AWS customers can access OpenAI's latest models and its Codex coding agent through Amazon Bedrock, initially in limited preview
OpenAI announced a limited-preview launch of its newest AI models and Codex coding agent on Amazon $AMZN Web Services' Bedrock platform; broader availability for all AWS customers is expected within weeks.
Also introduced was Amazon Bedrock Managed Agents powered by OpenAI, which gives enterprise customers tools to deploy customized, production-ready agents inside AWS environments; among its capabilities is the ability to retain context from prior interactions.
"This is what our customers have been asking us for for a really long time," AWS CEO Matt Garman said at a launch event in San Francisco, according to CNBC.
A scheduling conflict kept OpenAI CEO Sam Altman from attending in person — he was across the Bay in Oakland, where court proceedings in his lawsuit against Elon Musk are ongoing — so he delivered his remarks in a pre-recorded video, according to CNBC. In it, he expressed enthusiasm for what the AWS partnership would mean for customers.
The AWS news followed by just one day a renegotiated deal between OpenAI and Microsoft $MSFT under which OpenAI gained the ability to put a ceiling on revenue-share obligations and offer its services through competing cloud platforms. In an internal memo sent earlier this month and reported by CNBC, OpenAI revenue chief Denise Dresser acknowledged to staff that while the Microsoft arrangement had been essential, it "has also limited our ability to meet enterprises where they are — for many that's Bedrock."
Before Tuesday's announcement, the only OpenAI offerings available to AWS developers were open-weight models, which had been added to the platform last August.
The AWS expansion builds on a broader deepening of ties between the two companies. A $38 billion commitment between the two companies was disclosed in November, and the relationship deepened further when Amazon pledged $50 billion to OpenAI roughly three months afterward; as part of that deal, OpenAI committed to tapping two gigawatts of capacity from AWS's proprietary Trainium chips for model training, according to CNBC.
