OpenAI and Amazon announce strategic partnership

OpenAI 和 Amazon ( NASDAQ: AMZN ) 今日宣布达成一项多年期战略合作，旨在加速面向企业、初创公司与终端用户的人工智能创新。作为交易一部分， Amazon 将向 OpenAI 投资 500 亿美元，首笔 150 亿美元将立即到位，其后在满足一定条件时再追加 350 亿美元。

双方将联合推出面向生产级的生成式 AI 应用与智能代理的开发环境。具体来看， OpenAI 与 Amazon Web Services ( AWS ) 正在共同开发一个由 OpenAI 模型驱动的 Stateful Runtime Environment，该环境将通过 Amazon Bedrock 向 AWS 客户提供。所谓 Stateful 开发者环境，是下一代前沿模型使用方式，它能无缝访问算力、内存与身份等资源，保持上下文、记忆先前工作、在多个软件工具与数据源间协同并调用计算资源，适合长期项目与持续工作流。该 Stateful Runtime Environment 将针对 AWS 基础设施进行优化，并与 Amazon Bedrock 的 Bedrock AgentCore 及其他基础服务集成，确保客户的 AI 应用与代理能与其在 AWS 上运行的其他业务系统协同一致，预计将在未来数月内推出。

在企业级平台方面， AWS 将成为 OpenAI Frontier 的独家第三方云分发商，借此扩大 OpenAI 最先进企业平台的可及性，以应对各行业加速的 AI 部署需求。 OpenAI Frontier 可让组织构建、部署并管理能够在真实业务系统中运行、共享上下文且内置治理与企业级安全的 AI 代理团队，而无需自行维护底层基础设施。当企业从试验走向量产时， Frontier 有助于迅速、安全且全球尺度地将强大 AI 集成进现有工作流程。

在算力承诺上， OpenAI 与 AWS 将把已有的 380 亿美元多年合作协议在未来八年内再扩大 1000 亿美元。扩展内容包括 OpenAI 承诺通过 AWS 基础设施长期消耗约 2 吉瓦（gigawatts）的 Trainium 算力，以满足对 Stateful Runtime、 Frontier 及其他高阶工作负载的需求。这一安排既能降低大规模智能生产成本，也能提升效率。该承诺涵盖 Trainium3 与下一代 Trainium4 芯片； Trainium4 预计自 2027 年起交付，届时将带来显著的性能提升——包括更高的 FP4 计算性能、更宽的内存带宽与更大的高带宽内存容量，以支持更强大的大规模 AI 系统。

双方还将合作开发供 Amazon 客户端应用使用的定制模型。 Amazon 团队将可对 OpenAI 模型进行定制，用于直接面向客户的 AI 产品与代理，这些能力将补充 Amazon 已向开发者提供的模型（包括 Amazon 的 Nova family ），为团队在规模化构建与交付时提供更多工具选择。

OpenAI 联合创始人兼 CEO Sam Altman 表示：“ OpenAI 与 Amazon 共有一项信念：AI 应该以切实有用的方式惠及人们。将 OpenAI 的智能与 Amazon 的基础设施和全球覆盖相结合，能把强大的 AI 真正带到企业和用户手中，达到实际规模的应用。”

News:

• Amazon Web Services (AWS) and OpenAI will co-create a Stateful Runtime Environment powered by OpenAI models, available on Amazon Bedrock for AWS customers to build generative AI applications and agents at production scale.
• AWS will be the exclusive third-party cloud distribution provider for OpenAI Frontier, which enables organizations to build, deploy, and manage teams of AI agents.
• OpenAI to consume 2 gigawatts of Trainium capacity through AWS infrastructure to support demand for Stateful Runtime Environment, Frontier, and other advanced workloads.
• OpenAI and Amazon will develop customized models available to power Amazon’s customer-facing applications.
• Amazon will invest $50 billion in OpenAI.

OpenAI and Amazon (NASDAQ: AMZN) today announced a multi-year strategic partnership to accelerate AI innovation for enterprises, startups, and end consumers around the world. Amazon will also invest $50 billion in OpenAI, starting with an initial $15 billion investment and followed by another $35 billion in the coming months when certain conditions are met.

Partnering to bring new advanced AI capabilities to enterprises worldwide

OpenAI and Amazon are jointly developing a Stateful Runtime Environment powered by OpenAI’s models, which will be available through Amazon Bedrock.

Stateful developer environments are the next generation of how frontier models will be used, seamlessly enabling models to access elements like compute, memory, and identity. A Stateful Runtime Environment allows developers to keep context, remember prior work, work across software tools and data sources, and access compute. They're designed to handle ongoing projects and workflows.

These stateful developer environments will be trained to run optimally on AWS's infrastructure and integrated with Amazon Bedrock AgentCore and infrastructure services so customers’ AI applications and agents run cohesively with the rest of their infrastructure applications running in AWS. The Stateful Runtime Environment is expected to launch in the next few months.

Bringing OpenAI’s most advanced enterprise platform to AWS customers

AWS will serve as the exclusive third-party cloud distribution provider for OpenAI Frontier, expanding access to OpenAI’s most advanced enterprise platform as demand for AI deployment accelerates across industries.

Frontier enables organizations to build, deploy, and manage teams of AI agents that operate across real business systems with shared context, built-in governance, and enterprise-grade security, without managing underlying infrastructure. As companies move from experimentation to production AI, Frontier makes it straightforward to integrate powerful AI into existing workflows quickly, securely, and at global scale.

OpenAI to use Trainium compute to power growing Amazon customer demand

OpenAI and AWS are expanding their existing $38 billion multi-year agreement⁠ by $100 billion over 8 years. The expansion includes OpenAI committing to consume approximately 2 gigawatts of Trainium capacity through AWS infrastructure, which will support demand for Stateful Runtime, Frontier, and other advanced workloads. This agreement lowers the cost and improves the efficiency of producing intelligence at scale.

Under this structure, OpenAI secures long-term capacity while working with AWS to deploy purpose-built silicon alongside its broader compute ecosystem, enabling enterprises to consume intelligence on demand without managing underlying infrastructure.

This commitment spans both Trainium3 and next-generation Trainium4 chips and will power a broad range of advanced AI workloads. Trainium4, expected to begin delivery in 2027, will provide another major performance gain, including significantly higher FP4 compute performance, expanded memory bandwidth, and increased high-bandwidth memory capacity to support increasingly capable AI systems at scale.

Custom models available to power Amazon’s customer-facing applications

OpenAI and Amazon will collaborate to develop customized models available to Amazon developers to power Amazon’s customer-facing applications. Amazon teams will be able to tailor OpenAI models for use across AI products and agents that serve customers directly. These capabilities will complement the models already available to Amazon developers, including Amazon’s Nova family, offering another tool for teams to build and deliver at scale.

“OpenAI and Amazon share a belief that AI should show up in ways that are practical and genuinely useful for people. Combining OpenAI’s intelligence with Amazon’s infrastructure and global reach helps us put powerful AI into the hands of businesses and users at real scale.”

Generated by RSStT. The copyright belongs to the original author.

Source