Live Session: How to Find Your Moat and Become Investable
Over the past few months, we've worked with hundreds of founders inside AltaLab.
And a pattern kept repeating: the teams that made real progress could answer one question clearly:
Why does this product need to exist, and why are you the right team to build it?
That's the question we've been exploring with founders at AltaLab.
Some of them made enormous progress and are now in conversations with AltaIR Capital and other investors — a strong signal that the program is producing companies investors want to back. At some point, you start wondering whether it's time to organize a Demo Day.
Others struggled to reach this point.
The patterns behind both outcomes turned out to be surprisingly consistent.
Next week, I'm hosting a live session:
Inside AltaLab: How to Find Your Moat and Become Investable
I'll share behind-the-scenes insights from AltaLab's first cohorts — the actual founder journey, why some startups passed the bootcamp and secured funding, why others struggled as early as the online module, and what those lessons mean for your own path.
We'll discuss:
why AI itself is rarely the moat;
why workflows matter more than features;
why execution speed becomes an advantage early on;
why distribution and positioning matter as much as technology;
what separated the companies that moved forward from those that didn't.
This session is for:
founders looking for a clearer framework before launch;
early-stage teams seeking more structure;
experienced entrepreneurs who want to fix what's slowing them down.
If you want a straight answer on whether your company has a real moat and a real shot at becoming one of the companies that leads the category — this session is for you.
What’s Happening in AI: The Model Dependency Problem
Speaking of moats — here’s one place they don’t live: the model itself.
People spent the past few weeks on Polymarket betting on when access to Claude Fable 5 would be restored for US customers — right alongside World Cup predictions and Bitcoin price forecasts.
This episode is a good reminder that most AI startups don’t control the infrastructure they depend on. They just build on top of it. When pricing changes, rate limits appear, or access gets restricted, someone else’s product decision can suddenly become part of your roadmap.
This is also why we’re seeing growing interest in:
open-source models;
local models;
multi-model architectures.
Founders are designing products that can switch between models — routing tasks to the best performer, optimizing cost and latency, and maintaining continuity if any single provider becomes unavailable.
Building a company around a single model, no matter how capable it is today, is rarely a durable strategy.
Models will improve. Prices and policies will change.
The strongest companies own something deeper:
customer relationships;
proprietary data;
distribution;
workflows.
Those assets remain valuable regardless of which model sits underneath.
That’s the direction this is all heading. At some point, AI stops being a differentiator and starts being infrastructure — the same way electricity did.
Consumers rarely care which power plant generated their electricity. Eventually, they may care just as little about which model generated their answer.
AI access is turning into a utility, and the value is migrating upward.
Once AI is just infrastructure, the differentiation has to come from somewhere else.
So the question for founders is:
How are you building today so you don’t depend on the stability of any single model provider?
And where does your differentiation actually live once the model stops being one?
P.S. As for Anthropic — if your roadmap can influence a Polymarket contract, you probably qualify as critical infrastructure 😀