Companies

Amazon Web Services

Follow the latest coverage, related explainers and connected technology stories.

Latest coverage

CN
Amazon launches "Kiro bots" to enhance the autonomy of AI coding agents

Amazon launches "Kiro bots" to enhance the autonomy of AI coding agents

AWS has launched the new "Kiro bots" feature, which grants AI agents persistent memory and independent workflow management, aiming to reduce the need for developer intervention in managing coding sessions.

CN
Amazon: Record Growth in AWS Driven by AI and Infrastructure Expansion

Amazon: Record Growth in AWS Driven by AI and Infrastructure Expansion

Amazon announced a 39.3% revenue growth for AWS in the third quarter of 2026, with accelerating demand for AI services and expanded technical capabilities.

CN
Amazon S3 Arrives on Marketplace and v0 from Vercel to Support Storage, Analytics, and AI Applications

Amazon S3 Arrives on Marketplace and v0 from Vercel to Support Storage, Analytics, and AI Applications

Amazon S3, including S3 Tables and S3 Vectors, is now available directly through Vercel Marketplace and v0, enabling developers to create storage resources and connect them to their applications without manually configuring static access keys. The integration covers general-purpose storage, Apache Iceberg analytics, and vector search for AI applications.

CN
AWS Showcases Prime Day Figures and Introduces Updates for AI and Cloud Security

AWS Showcases Prime Day Figures and Introduces Updates for AI and Cloud Security

AWS revealed major operational metrics from Prime Day 2026, including DynamoDB processing more than 59 trillion requests. It also announced the availability of Claude Haiku 5.5 and Amazon Nova 2.5 Sonic, along with updates for database protection, Lambda functions, and Redshift.

CN
AWS Launches Lambda Web Functions to Run HTTP Applications Directly on Serverless Infrastructure

AWS Launches Lambda Web Functions to Run HTTP Applications Directly on Serverless Infrastructure

AWS announced Lambda Web Functions, a new execution model that enables the deployment of conventional HTTP servers built with frameworks such as Express and Next.js on Lambda without converting them into event handlers. The service supports synchronous execution, cross-Region distribution, and pricing based on processor usage time, with initial availability in five AWS Regions.

CN
AWS Launches Open-Source Toolchain for Building Perceptive and Adaptive Machines

AWS Launches Open-Source Toolchain for Building Perceptive and Adaptive Machines

AWS introduced the Physical AI Toolchain on AWS, an open-source framework that combines software architecture guidance, deployment automation, and ready-to-use code for developing robots and intelligent machines. The toolchain integrates with NVIDIA’s physical AI ecosystem and spans data generation and training through simulation, edge deployment, and continuous optimization.

CN
AWS Enables OpenAI-Powered Managed Bedrock Agents and Announces Changes to Its Cloud Services

AWS Enables OpenAI-Powered Managed Bedrock Agents and Announces Changes to Its Cloud Services

AWS announced a public preview of OpenAI-powered managed Amazon Bedrock Agents, running them within AWS and connecting them to existing identities, permissions, and governance. The updates included new models, data and artificial intelligence launches, and a timeline for ending or reducing the availability of several services.

CN
AWS Stops Using Non-Disclosure Agreements with Government Agencies Over Data Centers

AWS Stops Using Non-Disclosure Agreements with Government Agencies Over Data Centers

Matt Garman, CEO of Amazon Web Services, said the company no longer uses non-disclosure agreements with government agencies for data center projects, in an attempt to ease growing opposition in the United States. This comes amid demands for greater transparency and concerns about water and energy consumption, emissions, and the impact of these projects on electricity prices.

CN
California Court of Appeal Rejects Revival of Lawsuit Against Journalist Who Published Report on Former CEO’s Arrest

California Court of Appeal Rejects Revival of Lawsuit Against Journalist Who Published Report on Former CEO’s Arrest

The California Court of Appeal upheld the dismissal of a lawsuit filed by Maury Blackman, former CEO of Premise Data, against journalist Jack Poulson and platforms that hosted his report on a confidential arrest for criminal domestic violence. The decision relied on First Amendment protections and the state’s anti-SLAPP law, in addition to legal protections for the publisher and platforms.

CN
Amazon Aurora PostgreSQL Enables Direct Queries from Iceberg and Parquet Within the Data Lake

Amazon Aurora PostgreSQL Enables Direct Queries from Iceberg and Parquet Within the Data Lake

AWS has added to Amazon Aurora PostgreSQL the ability to combine live operational data with Apache Iceberg and Parquet data stored in Amazon S3 in a single PostgreSQL query, without ETL pipelines to move the data. The capability relies on embedded DuckDB and supports the AWS Glue Data Catalog and Iceberg catalogs compatible with IRC.

CN
Amazon S3 Vectors Adds Metadata Prefiltering to Improve Search Result Recall

Amazon S3 Vectors Adds Metadata Prefiltering to Improve Search Result Recall

AWS has added metadata prefiltering to Amazon S3 Vectors before similarity search, potentially increasing the recall rate of matching results by up to five times in queries with selective filters. The feature is available at no additional cost and without re-ingesting data, but it requires moving existing indexes to ENHANCED mode.

CN
Amazon S3 Tables Now Supports All Apache Iceberg V3 Data Types

Amazon S3 Tables Now Supports All Apache Iceberg V3 Data Types

AWS announced full support for the Apache Iceberg V3 specification in Amazon S3 Tables, including deletion vectors, row lineage, and new types for semi-structured and geospatial data. New V3 tables can be created or existing V2 tables can be upgraded, but the upgrade is one-way and requires access-engine compatibility.

CN
How AWS Supports STRADVISION’s Transition from Startup Idea to Physical AI Solutions

How AWS Supports STRADVISION’s Transition from Startup Idea to Physical AI Solutions

A MONOist interview with executives from AWS and STRADVISION examines the role of AWS startup programs and cloud infrastructure in developing computer vision and physical AI solutions. The company explains how it uses AWS services, including infrastructure dedicated to training and running models, to expand its applications in automotive and robotics.

CN
AWS Open-Sources Strands Harness for Building AI Agents Independent of a Single Language Model

AWS Open-Sources Strands Harness for Building AI Agents Independent of a Single Language Model

AWS has launched the open-source Strands Harness tool for building AI agents using models such as GPT, Claude, Gemini, and Ollama, with the ability to switch models and deploy the agent in any Linux container environment. The tool provides built-in capabilities for context and memory management, command execution, and access to files and the web.

CN
AWS Expands AI Options and Launches New Tools for Monitoring, Inference, and Agents

AWS Expands AI Options and Launches New Tools for Monitoring, Inference, and Agents

AWS presented a set of updates including the availability of GPT-6 Sol, GPT-6 Luna, and Claude Opus 5.5 through Amazon Bedrock, along with tools for monitoring agents, routing model inference, and building event-driven applications and software agents.

CN
Anthropic Launches Claude Opus 5.5 with Performance Comparable to Claude Fable 5.1 and 40% Lower Operating Cost

Anthropic Launches Claude Opus 5.5 with Performance Comparable to Claude Fable 5.1 and 40% Lower Operating Cost

Anthropic announced Claude Opus 5.5, the first model in the Claude 5.5 family, with improvements in coding, knowledge work, and safety, and an operating cost 40% lower than Opus 5. The model is immediately available through Anthropic, Amazon Web Services, Google Cloud, and Microsoft Azure, with additional restrictions and controls for critical-use cases and cybersecurity.

CN
AWS Expands Its Services Ecosystem with the Launch of the Builder Center App and Updates to AI, Java, and Cloud Computing

AWS Expands Its Services Ecosystem with the Launch of the Builder Center App and Updates to AI, Java, and Cloud Computing

AWS announced the global availability of the Builder Center mobile app, along with the general availability of Connect Talent, Amazon Corretto 27, and Kimi K3 on Bedrock, as well as updates to EC2, Elastic Beanstalk, and the onboarding experience for new users.

CN
AWS Launches Low-Cost EC2 T8i Instances with Burstable Performance

AWS Launches Low-Cost EC2 T8i Instances with Burstable Performance

AWS announced the general availability of Amazon EC2 T8i instances powered by sixth-generation Intel Xeon Scalable processors, offering up to 30% better price performance compared with T3 instances. The instances target low- to moderate-CPU-utilization workloads while retaining the familiar CPU credit model.

CN
AWS Presents Its Vision for AI and Unveils World Models Running on Trainium3

AWS Presents Its Vision for AI and Unveils World Models Running on Trainium3

During its global meeting on September 17, 2026, AWS presented a vision for building AI tools along the lines of cloud-computing architecture, showcasing Odyssey-3 and Agora-2 for modeling shared worlds. It also highlighted startups using Trainium in areas including robotics, code generation, and interactive environments.

CN
AWS Launches Cluster Mode in Elastic Beanstalk to Manage Multiple Applications Through Managed EKS Infrastructure

AWS Launches Cluster Mode in Elastic Beanstalk to Manage Multiple Applications Through Managed EKS Infrastructure

AWS announced Cluster Mode in Elastic Beanstalk, a managed operating mode that deploys, scales, monitors, and updates applications on Amazon EKS infrastructure without the team directly managing the underlying infrastructure. The mode enables resource sharing among multiple applications, but adds EKS costs and related AWS resources and is not suitable for every workload.