Inside OpenAI’s DeployCo: Scaling Enterprise AI Through Engineering

OpenAI is moving beyond simple API access to deeply embed its intelligence within the core business processes of global corporations. Through its new subsidiary, DeployCo, the company is deploying specialized engineers to bridge the gap between cutting-edge research and real-world industrial application.

The Rise of DeployCo and the FDE Model

Led by CTO Arnaud Fournier, OpenAI’s Deployment Company (DeployCo) represents a strategic shift toward "Forward Deployed Engineering" (FDE). Unlike traditional software sales, the FDE model places engineers directly at the interface of the customer and the research team. This approach ensures that AI models are not just accessed via an endpoint but are fully integrated into compliant, monitored, and scalable business workflows.

To accelerate this global expansion, DeployCo recently acquired the British consulting firm Tomoro, adding approximately 150 deployment specialists to its ranks. This move establishes a significant leadership presence in Europe, with key operational sites in Paris, London, and Munich. This strategy differentiates OpenAI from traditional consultancies like Deloitte or McKinsey; while the latter provide strategic oversight, OpenAI’s engineers provide the technical "state of the art" required to keep pace with rapid model iterations.

Driving Innovation Through the Feedback Loop

A critical component of DeployCo’s mission is the bidirectional flow of information. While OpenAI maintains strict privacy standards—stating they do not train on customer data without explicit research partnerships—a powerful secondary feedback loop exists.

When deployment engineers encounter model weaknesses, such as deficiencies in document understanding, this data is funneled back to the research teams to improve future iterations. A prime example is the collaboration with the bank BBVA. By identifying specific needs, the solution evolved significantly from GPT-5.0 to 5.5. Furthermore, real-world demands for multi-agent orchestration directly led to the creation of the Swarm open-source repository and the subsequent Agent SDK.

This feedback loop also enables companies to move from simple automation to fundamental process reshaping. For BBVA, instead of merely automating annual credit documents, OpenAI engineers helped implement a system for continuous, daily credit risk assessment, allowing the bank to react instantly to geopolitical shifts.

Explosive Codex Growth and Market Adoption

Despite concerns regarding the EU AI Act and strict data privacy regulations, AI adoption in Europe is accelerating. OpenAI has mitigated regulatory friction by introducing EU data residency and enterprise key management. This has positioned France, Germany, and the UK among OpenAI's ten largest global markets.

The growth of Codex, OpenAI’s developer-centric tool, serves as a bellwether for this momentum:

  • Global Scale: Codex has reached over four million weekly users, representing a fivefold jump in just three months.
  • German Dominance: In Germany, weekly active Codex users have grown more than sevenfold since January 2026.
  • Developer Adoption: Germany currently ranks in the global top five for weekly active users and is among the top three globally for both paid subscriptions and developer activity.

Key Takeaways

  • Strategic Integration: Through DeployCo, OpenAI is using Forward Deployed Engineers to move AI from experimental tools to embedded enterprise infrastructure.
  • Continuous Improvement: Feedback from real-world deployments drives rapid model iterations and the development of new tools like the Agent SDK.
  • Rapid Developer Scaling: Codex is seeing massive global growth, particularly in Europe, signaling a massive surge in AI-driven software development.