattern that signals pure strength rather than weakness. The company is becoming an economic powerhouse with strong cash generation potential. This definitely adds to your findings.
Part of the appeal of enterprise software is the product does what it says on the tin or you can sue someone over it. Vibe coded software doesn't provide a litigation target. That's a bit of a headwind to AI products trying to compete with enterprise software.
The 'Laughter Indicator' is honestly one of my favorite buy signals. When a pitch for a cash-gushing monopoly gets laughed at, you know the sentiment is washed out.
You nailed the metric mismatch: The market is currently obsessed with the speed of the engine (tokens per second), while completely ignoring the weight of the fuel (data gravity).
Let them chase the speed. We'll stick to the gravity. Great to have you reading.