Training Data
Training Data

Factory's Matan Grinberg: The Coming ‘Dark Factory’ Where Software Builds Itself

July 21, 2026

AI Summary

5 min read

Factory's Matan Grinberg: The Coming ‘Dark Factory’ Where Software Builds Itself

Matan Grinberg, co-founder and CEO of Factory, started the company in April 2023—two years before the enterprise market was ready for fully autonomous software development agents. "It was our journey in the desert," he says. "We were focused on fully autonomous agents, but engineers weren't ready. Procurement teams at the enterprise weren't ready." That early timing forced Factory to learn hard lessons about building for developers in large organizations, and those lessons now define the company's operating principles and product strategy.

The Two Years in the Desert

Factory spent its first two years selling a vision that customers weren't ready to receive. Grinberg, a physicist who dropped out of his PhD and had never held a job before starting the company, convinced twenty of the smartest people he knew to join him. They went customer after customer and found that enterprises barely wanted GitHub Copilot, let alone fully autonomous agents.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (00:00) **"Create Obsessed Customers" vs. Customer Obsession** - Matan explains the core operating principle that differentiates Factory from Amazon's famous mantra.
  • 2 (01:36) **Why Factory is a "Dark Horse" in the AI Coding Market** - Matan lays out the competitive landscape and Factory's key differentiation: model independence.
  • 3 (05:40) **The "Two Years in the Desert" and the Decision to Return All Revenue** - A deep look at the low point of the company's journey and the radical honesty that saved it.
  • 4 (11:46) **How a "Ruthless" Hiring Culture Built Resilience Through Rock Bottom** - Matan explains how the team's shared mission and brutal honesty held them together during the crisis.
  • 5 (13:52) **The Pivot: Why Model Performance Wasn't the Most Important Thing** - The turning point came from a change in developer behavior, not just better AI.
  • 6 (16:53) **The "Harness" is the Moat: Why Multi-Model is Better Than Co-Designed** - Factory's technical advantage comes from building a harness that supports many models, not just one.
  • 7 (23:34) **Token Maxing & The Factory Router: From Adoption at All Costs to Cost Rationalization** - Matan describes the market's evolution from trying to get anyone to use AI to needing to manage token spend intelligently.

+ Full timestamped outline available in the app

Show Notes

Factory started building fully autonomous coding agents in April 2023, two years before enterprises were ready. Matan Grinberg now says this is indistinguishable from being wrong. The Factory co-founder and CEO explains how the company survived its "journey in the desert," including the decision to hand nearly all of its revenue back to customers when the product wasn't making developers obsessed. Matan makes the contrarian technical case that a model-agnostic harness beats the model-and-harness co-design that labs like OpenAI and Anthropic favor, because exposing a harness to many models keeps it from overfitting to any single one. He argues open-weight models like GLM will capture the majority of tokens by staying one generation behind the frontier at a fraction of the cost, and that CIOs will soon justify every incremental token the way they justify headcount. Looking ahead, he predicts 90% of coding tokens will run asynchronously—the "dark factory" where software builds itself.

Hosted by Sonya Huang and Pat Grady, Sequoia Capital


Training Data

More from this podcast

Training Data →