OpenAI’s frontier models — including GPT-5.5 — and its Codex coding agent are now available on Amazon Web Services, giving enterprise customers a way to run leading OpenAI technology inside their existing AWS environment. The announcement landed in early June 2026, extending OpenAI’s reach across the three major cloud providers.
The Deal
Rather than forcing enterprises to open a separate account with OpenAI, the AWS integration lets companies consume GPT-5.5 and Codex under their existing AWS contracts and billing. That removes a major procurement hurdle: many large organizations have standardized buying, security review, and compliance on one cloud vendor.
What’s Available
- GPT-5.5 — OpenAI’s reasoning powerhouse, usable for analysis, customer support, and agent workflows
- Codex — the AI coding agent, deployed for development tasks inside corporate GitHub and CI environments
- Standard AWS tooling — identity management, encryption, VPC networking, and CloudWatch-style logging around model calls
Why It Matters for Enterprises
For customers, the appeal is straightforward: frontier AI quality without changing cloud vendors or procurement processes. Security teams get model traffic inside governed, auditable AWS environments instead of a separate vendor pipeline, and cost shows up in the cloud bill finance already reviews. For developers, swapping between AWS and other clouds becomes easier, reducing lock-in concerns.
The Cloud Wars
The move is notable because Amazon is simultaneously a huge investor in and cloud partner of Anthropic, OpenAI’s chief rival. OpenAI’s expansion onto AWS shows that in practice the AI platform business runs multi-cloud, and customers increasingly want to mix frontier models from several labs. It also pressures every cloud to carry the broadest catalog of top models to stay competitive.
What This Means
OpenAI becomes more deeply embedded in the enterprise IT stack at a moment of peak competition for agent and foundation-model workloads. For AWS customers, it means frontier AI on demand without a single-vendor commitment — and for the industry, further evidence that model distribution through established clouds is the fastest route to enterprise scale.