Ubisoft scales AI coding agents to 4,000+ developers with Unity Gateway

Ubisoft is a global video game creator and publisher behind franchises including Assassin’s Creed, Far Cry and Tom Clancy’s Rainbow Six. As AI coding adoption accelerated, Ubisoft faced a platform challenge: how to give thousands of developers access to rapidly evolving models and coding tools without multiplying provider integrations, configuration and cost controls. With Unity Gateway, Ubisoft scaled from roughly 200 to more than 4,000 active developers in five months, while centralizing AI governance, managing consumption and accelerating access to frontier models.

Building an AI platform that could scale with adoption

AI coding agents at Ubisoft started from the bottom up. Development teams experimented with Claude Code, GitHub Copilot and other tools through individual licenses and proofs of concept, while Ubisoft used LiteLLM as a gateway for Claude Code traffic. But as adoption accelerated and Ubisoft pursued a CEO-sponsored initiative to make AI broadly available to employees, onboarding thousands of developers and managing gateway performance became increasingly difficult.

Already using Databricks as its Data and AI Platform, Ubisoft turned to Unity Gateway to establish a scalable foundation for developer AI. Through a common endpoint, Ubisoft can provide access to Anthropic, OpenAI and open models such as DeepSeek, Kimi, GLM and Qwen without managing separate provider integrations and contracts. “We certainly didn’t want to have to manage relationships and integrations with five different providers. With Unity Gateway, all of those models are available behind a single endpoint,” said Thomas Donadei, Technical Product Manager, Ubisoft.

This approach also helps Ubisoft keep pace with rapid model innovation. When Databricks makes a new model available, Ubisoft works with its account team to provision the capacity needed to support thousands of users. Developers can typically gain full access to newly released models within 24–48 hours, without each team having to configure a new provider integration.

Scaling from 200 to 4,000 developers without adding friction

Ubisoft initially focused its rollout on game production, including level designers, gameplay developers, and other teams directly involved in building games. For many of these developers, Unity Gateway was their first interaction with Databricks, and Ubisoft wanted that interaction to be nearly invisible.

The Ubisoft Data Office (UDO) pre-provisions developers with Databricks accounts, and with the tools team deploys a custom installer to their desktops. The installer automates personal access token (PAT) creation, Unity Gateway connection and coding agent configuration, removing the need for developers to manually manage authentication or gateway endpoints. Developers can simply launch their preferred coding agent and get to work. “From the user’s perspective, the process is transparent. They don’t need to understand Databricks. The Gateway can essentially run in the background,” said Donadei.

By abstracting that infrastructure away from individual developers, Ubisoft grew from approximately 200 active users in March to more than 4,000 by August, a 20x increase in five months. The company is targeting 5,000 developers as it expands across game production, with plans to eventually bring governed AI access to functions such as marketing, finance and HR.

Managing AI tokenomics at engineering scale

The rapid growth in adoption also created a new AI tokenomics challenge. As thousands of developers began using coding agents every day, their AI spend increased 158%. The platform team needed visibility into who was consuming models, which teams were driving usage and how AI spend was growing across the organization.

Unity Gateway maintains the identity associated with each request, even when thousands of developers access models through the same centralized endpoint. Ubisoft combines this consumption data with Databricks AI/BI Dashboards to track usage across developers and internal teams, giving the platform team a foundation for understanding and allocating AI costs.

To help manage growing consumption, Ubisoft has also built custom monitoring and email alerts that notify teams when usage reaches defined thresholds. Looking ahead, Ubisoft plans to expand these capabilities with team-level budgets and chargeback, giving managers greater visibility and control over AI spending while allocating consumption costs to individual Units of Work (UoWs).

Together, these capabilities are helping Ubisoft move toward a model where developers retain access to the AI tools they need, while platform and engineering leaders gain greater accountability for consumption across the organization.

A foundation for AI across Ubisoft

For Ubisoft, Unity Gateway is becoming more than the infrastructure behind coding agents. Together with Unity Catalog, it gives the company a common foundation for governing the data, models and AI tools its teams use.

“With Databricks Unity Gateway, the core Gateway capabilities are already there. We can spend less time building and rolling out the Gateway itself, and more time making sure developers are using these tools effectively and getting real value from them,” said Donadei.

That shift is increasingly important as Ubisoft moves from thousands of early adopters toward broader AI adoption. Instead of devoting its platform team to maintaining gateway infrastructure and individual provider integrations, Ubisoft can focus on the higher-level questions: which models deliver the best results, where developers are getting value, how to manage rapidly growing AI spend, and how to make the next generation of AI tools available safely across the organization.

With more than 4,000 active users today and 5,000 targeted across game production, Ubisoft now has an AI platform designed to evolve alongside the models, coding agents and providers its developers use.