Freight brokerages in 2026 look like stock trading floors in 1989. Rooms of people on phones, quoting prices from memory, faxing paperwork across the country. Then electronic trading showed up and those floors emptied out. Peer (YC S26) is that moment for trucking.

https://drive.google.com/file/d/1IsbGimgbdogUQdADAzDloIuPn9zQgEJi

Trucking is a $900 billion industry, and 91.5% of American trucking companies run ten trucks or fewer. Nobody shipping freight can manage hundreds of those relationships, so they hand the load to a broker, who finds the right truck, in the right place, at the right price.

Peer is an AI-native freight brokerage. A shipper sends Peer a load and agents do everything after that: price it, find and vet the trucking company, handle the paperwork, track the delivery, and fix what goes wrong on the road. Even the most tech-forward incumbents need thousands of people to run a brokerage. Peer runs on agents, doing far more loads with far fewer people.

Shippers want a fair price fast, a truck that is insured, and a trucking company that shows up. An agent quotes market rates in minutes instead of days of emails, because it reads the request the moment it lands. It catches a lapsed insurance policy before the freight moves, because it rechecks continuously, not once at signup. And it drops a trucking company that keeps failing, because it has no relationship to protect.

Aaron and Felix are both technical and both dropped out of Harvey Mudd to build Peer. Aaron built voice AI for trucking fleets at Waylens and researched AI agents at the MIT Media Lab. Felix worked in machine learning at Magnefy and shipped AI tools to the LA Department of Water and Power. Together they won the Google DeepMind Hackathon at UCLA.