Before we get started this week, a quick shameless ad:
If you're enjoying this newsletter then check out my book, Startup Different, which goes deeper on a lot of the business topics I cover here. You can find it on Amazon, Audible and Kindle.
And now, onto this week's newsletter.
Right now, AI pricing doesn’t reflect reality.
Why?
Because it’s being propped up by massive amounts of venture capital.
Billions of dollars are flowing into companies building these models and platforms. That money is covering the true cost of compute, infrastructure, and development.
Which means:
You’re not paying the real price. Not yet.
But that won’t last forever.
What Happens Next
We’re already starting to see the shift.
Pricing models are changing:
Flat-rate plans → usage-based pricing
“Unlimited” → metered tokens
Low entry cost → scaling penalties
And this is just the beginning.
Because eventually, every AI company has to answer the same question:
How do we make this profitable?
And when that happens, costs move downstream to you.
Here’s where it gets dangerous.
AI doesn’t just add cost, it scales cost.
The more you integrate it:
The more tokens you use
The more workflows depend on it
The harder it becomes to remove
So what starts as:
“Let’s test this tool.”
Quickly turns into:
“We can’t operate without this.”
And now pricing changes hit differently.
Because you’re no longer experimenting.
You’re dependent.
The Vendor Lock-In Problem
This is the part almost nobody is talking about.
Once you’ve:
Built workflows around a specific model
Trained your team on a specific tool
Integrated it into your systems
Switching becomes painful.
Not impossible, but costly.
And AI providers know this.
Which means over time, pricing power shifts away from you… and toward them.
What Smart Operators Are Doing Right Now
The companies getting ahead of this aren’t avoiding AI.
They’re just being more intentional about how they use it.
Here’s what that looks like:
1. They Track Usage Early
Not just “are we using AI?”
But:
How often?
For what?
At what cost per outcome?
If you don’t know your usage, you don’t know your exposure.
2. They Tie AI to ROI
Every use case answers a simple question:
Is this saving time, making money, or reducing risk?
If not, it’s noise.
3. They Avoid Over-Centralization
They don’t rely entirely on one tool or provider.
They:
Test alternatives
Stay flexible
Avoid building fragile dependencies
Because optionality matters.
4. They Design for Efficiency, Not Just Speed
Just because AI can do something faster…
Doesn’t mean it’s worth doing at scale.
Smart companies optimize for:
Cost per result
Not cost per action
AI is one of the most powerful tools businesses have ever had.
But it’s also one of the easiest to misuse.
Because right now, it feels cheap.
Accessible.
Almost free.
And that’s exactly what makes it dangerous.
The Upcoming Shift
Over the next 12–24 months, we’re going to see a transition:
From:
“How do we use more AI?”
To:
“How do we use AI efficiently?”
The winners won’t be the ones who adopted the fastest.
They’ll be the ones who understood the economics early.
AI isn’t cheap. It’s just temporarily underpriced.
And the businesses that build with that in mind?
They’re the ones that won’t get caught in the trap.
Thanks for reading!
Dave
