Why has Amazon Web Services become the top AI development platform in China?

Text/Yang Jianyong

With generative AI on the stage of the times, enterprises have taken advantage of the trend and accelerated their upgrade to intelligence. As a result of the embrace of generative AI, generative AI will reshape the entire business of enterprises, whether it is customer experience, efficiency improvement, rapid decision-making and creative innovation.

As a global leader in cloud computing, Amazon Web Services injects innovation vitality into all walks of life, from IaaS infrastructure around the world to rich cloud services and ecosystems, whether it is artificial intelligence, machine learning, serverless, etc., its breadth and depth of services are far ahead of the industry.

Recently, the latest "2023 China AI Development Platform Market Report" released by Frost & Sullivan (Sullivan), a world-renowned market research institution, and Toubao Research Institute confirms Amazon Web Services' leading position in AI technology promotion and services. Amazon Web Services' outstanding performance and innovation capabilities in the field of artificial intelligence have won the "top spot" of the report for the third consecutive year.

Not only is it at the top of the Leaders Quadrant in both the Innovation Index and the Growth Index, but Amazon Web Services also received the highest scores in six key indicators: data labeling, model development capabilities, model management capabilities, data preprocessing capabilities, algorithm management capabilities, and model deployment capabilities.

In the past three years, why has Amazon Web Services become the top AI development platform in China? With the advent of the generative AI era, Amazon Web Services has proposed a three-layer technology stack in the field of generative AI, including an infrastructure layer for basic model training and inference, a tool layer for building with basic models, and an application layer for building with basic models.

First, the underlying capabilities. Amazon Web Services provides infrastructure for basic model training and inference, including self-developed chips and chips from AMD, Intel, and Nvidia, as well as related services, including Amazon SageMaker.

Take the unique advantage of Amazon Web Services - self-developed chips as an example. At the moment of computing power anxiety, self-developed chips have become the new focus of competition among global technology giants, and Amazon Web Services has been innovating continuously on this track for more than ten years, and each generation of self-developed chips has greatly improved the cost performance and energy efficiency of cloud workloads. The world's leading generative AI companies, including Anthropic, are also leveraging Amazon Trainium training chips and Amazon Inferentia inference chips to build, train, and deploy their future foundational models.

In terms of model deployment capabilities, Amazon SageMaker Model Deployment supports a variety of inference service deployment modes, including serverless inference, single-node multi-model, and asynchronous inference, which is suitable for various inference scenarios with high performance and low cost. Amazon SageMaker provides comprehensive inference service monitoring, including invocation, resource utilization, and model behavior analysis, as well as model hot updates to help optimize model performance. Amazon SageMaker Edge Manager and Amazon SageMaker Neo also support model deployment and optimization for edge devices, enhancing deployment flexibility and efficiency.

Second, in the middle tier of tooling, Amazon Bedrock is the easiest way to build and scale generative AI applications with large models.

Amazon Bedrock is a fully managed generative AI service that provides high-performance foundational models from leading AI companies, including AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and Amazon, to name a few, as well as a set of capabilities for enterprises to build generative AI applications, simplifying development while ensuring privacy and security.

In the view of Amazon Web Services, the implementation of generative AI by enterprises is not only about the model, but also needs tools suitable for users, and at the same time, in the process of landing, different application scenarios must have different tools to do it. Amazon Web Services from the infrastructure to generative AI, and then to the underlying data platform that supports generative AI, so that everyone can really easily put the concept of generative AI in the enterprise, have their own generative AI applications, and have their own large models, so that every enterprise can use generative AI to innovate.

With the help of Amazon Web Services, Kingsoft Office is using a series of services, including Amazon Bedrock in overseas regions, to accelerate the implementation of generative AI technology in its products, and continue to enhance users' smart office experience. As one of the first users of Amazon Bedrock, Kingsoft Office uses Amazon Bedrock to conduct comprehensive tests on various office application scenarios, covering innovative scenarios such as text polishing, text rewriting, grammar checking, and language translation.

Finally, at the top-level application layer, applications built using basic models, including Amazon Q and Amazon CodeWhisperer, are provided.

Amazon Web Services is committed to reducing the cost of enterprises to implement generative AI, iterating models and deploying applications at scale with high efficiency and low cost, and helping enterprises build their own unique competitiveness in the era of generative AI with end-to-end data processing capabilities Bedrock helps enterprises choose the most appropriate model and build method for their business scenarios, provides out-of-the-box solutions to improve the ease of use of generative AI, and is committed to helping enterprises build responsible AI.

Amazon Web Services has accumulated deep global experience in AI development platform services, and is deeply involved in China, helping enterprises of all industries and sizes accelerate the innovation of generative AI applications. In this regard, Li Qing, senior director of Sullivan and vice president of Toubao Research Institute, said: "Amazon Web Services is a leader in China's AI development platform, which not only provides users with the ability to manage the full lifecycle of AI development, but also launches a complete generative AI technology stack to help users simplify the development of generative AI applications while ensuring the security and privacy of enterprises. Amazon Web Services can provide diversified models and enhanced data customization capabilities, support a wide range of customer bases with advanced cloud infrastructure, and deepen the Chinese market, helping enterprises quickly apply artificial intelligence and machine learning. ”

Yang Jianyong, a writer for Forbes China, expresses his views on an individual basis. It is committed to in-depth interpretation of cutting-edge technologies such as the Internet of Things, cloud services, artificial intelligence and smart home.