OpenAI models on AWS Bedrock as OpenAI ends Microsoft exclusivity
AWS announces OpenAI models, Codex, and Bedrock Managed Agents after OpenAI revised its Microsoft agreement on April 27, 2026, enabling broader cloud availability.
OpenAI models on AWS Bedrock became available this week after OpenAI and Microsoft revised their partnership, lifting previous exclusivity and clearing the way for Amazon Web Services to host the company’s latest models and tools. The move, confirmed on April 27 and followed by an Amazon announcement on April 28, 2026, brings OpenAI’s reasoning models and its code-writing Codex service to AWS’s Bedrock platform. The shift marks a significant redistribution of AI infrastructure access among the major cloud providers and is likely to reshape enterprise AI options.
Amazon adds OpenAI models to Bedrock
Amazon disclosed that its Bedrock service will now offer OpenAI’s newest foundational models and the Codex code-generation models to customers. The company said the integration includes tools and runtime support designed to let developers select OpenAI models directly from Bedrock’s model catalog.
AWS framed the availability as part of a broader effort to give enterprises more choice in model selection and deployment options. The announcement described the update as a first step toward deeper collaboration with OpenAI and positioned Bedrock as a hub for AI application development across multiple vendor models.
Bedrock Managed Agents introduced for OpenAI reasoning models
Alongside model access, Amazon unveiled Bedrock Managed Agents, a new service built to run and manage AI agents that use OpenAI’s reasoning-focused models. The offering includes agent steering controls, security features, and lifecycle management aimed at enterprise deployments.
AWS said Bedrock Managed Agents will provide governance controls and operational tooling to help organizations run stateful, multi-step agent workflows securely. The managed approach is intended to lower the barrier for companies building AI-driven automation while imposing guardrails around agent behavior.
Revised OpenAI–Microsoft agreement removes exclusivity
The availability of OpenAI models on AWS followed a revision to OpenAI’s agreement with Microsoft announced April 27, 2026, that removed exclusivity rights Microsoft previously held over OpenAI’s products. Microsoft had been a major investor and cloud partner, and the contractual change resolved a key obstacle that had prevented other cloud providers from offering OpenAI services directly.
The amendment enabled OpenAI to expand commercial relationships with multiple cloud vendors, a shift that came after OpenAI negotiated a multibillion-dollar commercial arrangement with Amazon earlier this year. The revised terms are likely to affect how major clouds carve out product bundles and differentiate on integration, pricing, and enterprise features.
Cloud alliances shift as companies pursue alternative partners
The emergence of OpenAI models on AWS highlights broader strategic re-alignments among leading cloud and AI firms. OpenAI has broadened its infrastructure partnerships beyond a single cloud, while other technology companies have sought their own preferred AI suppliers and model partners.
These realignments include Microsoft’s growing engagement with rival model providers and development of its own agent products, and other large providers forming dedicated arrangements to host or integrate competing models. The shifting network of alliances reflects competition over where AI workloads run and who controls the developer and enterprise ecosystems.
Implications for developers and enterprise customers
For developers, the arrival of OpenAI models on Bedrock means more options for latency, data residency, and integration into existing AWS services. Enterprises that already use AWS can now evaluate OpenAI-powered workflows without a cross-cloud migration, potentially accelerating proof-of-concept work and production deployments.
At the same time, broader availability raises operational questions about model governance, data protection, and cost management. Organizations will need to compare implementation differences across clouds, including security controls, service-level guarantees, and how each provider handles model updates and versioning.
Competitive dynamics and what to watch next
Market observers say the immediate effects will be seen in how cloud providers package model access and agent tooling as enterprise differentiators. Pricing, contractual terms, and performance SLAs are likely to become focal points as large accounts evaluate multi-cloud AI strategies.
Key items to watch include the rollout timeline for Bedrock Managed Agents across regions, the range of OpenAI models available through Bedrock, and how Microsoft and other cloud rivals respond with their own agent and model offerings. Enterprise procurement cycles and systems integrators’ recommendations will further shape adoption patterns over the coming months.
OpenAI’s decision to broaden cloud partnerships and AWS’s rapid inclusion of OpenAI models into Bedrock mark a pivotal moment in the commercialization of generative AI, offering enterprises more choice while intensifying competition among the cloud giants.
