OpenAI announced a strategic partnership with Dell Technologies on May 18, 2026 to bring Codex — its AI coding and task-execution agent — to hybrid and on-premises enterprise environments. The announcement was made at Dell Technologies World 2026 in Las Vegas, the same event where Dell and NVIDIA unveiled the AI Factory 2.0 infrastructure platform.

The partnership is a significant shift for Codex, which has until now been primarily a cloud-delivered product. Businesses that operate under strict data residency requirements, run sensitive intellectual property through their development workflows, or simply need lower latency than cloud APIs can deliver will now have a pathway to deploy Codex on infrastructure they control.

What the Partnership Actually Does

The integration has two components. First, Codex will connect with the Dell AI Data Platform, which many businesses already use to store and govern their enterprise data on-premises. This matters because Codex’s usefulness scales with how much internal context it has access to — codebases, documentation, operational knowledge, business systems — and getting that context to the model without routing it through public cloud APIs has been a blocker for security-conscious enterprises.

Second, Dell and OpenAI are exploring how Codex, ChatGPT Enterprise, and other OpenAI API products can interface with Dell AI Factory to prepare data, manage systems of record, run tests, and deploy applications in hybrid infrastructure setups.

The combination means an enterprise could run Codex agents against private codebases, internal documentation, and business data without that information ever leaving their own hardware.

The Scale Behind the Partnership

More than 4 million developers now use Codex every week. That number represents a notable growth trajectory — Codex has expanded from a developer novelty to a production tool that teams are embedding across the software development lifecycle.

Use cases already in production at enterprise customers include code review automation, test coverage expansion, incident response reasoning across large repositories, and legacy system analysis. That last category is particularly relevant for large enterprises, where the cost and risk of maintaining old codebases often outweighs the cost of the AI investment required to modernise them.

Codex’s use cases are also expanding well beyond software. Teams are using Codex-powered agents to gather context across business tools, prepare internal reports, route product feedback, qualify leads, write follow-ups, and coordinate workflows across business systems. The coding origin is becoming less relevant as the agent capability becomes more broadly applicable — a pattern described in more detail in what AI agents actually do all day.

Why On-Premises Matters for Enterprise Adoption

Cloud-delivered AI has been the default deployment model for the past several years, and for many use cases it works well. But cloud dependency creates real friction for certain categories of enterprise customers.

Financial services firms dealing with client data, healthcare organisations under HIPAA, legal departments handling privileged communications, and government contractors subject to data sovereignty requirements have all faced the same dilemma: the AI tools they want to use require sending sensitive data to third-party cloud environments. That is often a compliance or legal problem that has no clean workaround.

The Dell partnership is designed to solve that problem for Codex specifically — and by extension for the broader range of OpenAI products that will follow this integration path.

This is not unique to OpenAI. The entire enterprise AI infrastructure market is moving in the same direction. Dell’s AI Factory 2.0, announced at the same event, is built around on-premises NVIDIA Blackwell hardware. Microsoft’s Azure Local is expanding on-premises AI capabilities. The message from the infrastructure layer is clear: cloud-first is giving way to hybrid-by-default in enterprise AI deployment.

What This Means for Business

For software teams: Codex running against your actual codebase on your own infrastructure is a different proposition from Codex with limited context access. The quality of AI assistance improves significantly when the model has complete access to internal documentation, architecture decisions, and business logic. Organisations that have been cautious about cloud-based AI coding tools because of IP concerns now have a cleaner path. (The build vs. buy question for AI development tools looks very different when on-premises deployment is on the table.)

For regulated industries: Finance, healthcare, legal, and government organisations now have a real option for deploying one of the most capable AI development agents in production. The compliance conversation changes when the data never leaves your infrastructure.

For organisations running AI agents at scale: The partnership points toward a broader pattern. OpenAI is building out an enterprise distribution strategy that goes well beyond direct API access — partnering with systems integrators (Accenture, Cognizant, Infosys, TCS, PwC), hardware vendors (Dell), and cloud providers to meet enterprise customers where their infrastructure already lives. That distribution push signals how seriously OpenAI is treating enterprise revenue as a long-term strategic priority.

The timing of this announcement alongside Dell AI Factory 2.0 is deliberate. Both launches are making the same argument: the infrastructure is ready for production-grade enterprise AI, and the remaining question is how quickly your organisation decides to use it.

---

The practical next step is the free Working With Claude field guide. Thirty-two pages covering the ecosystem, Claude Code, and how to govern a rollout properly. Get your copy.