Transcript
Speedrun doesn’t pretend to be for everyone. The program is built around a very specific kind of founder: someone who’s already shipping, not just theorizing. Its public FAQ says it looks for teams that demonstrate a proven capacity to go from zero to one without waiting for permission, and that are already building a working product before the accelerator starts. In practice, that means Speedrun is not the place for pure idea‑stage founders; it’s for builders who can ship a working product in 12 weeks, not just a slide deck. The investment focus is unusually clear: early‑stage, capital‑efficient teams working in AI‑native categories, with a strong bias toward developer tools, infrastructure, and platforms that can leverage the broader a16z AI‑software‑development‑stack thesis. Companies like Looop, Replit, Cursor, and Lovable are examples of that thesis in practice; they’re not the category itself, just prominent expressions of the same underlying stack. Speedrun’s asks of founders are explicit and intense. The program runs 12 weeks in person, with modules dedicated to rapid product development, go‑to‑market, fundraising, and operational scaling. Founders are expected to iterate weekly, ship concrete user‑validated features, and prepare for a high‑pressure Demo Day in front of a16z‑curated investors. Beyond shipping, Speedrun expects you to lean into the a16z network: using the 5 million dollars in vendor credits aggressively, tapping into a16z growth and marketing advisors, and aligning your product roadmap with the broader AI‑software‑stack narrative. In other words, Speedrun isn’t just giving you 1 million dollars; it’s treating that 1 million as a forcing function to compress 18–24 months of learning into 12 weeks.