Why Durable Companies Avoid the AI Label
Durable companies don’t call themselves AI companies. They pick a clear problem, solve it with technology, and move on rather than starting with tech and working backwards.
Here are seven examples of what I mean with companies we’ve invested in.
Pave Robotics
They build robots that seal cracks in asphalt. Operate 24/7. Replace a crew of six. Nobody calls them an AI company. They’re a road repair company that uses robotics and machine vision.
[P] 42% of American roads are in poor condition. The U.S. spends $27B a year on maintenance. Labor shortages keep worsening. Someone asked, why are humans still doing this? That question, not the technology itself, is what matters.
Nuvocargo
Nuvocargo does freight forwarding between the U.S. and Mexico, a market worth almost a trillion dollars that still ran on paper, phone calls, and net-90 payments. They integrated freight, customs, insurance, and financing into one platform. Worth $250M. The founder grew up watching his father’s freight business suffer from those inefficiencies. Again, not an AI company. A logistics company using modern tools.
Cents
Cents operates in the most boring industry you can imagine: laundry. They power 1 in 7 laundromats in America. Built an AI call center that handles customer service, but they’re not selling “AI.” They’re selling operational efficiency for small business owners who were drowning in phone calls.
SkyFi
SkyFi made satellite imagery accessible to anyone. What used to require a hedge fund or the NSA can now be used by a tree trimmer checking which homes need service.
Fig
Fig is building Shopify for mobile in developing countries. The founder’s dad was a pool cleaner. He watched small business owners struggle without access to basic business tools. Now a food stand operator in Mexico City can run their entire business from a phone.
Saronic
Saronic is building autonomous maritime vessels. The founder saw the risks in maritime logistics: lives lost at sea, fragile ports, vulnerable supply chains, and built around that.
Figure AI
Figure AI is building general-purpose humanoid robots for real work: factories, warehouses, and even household environments. The goal is straightforward: automate repetitive and dangerous jobs, fill labor gaps, and increase output.
The Pattern
In each case, the pattern is the same:
Start from a large, visible inefficiency
Ask why it still exists
Use technology to rebuild the system
Compare this approach to the AI horoscope apps and the “spiritual wellness marketplace” I got pitched a couple of weeks back. They’re just chasing novelty, and they’ll peak with the hype cycle.
Conclusion
Ten years from now, nobody will remember the AI horoscope startup. But we’ll still need roads, freight, laundry, and supply chains. The companies solving those foundational problems will be the ones still here.
We’re investing across the dumb, dull, dirty, and dangerous. DM’s open.
