Amazon rushes into the AI Agent battlefield

Author | Huang Yu

Editor | Wang Xiaojun

In 2025, regarded as the "Year of AI Agents," the competition for Agents is in full swing. Global cloud providers are also making efforts to become the preferred "black land" for developers to build AI Agents.

On July 17, at the Amazon Web Services summit in New York, Amazon Web Services, as the world's largest cloud provider, officially launched a series of new capabilities and tools for AI Agents, the most important of which is Amazon Bedrock AgentCore, an enterprise-level AI Agent building toolkit.

Swami Sivasubramanian, Vice President of Agentic AI at Amazon Web Services, stated that AI Agents, as autonomous software systems that leverage AI for reasoning, planning, adjusting, and completing tasks, will significantly accelerate innovation across various industries and enhance overall productivity.

"This is a profound transformation across multiple dimensions," he pointed out. AI Agents have disrupted the way software is built and have also brought new challenges for deployment and operation. The more far-reaching potential impact is that it changes the way software interacts with the world and reshapes the way humans interact with software.

To embrace this transformation, Amazon Bedrock AgentCore is a powerful tool launched by Amazon Web Services.

According to reports, Amazon Bedrock AgentCore aims to provide a complete set of capabilities for deploying and efficiently operating AI Agents (customers can build using any framework and model).

Amazon Web Services has recognized that as enterprises accelerate their embrace of AI Agents, they face key challenges: the systems they build must autonomously operate across digital boundaries while meeting the security, reliability, and governance standards required for enterprise-level deployment.

It is reported that Amazon Bedrock AgentCore can help developers bridge the critical gap between proof of concept and production deployment for AI Agents. Through a composable capability system, enterprises can transition agents from prototypes to applications, scaling to millions of end users.

At the same time, Amazon Bedrock AgentCore has seven core capabilities, covering runtime, memory, identity verification, gateways, code interpreters, browser tools, and visual dashboards.

For example, AgentCore Runtime supports low-latency interactive experiences and can handle complex asynchronous workloads lasting up to 8 hours, a duration that is industry-leading. It is also currently the only runtime environment that supports the use of various open-source and commercial frameworks while achieving complete session isolation.

Additionally, AgentCore Memory provides industry-leading long-term and short-term memory accuracy, making it more efficient for developers to build context-aware Agents; AgentCore Identity also offers seamless and secure Agent identity verification capabilities, which can integrate with Amazon Cognito, Integration of identity services such as Microsoft Entra ID and Okta.

Clearly, through the Amazon Bedrock AgentCore "smart brain," Amazon Web Services aims to lower the threshold for enterprises to develop agents and achieve large-scale deployment and operation of AI agents.

The future market space for AI agents is vast. Not only Amazon Web Services, but also Google Cloud, Tencent Cloud, Alibaba Cloud, and others have previously provided related solutions.

Minsheng Securities released a research report stating that it firmly believes 2025 will be the inaugural year for AI agents and the starting point for a software revolution: Agents may become an important catalyst for the revaluation of software, and the target market for software vendors is expected to expand to trillions of dollars in the labor market. AI agents are also expected to enhance the consumption attributes of software, further opening up the valuation ceiling for software vendors.

CICC believes that the innovation of product forms like AI agents is equally important as the evolution of underlying models, and it has shown strong potential for commercialization and implementation. Looking ahead, AI agents may reshape the internet ecosystem by taking over mobile devices.

The demand for AI agent development is bound to grow increasingly strong, and Amazon Web Services has also proposed the vision of "building the world's most practical AI agents platform." With the empowerment of global cloud giants, this battlefield will see more players emerge