OpenAI made ChatGPT Enterprise available through AWS Marketplace on November 9, giving organizations another way to purchase its workplace AI products through an existing cloud supplier.
The launch changes procurement rather than introducing a new model. OpenAI says Marketplace buyers receive the same models and functionality as direct customers, with new features arriving at the same time. Eligible spending can count toward customers’ AWS commitments.
The offering includes ChatGPT, ChatGPT Work, Codex and dots. OpenAI presents the bundle as a way for employees to use AI for everyday work alongside the OpenAI models developers can access through Amazon Bedrock to build applications and agents.
A different route through procurement
For businesses already buying software through AWS, the listing creates an additional purchasing and billing option. OpenAI says eligible customers can also obtain ChatGPT for Healthcare and ChatGPT for Financial Services through the marketplace, subject to eligibility checks.
The AWS Marketplace listing names OpenAI as the seller and directs buyers to request a private offer. It does not publish a universal per-user price; pricing depends on the organization’s requirements and eligibility.
The listing says eligible consumption can count toward an AWS Enterprise Discount Program or Private Pricing Agreement commitment, subject to the customer’s agreement. That condition matters: the existence of the listing does not establish how an individual customer’s contract will treat a purchase.
Billing location is not data location
A prominent caveat in the AWS listing separates procurement from hosting. Purchasing through Marketplace does not guarantee that all customer data remains on AWS infrastructure or inside the customer’s AWS environment.
OpenAI remains the data processor and uses subprocessors, including AWS. The listing describes encryption, administrative controls and a default commitment not to use business data for model training. Those protections do not remove the need to evaluate the service’s processing locations and contractual terms.
Data residency describes where supported content is stored. Inference residency addresses where supported model calculations run. They are different controls, even when a product is purchased through a single cloud marketplace.
A specific limit for international buyers
OpenAI’s Help Center says ChatGPT Enterprise workspaces purchased through AWS Marketplace can use its supported storage regions, but inference residency for those workspaces is available only in the United States.
The documentation also warns that regional inference does not cover every operation. Some processing outside the model’s graphics processors, along with external integrations and system information, may fall outside the selected region. Buyers therefore should not interpret a storage choice as a guarantee that every step stays in the same country.
For an international organization, the immediate decision is whether the purchasing convenience fits its existing cloud agreement and data requirements. OpenAI directs prospective customers to their OpenAI or AWS account teams to confirm eligibility. The final price, treatment of committed spending and applicable residency configuration remain questions for that specific offer.
