I was recently at a SaaStr Fund portfolio company event, and a CTO came up to me, clearly excited. He told me proudly how 30% of their code was now written by AI and Cursor.
I shrugged.
Not because it wasn’t impressive. Not because they weren’t doing good work. But because…
I no longer care how much of your code is made with AI. It’s the base case now. No extra points.
I still thought this was amazing when another CTO told me the same thing just back in April. I was genuinely impressed. We talked about it for 20 minutes and even on stage at SaaStr Annual and AI Summit in May.
Today? Less than 6 months later? It’s just how we code now.
Look, Cursor is at $1B+ ARR and that’s just amazing. Truly incredible. But the conversation has fundamentally changed.
It’s Just How We Code Now
Because AI coding tools have crossed the chasm. They’re not experimental anymore. They’re not a competitive advantage. They’re baseline.
The fact that 30% or 50% or even 70% of your code is written by AI now no longer really matters. As incredible as Cursor, Claude Code, etc. are — they are now just tools. Tools we pay $1B+ for, gladly. Disruptive tools. But just tools in everyone’s toolkit now.
Every engineering team worth their salt is using AI coding assistants. Copilot. Cursor. Claude. Whatever. The specific tool doesn’t matter as much as you think.
What matters is what you’re building and how fast you’re shipping.
The conversation has shifted from “Are you using AI to code? Really? How much?” to “What are you shipping that your competitors can’t?”
{“model_id”: “text2image_v1/prod/20250912-2317”}
So What Actually Matters Now?
If AI code percentage isn’t a differentiator anymore, what is?
- Velocity. Are you shipping faster than you were 12 months ago? If not, you’re doing it wrong.
- Quality. AI can write code, but is your product actually better? Are you catching bugs earlier? Is your architecture cleaner?
- Product intuition. AI can generate code, but it can’t tell you what to build. The companies winning right now have founders and PMs who know exactly what features will move the needle.
- Talent leverage. The best companies aren’t replacing engineers with AI. They’re making their best engineers 2-3x more productive. Big difference.
The Bottom Line
Cursor hitting $1B+ ARR and a $30B+ valuation isn’t just a company milestone. It’s a market signal.
It tells us that AI coding tools have become infrastructure. Like AWS. Like GitHub. Like your IDE.
You don’t get credit for using AWS. You don’t get credit for using GitHub. And you definitely don’t get credit anymore for what percentage of your code is AI-generated.
The only question that matters: What are you building with it?
Because everyone else has access to the same tools you do.
The moat isn’t the AI. The moat is what you do with it.
What’s Actually Disruptive Now?
You want to know what I do find fascinating? Tools like Replit and Lovable.
Because they let non-coders “code”.
That’s something you genuinely couldn’t do before. That’s a real unlock. That’s expanding the market, not just making existing developers more efficient. I’ve written 9 apps in last 100 days on Replit myself, without a developer. That’s pretty cool. And something that literally could never even be done 12 months ago.
Don’t get me wrong—using AI to make developers more efficient is extremely powerful. Game-changing, even.
But it’s no longer a game changer when everyone is using these tools.
It’s no longer profound. It’s just baseline.
The profound stuff? That’s happening at the edges. Where AI is enabling entirely new people to build entirely new things.
