OpenAI and Dell Expand Codex Deployment to Hybrid and On-Premises Environments
OpenAI and Dell have announced a significant expansion of Codex deployment capabilities to hybrid and on-premises environments. This integration targets the Dell AI Data Platform and the Dell AI Factory stack, aiming to bring Codex agents closer to customer data repositories that reside within client perimeters. OpenAI currently boasts over 4 million weekly users of Codex among developers.
This joint announcement positions Codex as the first OpenAI agent that can be deployed outside of the public OpenAI cloud infrastructure and in proximity to business data and systems of Dell’s enterprise clients. The scope of their partnership includes the integration of Codex with the Dell AI Data Platform, a comprehensive orchestration and governance framework for enterprise data that can be deployed on-site. Furthermore, the collaboration explores an interface connecting Codex with ChatGPT Enterprise, OpenAI APIs, and the Dell AI Factory—Dell’s hardware and software platform that competes with Nvidia configurations, HPE Private Cloud AI, and IBM Watsonx. Ihab Tarazi, Senior Vice President and CTO of Dell’s Infrastructure Solutions Group, stated that this initiative would enable clients to “deploy AI where enterprise data already resides, on their own premises, and to industrialize AI agents at scale securely.”
Codex, highlighted by OpenAI as one of its fastest-growing enterprise products, is utilized weekly by over 4 million developers for various tasks, including code review, test coverage, incident response, and contextual analysis of large repositories. OpenAI is now aiming for broader use beyond development, extending into the collection of context across internal tools, report preparation, product feedback routing, lead qualification, drafting follow-ups, and task coordination between business systems.
A Partial Return to On-Premises Solutions for Sensitive Workloads
This expanded usage positions Codex within the productivity agent landscape, a competitive space occupied by Microsoft Copilot Studio, Salesforce Agentforce, Google Agentspace, and ServiceNow AI Agents. For European IT departments, the introduction of an on-premises deployable OpenAI agent redefines governance perspectives: inference streams and contextual data processing can remain within the client’s Dell perimeter without the need for data transit through OpenAI’s data centers. This direction aligns with a market trajectory that has been gaining momentum since 2025, advocating a partial return to on-premises solutions for sensitive generative and agent AI workloads, following an initial phase of cloud-only adoption deemed incompatible with data minimization and European localization requirements.
The partnership directly responds to concerns raised by European IT leaders since the expansions of ChatGPT Enterprise and Codex: the need to share data with OpenAI’s data centers raises issues related to localization, governance, and data minimization as mandated by GDPR, NIS2, and DORA, while also being subject to the extraterritorial scope of the Cloud Act, given OpenAI’s status as a U.S. company. With the Dell AI Data Platform and Dell AI Factory, clients retain their data repositories while bringing the model closer to their environment, contrasting with the previous cloud-only framework that necessitated data export.
However, the arrangement raises important questions on three critical factors for security and compliance leaders. Firstly, the precise modalities for distinguishing between the OpenAI control plane (model updates, telemetry, centralized logging) and the Dell data plane (client data, inference contexts, local logs) have not been detailed in the announcement. Secondly, the implications of the Cloud Act on artifacts that pass through the OpenAI control plane (such as model weights, aggregated usage logs, updates) will continue to apply, even while business data remains within the client’s Dell perimeter. European Chief Information Security Officers (CISOs) will require specific contractual terms and architectural clarifications before making decisions on moving a pilot to production.
A Platform Response to Nvidia and HPE Configurations
On Dell’s side, the partnership with OpenAI extends the company’s platform strategy built around the Dell AI Factory, which was launched in 2024 and has since expanded through collaborations with Nvidia, Meta, and Hugging Face. The addition of Codex and ChatGPT Enterprise provides Dell with a unique selling point against direct Nvidia AI Enterprise configurations, HPE Private Cloud AI, and IBM Watsonx offerings by incorporating a deployable OpenAI agent that operates behind the client’s firewall.
For European integrators and resellers within the Dell channel, this partnership paves the way for a new cycle of qualification and marketing of on-premises deployed OpenAI agents, with use cases that reach beyond software development. Likely sales angles include modernizing shared service centers, automating finance and legal support functions, and industrializing incident response within existing Dell environments. The potential success will depend on how swiftly Dell and OpenAI can release reference architectures and deployment kits that meet European compliance requirements, as well as the channel’s ability to construct offerings that adhere to sector-specific obligations, particularly those associated with essential entities under NIS2 and financial establishments governed by DORA.



