Wrapping Up Our 2026 True Ventures Fellowship

That's a wrap! Our eighteenth True Ventures Fellowship came to a close this month, and eleven fellows have headed back to campus after a summer inside AI-native True portfolio companies.

This year's program had one organizing idea: AI fluency. 

When we kicked things off in June, we said the goal wasn't just to place talented students at great companies. The goal was to send them back out with accelerated AI-building skills they could put on their resumes and use to land jobs after college.

The gap between people who build at the frontier with AI and people who don't is widening fast, and the window for developing those skills naturally, rather than retrofitting them later, is closing. So we went all in on theming this year's program to suit the needs of today's fast-growing AI companies.

What They Built

Eleven fellows worked across ten True companies: Sailplane, Zephyr Cloud, Concentrate AI, Riley, Chatsee, Moneta Health, GT Bio, Daydream, Item, and Nectar Social. Every week, they posted what they'd shipped. Here are some of the projects they worked on, so you can see the caliber of their talent:

• A monthly regulatory compliance agent at a digital health company that pulls federal and state rule changes from official sources, assesses each one against the company's specific setup, and drafts a report a teammate reviews instead of researching from scratch.

• A pre-call briefing tool for AI strategists each juggling more than twenty customer accounts — it reads the calendar, finds the next customer call, and surfaces that account's open items and relationship history.

• A production system for a company's new podcast, built inside the tools the team already used, plus a weekly analytics brief that replaced more than an hour of manual data pulling.

• A lab inventory automation that files reagent arrivals and storage locations out of a Slack channel and into a searchable database, so scientists can ask where something is instead of scrolling back through months of messages.

• The chat feature for a parenting app's iOS build, owned end to end from parent interviews through prototype to implementation — and a mobile CRM chatbot built by a fellow who had never written iOS before this summer.

A pattern ran through nearly all of it. They built into the tools their teams already used, favored existing infrastructure over another subscription, and kept a human in the loop on purpose. Several shipped a first version, realized it had only moved the friction rather than removed it, and said so before rebuilding; a rare instinct that usually takes years to develop.

Encouraging AI Fluency

Our goal in June was to accelerate AI fluency. By August, reaching for these tools wasn't something this class had to remember to do. It was simply how they worked.

That isn't a curriculum outcome. It's a repetition outcome — and it's the same way we build fluency inside the firm. We ran three AI builder events this year with True founders and their networks: working sessions, not panels. And we host Scrappy Sessions weekly in our SF office to welcome builders from all domains to demo what they're working on. 

That work keeps going now that the fellowship has wrapped. On September 10th, Scrappy Sessions comes to the main stage at Connected Stack at the Presidio Theatre in SF, where fifteen builders will demo in front of our team and an audience of hundreds. Apply to demo if you have an AI-built project to show off, or join us and get inspired by what other builders are inventing.

Thank You

To the founders and teams who hosted a fellow this summer, and to everyone at True who made room for this class — thank you. This program only works because so many people say yes to it.

And to Oni, Thu, Oscar, Aditya, Pear, Shantanu, Thalia, Greeshma, Charlotte, Alex, and Sophia: eighteen classes in, you set a new bar. We can't wait to see what you build next.

Thinking about next summer? Sign up here to be the first to hear when our 2027 program opens.