Amazon is already offering new OpenAI products on AWS

```json { "title": "Amazon AWS Now Offers OpenAI Models on Bedrock", "metaDescription": "AWS launched OpenAI models, Codex, and Managed Agents on Amazon Bedrock in limited preview on April 28, 2026, one day after Microsoft dropped its exclusivity.", "content": "<h2>Amazon Brings OpenAI Models to AWS Bedrock One Day After Microsoft Drops Exclusivity</h2><p>Amazon Web Services and OpenAI expanded their strategic partnership on April 28, 2026, launching three new offerings on Amazon Bedrock in limited preview: OpenAI models on Amazon Bedrock, Codex on Amazon Bedrock, and Amazon Bedrock Managed Agents powered by OpenAI. The announcement, made at a launch event in San Francisco by AWS CEO Matt Garman, came just 24 hours after Microsoft and OpenAI restructured their long-standing partnership to make Microsoft's license to OpenAI intellectual property non-exclusive — clearing the legal path for OpenAI to distribute its products across any cloud provider, including AWS and Google Cloud.</p><p>The timing was not coincidental. OpenAI's flagship frontier model GPT-5.4 became immediately available in limited preview on Amazon Bedrock, with GPT-5.5 slated to follow shortly. According to CNBC, the new AWS services are expected to become generally available within the next few weeks.</p><h2>What's Now Available on Amazon Bedrock</h2><p>The April 28 launch introduced three distinct offerings for enterprise customers on Amazon Bedrock, all currently in limited preview.</p><p><strong>OpenAI Models on Amazon Bedrock</strong> gives customers access to OpenAI's frontier models — beginning with GPT-5.4 immediately and GPT-5.5 as the flagship model in the near term — through the same Bedrock interface they already use to access models from Anthropic, Meta, Mistral, Cohere, Amazon, and other providers. Enterprise controls including IAM, AWS PrivateLink, guardrails, encryption, and CloudTrail logging apply to OpenAI model usage by default.</p><p><strong>Codex on Amazon Bedrock</strong> brings OpenAI's coding agent to AWS infrastructure. More than 4 million people use Codex every week, according to OpenAI's April 28, 2026 announcement. Customers can apply Codex usage toward their existing AWS cloud commitments.</p><p><strong>Amazon Bedrock Managed Agents powered by OpenAI</strong> is a new service designed to help enterprises deploy production-ready AI agents. This offering is particularly significant because AWS had been named the exclusive third-party cloud distribution provider for OpenAI Frontier — OpenAI's enterprise agent platform — under the February 2026 strategic partnership. The previous Microsoft exclusivity arrangement had created legal uncertainty about whether that AWS-exclusive promise was permissible. The revised Microsoft-OpenAI deal resolved that conflict.</p><p>According to the official AWS announcement, customers can evaluate and deploy OpenAI models alongside models from other providers through a unified interface with consistent security, governance, and cost controls. OpenAI model usage and Codex usage can both be applied toward existing AWS cloud commitments.</p><h2>The Microsoft Deal That Made This Possible</h2><p>The AWS announcement would not have been possible under the terms of OpenAI's previous agreement with Microsoft. According to TechCrunch, OpenAI's original agreement with Microsoft had prevented OpenAI from selling its Frontier agent-making tool exclusively on AWS — creating a direct conflict with the February 2026 Amazon partnership.</p><p>On April 27, 2026, Microsoft and OpenAI restructured their partnership. According to Bloomberg, Microsoft and OpenAI agreed to drop Microsoft's exclusive right to sell OpenAI's AI models. In exchange, according to Bloomberg, Microsoft will no longer pay a revenue share on OpenAI products it resells on its cloud. According to Microsoft's official blog and OpenAI's official statement, Microsoft will continue to hold a license to OpenAI intellectual property for models and products through 2032, but that license is now non-exclusive.</p><p>Microsoft has invested more than $13 billion in OpenAI since 2019, according to CNBC. As of October 2025, Microsoft's investment in OpenAI's for-profit arm was valued at approximately $135 billion, or roughly 27% of the company on an as-converted diluted basis. OpenAI is required to pay Microsoft a revenue share of 20% through 2030, now subject to an undisclosed cap, according to Dataconomy.</p><p>According to CNBC, an internal memo from OpenAI's revenue chief described the prior arrangement as having "limited our ability to meet enterprises where they are."</p><p>With the revised agreement in place, OpenAI is now free to distribute its products across any cloud platform. According to Axios, Google is studying how it might be able to work with OpenAI given the revised deal with Microsoft.</p><h2>The Broader Amazon-OpenAI Partnership</h2><p>The April 28 Bedrock launch builds on a substantial financial and infrastructure relationship between Amazon and OpenAI that was formalized in February 2026. According to aboutamazon.com, Amazon agreed to invest up to $50 billion in OpenAI as part of that strategic partnership, with OpenAI committing to use two gigawatts of AWS's custom Trainium chips for training AI models.</p><p>According to the AWS Weekly Roundup blog from March 2, 2026, OpenAI and AWS had already expanded their existing $38 billion multi-year agreement by $100 billion over eight years, with OpenAI committing to consume approximately 2 gigawatts of Trainium capacity. Under the February 2026 agreement, AWS was designated as the exclusive third-party cloud distribution provider for OpenAI Frontier, OpenAI's enterprise platform.</p><p>The scale of the infrastructure commitment underscores how central AWS has become to OpenAI's operations — and why the resolution of the Microsoft exclusivity issue mattered practically, not just strategically.</p><h2>What Executives Said</h2><p>AWS CEO Matt Garman, speaking at the San Francisco launch event, framed the announcement around customer demand. "This is what our customers have been asking us for for a really long time," Garman said, according to CNBC.</p><p>Amazon CEO Andy Jassy had telegraphed the news on April 27, 2026, posting on X: "We're excited to make OpenAI's models available directly to customers on Bedrock in the coming weeks, alongside the upcoming Stateful Runtime Environment." In a statement published on aboutamazon.com, Jassy added: "We have lots of developers and companies eager to run services powered by OpenAI models on AWS."</p><p>OpenAI CEO Sam Altman was not physically present at the San Francisco event. According to CNBC, Altman was in federal court in Oakland defending OpenAI in a lawsuit brought by Elon Musk. He sent a recorded video message to the event instead. "I wanted to say thank you to Matt and the whole AWS team," Altman said in the recorded message, according to CNBC. In a separate portion of his remarks, Altman said: "AWS is one of the companies that change what builders can do," according to Axios.</p><p>In a statement published on aboutamazon.com, Altman said: "Combining OpenAI's models with Amazon's infrastructure and global reach helps us put powerful AI into the hands of businesses and users at real scale."</p><h2>Industry Implications</h2><p>The AWS-OpenAI availability announcement reshapes the competitive dynamics of the enterprise AI cloud market in a meaningful way. For much of the past several years, Microsoft Azure held a structural advantage as the primary — and in many cases exclusive — commercial distribution channel for OpenAI's most capable models. That advantage no longer exists in its previous form.</p><p>Enterprise customers who have standardized on AWS infrastructure can now access OpenAI's frontier models without switching clouds or managing separate vendor relationships. The unified security and governance controls — IAM, AWS PrivateLink, guardrails, encryption, and CloudTrail logging — mean that adding OpenAI models to an existing AWS environment does not require new compliance architectures. The ability to apply OpenAI model usage toward existing AWS cloud commitments further reduces friction for large organizations with established AWS spending agreements.</p><p>The Managed Agents offering is particularly notable for enterprise AI deployments. The ability to run production-ready AI agents on AWS infrastructure, governed by AWS's existing enterprise controls, addresses one of the persistent concerns that has slowed agent adoption in regulated industries: the question of who is responsible for security, logging, and compliance when an AI system takes autonomous actions.</p><p>According to The Register, AWS also announced alongside the OpenAI deal a new personalized assistant called Quick and new versions of its Connect product, expanding to HR, health care, and supply chain management — signaling that AWS is simultaneously moving to compete with business software providers, not just serve as infrastructure for them.</p><p>According to Axios, Google is now studying how it might work with OpenAI under the revised Microsoft terms. If that materializes, it would mean OpenAI's models are available across all three major cloud platforms — a fundamentally different distribution model than existed even a week ago.</p><h2>What Comes Next</h2><p>According to CNBC, the new AWS services are expected to become generally available within the next few weeks. Currently, OpenAI models on Amazon Bedrock, Codex on Amazon Bedrock, and Amazon Bedrock Managed Agents powered by OpenAI are all in limited preview.</p><p>GPT-5.4 is immediately accessible in the limited preview. GPT-5.5, described in OpenAI's official announcement as the flagship frontier model for the Amazon Bedrock partnership, is listed as coming soon. The timeline for general availability has not been specified beyond "the next few weeks," per CNBC's April 28, 2026 reporting.</p><p>Andy Jassy's post on X also referenced an "upcoming Stateful Runtime Environment" alongside the OpenAI model availability announcement, though details on that offering were not disclosed in the launch materials reviewed for this article.</p><p>For enterprises currently evaluating AI infrastructure decisions, the limited preview period represents an opportunity to test OpenAI model performance within AWS environments before committing to production deployments. Organizations interested in the Managed Agents offering in particular may want to assess how it integrates with their existing AWS governance frameworks ahead of general availability.</p><p>For more tech news, visit our <a href=\"/news\">news section</a>.</p><h2>Why This Matters for Health and Productivity</h2><p>The arrival of OpenAI's most capable models on Amazon Bedrock — with enterprise-grade security, unified governance, and agent capabilities built in — accelerates the timeline for deploying AI in industries where data sensitivity has been a barrier. Health care, which AWS specifically named as an expansion target for its Connect product line, stands to benefit from AI tools that meet enterprise compliance requirements out of the box. For knowledge workers and teams already operating on AWS infrastructure, access to frontier AI models and coding agents through a single, governed platform reduces the operational complexity that has historically slowed AI adoption. Join the <a href=\"/#waitlist\">Moccet waitlist</a> to stay ahead of the curve.</p>", "excerpt": "Amazon Web Services and OpenAI launched three new offerings on Amazon Bedrock in limited preview on April 28, 2026, including OpenAI frontier models, Codex, and Managed Agents powered by OpenAI. The announcement came one day after Microsoft restructured its partnership with OpenAI, making its license to OpenAI intellectual property non-exclusive and clearing the path for OpenAI to distribute its products across any cloud platform.", "keywords": ["OpenAI Amazon Bedrock", "AWS OpenAI models", "Amazon Bedrock AI", "OpenAI GPT-5", "enterprise AI cloud"], "slug": "amazon-aws-openai-models-bedrock-2026" } ```