AWS Launches OpenAI Models After Microsoft Abandons Azure Exclusivity

AWS Launches OpenAI Models After Microsoft Abandons Azure Exclusivity

OpenAI has sharply expanded the distribution channels for its models: as of April 28, 2026, AWS announced that it will begin offering them to customers through Amazon Bedrock. The news came one day after Microsoft and OpenAI officially renegotiated their partnership terms and removed the previous Azure exclusivity on the resale of OpenAI technologies.

What Changed

Until this week, corporate customers who needed OpenAI models in their preferred cloud infrastructure had effectively only one option: Azure. On April 27, Microsoft announced that its license to OpenAI's intellectual property remains in force until 2032, but it is now non-exclusive. Microsoft, however, retains its status as primary cloud partner, and OpenAI products are still expected to launch first on Azure—unless Microsoft cannot or does not wish to support the required capabilities.

In practice, this represents a significant shift: OpenAI can now sell its models not only through its own services and Azure, but also through other clouds. For AWS, this closes one of the major gaps in its Bedrock lineup. For OpenAI, it opens access to companies that have already built processes, security, procurement, and compliance around Amazon Web Services and do not want to migrate everything for a single model provider.

The deal also removed the old AGI clause that tied Microsoft's commercial rights to the hypothetical achievement of artificial general intelligence. In real business, it seemed too vague and hindered the development of long-term, predictable agreements. The new contract makes the relationship between the two companies less ideological and far more market-driven: Microsoft remains a major shareholder and partner, but OpenAI gains freedom in distribution and space for new alliances in the future.

What Appeared in AWS

AWS has launched access to OpenAI products in limited preview within Amazon Bedrock. According to the company, several current models became available for testing immediately on April 28, with the most powerful GPT versions to arrive in the coming weeks. This extends beyond the models themselves to a broader set of services designed to make OpenAI scenarios convenient for large teams already living within the AWS ecosystem.

• Latest OpenAI models in Amazon Bedrock
• Codex for work in CLI, desktop applications, and VS Code
• Bedrock Managed Agents based on OpenAI
• Usage through AWS accounts, IAM, and PrivateLink
• Cost tracking in existing AWS cloud contracts

According to AWS's description, OpenAI models in Bedrock inherit mechanisms familiar to corporate clients: encryption, logging via CloudTrail, guardrails, and network isolation. Special emphasis is placed on agents: OpenAI and Amazon announced back in February a joint development of Stateful Runtime Environment—an environment where an agent can preserve context, tool state, and memory between steps, rather than operating in a "one request–one response" mode.

"Our customers have been asking for this for a long time," said AWS head

Matt Garman, commenting on the launch.

For the market, this is not simply another way to call an API. AWS is trying to demonstrate that it can sell not just a collection of disparate models, but a ready-made platform for production scenarios: with security, orchestration, agents, and billing within a familiar framework. This is exactly what large companies need—those already using Bedrock and wanting to add OpenAI without migrating to Azure, changing procurement processes, and reassembling compliance requirements.

Where the Risk for OpenAI Lies

The expansion into AWS looks like a strong commercial move, but it coincided with an uncomfortable moment for OpenAI itself. According to The Wall Street Journal, the company failed to meet some internal revenue growth and user base targets in 2026, and ChatGPT did not reach its internal goal of 1 billion weekly active users by the end of 2025. OpenAI and its leadership publicly dispute this interpretation, but the fact of the discussion itself matters: the market has begun looking not only at model quality but also at the economics of their scaling.

The problem is that OpenAI is simultaneously taking on enormous infrastructure commitments. The company previously announced a multi-year partnership with Amazon, Amazon's $50 billion investment, and expanded computing purchases from AWS by another $100 billion over eight years. Microsoft, meanwhile, remains a key Azure partner.

The wider the distribution, the more potential sales channels OpenAI has—but the sharper the question becomes: can demand cover such a volume of computing? This is precisely why the main point of the news is not that OpenAI models are "now everywhere." What matters far more is something else: multi-cloud is no longer theory but OpenAI's working strategy.

The next test is turning this availability into sustainable corporate revenue and proving that new sales channels actually accelerate business rather than simply expanding reach on paper, while Anthropic, Google, and Microsoft itself simultaneously strengthen their positions.

What This Means

For the corporate AI market, this is the end of an era in which access to OpenAI was a strategic advantage for Azure. Competition now shifts from exclusivity to quality of integration, security, and ease of deployment: the winner will not be whoever has OpenAI models available, but whoever integrates them best into actual workflows and fastest turns models into working products for business in the coming years.