Skip to content
Y CombinatorY Combinator

He Built the World's #1 Open-Source Coding Agent

Jay V is the founder and CEO of Opencode, an open-source alternative to Claude Code that works with any model you want. It's one of the fastest-growing products in AI: 13 million monthly active users, 20X growth this year, and more tokens processed daily than all of OpenRouter. But the overnight success took 16 years, one legal entity, and nine YC applications. In this episode of the Lightcone, Jay explains how Anthropic's attempt to block Opencode accidentally fueled its rise, how 16 years of near-misses prepared him to catch lightning in a bottle, and why most of the world still hasn't experienced the magic of a coding agent. https://opencode.ai Apply to Y Combinator: https://www.ycombinator.com/apply Work at a startup: https://www.ycombinator.com/jobs Transcript: https://ycrootaccess.substack.com/p/how-opencode-became-the-worlds-most Chapters: 00:00 — Intro 00:44 — OpenCode's Explosive Growth 01:16 — 20x Growth, 13M Users, and 7 Trillion Tokens 03:39 — The Anthropic Controversy That Changed Everything 05:43 — Bringing AI Coding Agents to the World 06:39 — When Open Source Models Became Good Enough 08:56 — What Millions of Developers Are Actually Using 13:31 — Why OpenCode Is Huge Outside the US 15:27 — Why Fortune 500 Companies Choose OpenCode 16:36 — The Economics of AI Tokens 20:02 — How Enterprises Are Using Coding Agents 22:58 — AI's New Unit Economics 24:56 — Why Model Choice Matters 29:55 — The Product Decisions Behind OpenCode 34:21 — A 16-Year Overnight Success 41:16 — Why Jay Never Gave Up

Jay VguestJared Friedmanhost
Jul 24, 202644mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

OpenCode scales global open-source coding agents with model choice economics

  1. OpenCode reached ~13M monthly active users, ~4.6M weekly active users, and processes ~7T tokens/day after ~20x growth this year.
  2. A January Anthropic clampdown on using Claude Code subscriptions via OpenCode backfired by legitimizing OpenCode and accelerating discovery and adoption.
  3. Open-source models became “good enough” for real work, enabling a $10/month subscription that lets users switch among many models based on cost, speed, and strengths.
  4. Usage data shows developers actively “budget tokens,” often switching to ultra-cheap models (e.g., DeepSeek Flash) near limits and choosing models for specific tasks (e.g., GLM for frontend).
  5. Enterprise adoption is largely bottom-up: employees adopt first, then procurement/security teams request agreements, driven by desire for model/vendor flexibility and spend controls.

IDEAS WORTH REMEMBERING

5 ideas

Model choice—not a single model—became OpenCode’s core wedge.

OpenCode positioned itself as a neutral “harness” supporting 70+ models/providers, letting users and companies avoid lock-in and benefit from lab-to-lab competition.

A competitor’s enforcement action can function like an endorsement.

Anthropic’s attempt to block requests referencing “OpenCode” signaled OpenCode was important enough to target, putting it on the same pedestal as Claude Code and driving new user trials.

Open models crossing the “real work” threshold unlocked a viable low-cost subscription.

Once models like Kimi/GLM/DeepSeek felt close enough in capability (and sometimes faster), OpenCode could offer a $10 plan that made agent workflows accessible beyond wealthy markets.

Token budgeting behavior is a major product driver outside the US.

Many users extend usage by switching to very cheap models (e.g., DeepSeek Flash) as they hit limits—an optimization mindset that differs from “throw money at it” usage patterns.

Speed and specialization can beat raw frontier quality for daily workflows.

Users adopted models due to tokens-per-second responsiveness (Kimi) or perceived task strengths (GLM for frontend), suggesting routing and UX around speed/task fit matters.

WORDS WORTH SAVING

5 quotes

Most people in the world still haven't experienced the magic of a coding agent.

Jay V

We also recently started processing around 7 trillion tokens per day.

Jay V

What it inadvertently did was it put OpenCode and Claude Code on the same sort of pedestal. It, like, equated the two products in some ways.

Jay V

You really know you have product market fit when like enterprises are bugging you to sign the security agreement so they can use your product.

Jared Friedman

It took more than a decade to get in, let's just put it that way.

Jay V

Explosive user and token-volume growth metricsAnthropic/Claude Code controversy as growth inflectionOpen-source model quality catching up to frontierModel marketplace positioning and multi-model pickerGlobal adoption in developing countries and ChinaEnterprise bottom-up adoption and governance needsToken economics: subsidies, whales, volume discounts, GPU utilization

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.