Stripe Engineers Boosted Output Tenfold With AI Tools

The company saw a massive surge in coding productivity using in-house agentic software throughout early 2026.

Updated on Sept. 28, 2026 in Artificial Intelligence

Isometric editorial illustration of a complex crystalline structure built from metallic and translucent conduits, representing scaled software engineering productivity. — Stripe engineers reported a tenfold increase in software development output during the first half of 2026, driven by the adoption of in-house AI agentic tools. AI Illustration. Upload story photo >

Live Poll

Do you believe AI tools will significantly increase your own productivity at work?

Stripe engineers achieved a tenfold increase in output during the first half of 2026 by utilizing in-house agentic coding tools known as Minions. One engineer alone successfully merged more than 600 pull requests during this period.

Why it matters

The shift highlights how automated coding tools can drastically scale software development speed and capacity. By integrating AI, the firm has accelerated project completion timelines and altered traditional software engineering workflows.

Stripe now merges 1,300 pull requests per week written by AI agents and conducts over six billion test runs per day. This infrastructure relies on data guardrails implemented in 2017 to maintain codebase integrity.

The players

Stripe

Stripe is a multinational financial services and software company that provides payment processing software and application programming interfaces for e-commerce websites.

Patrick Collison

Patrick Collison is the co-founder and CEO of Stripe who reported on the company's internal software development productivity benchmarks.

The details

Engineers utilize the specialized Minions system to automate code generation, resulting in a volume of production pull requests that reached over 1,000 per day in 2025. While productivity has soared, the company noted that incidents per unit time also increased following the widespread implementation of these AI tools.

Stripe began implementing data guardrails in 2017.
In 2024, the average developer output was 2.3 pull requests per week.
Stripe produced over 1,000 production pull requests daily in 2025.
During the first half of 2026, one engineer merged over 600 pull requests.

The Tech Race

This move marks a departure from traditional human-centric coding paradigms by shifting the bulk of production tasks to agentic systems. It positions the firm at the forefront of the industry-wide transition toward AI-augmented software development lifecycles.

Users may experience faster service updates and new feature deployments as project completion speeds accelerate. However, the reported increase in incident rates could potentially impact the reliability of the software platforms that rely on Stripe's infrastructure.

The takeaway

The rise of agentic tools indicates that software quality control is becoming just as critical as code generation speed. Organizations adopting these tools must balance high-velocity output with robust automated testing to prevent reliability issues.