Core Viewpoint - Amazon Web Services (AWS) is adopting a differentiated strategy in the AI arms race by not betting on a single large language model (LLM) but instead creating a model marketplace that allows customers to choose their technology paths, thereby building unique barriers in the cloud computing battlefield [1][5] Group 1: AWS AI Strategy - The Bedrock service, launched two years ago, has become a growth engine for AWS, contributing over 18% of AWS's total revenue in Q1 of this year, with its core competitiveness lying in offering an "AI toolbox" consisting of over 100 models [1][3] - AWS's Vice President of Compute and Networking, Dave Brown, emphasized that no single model can meet all scenarios, a realization underscored by the rapid deployment of a managed version of the DeepSeek model on Bedrock within a week of its launch [1][2] Group 2: Investment and Partnerships - AWS's Bedrock was conceived in 2020, with a shift in understanding customer needs leading to the development of a multi-model choice platform, making AWS the first cloud service provider to offer such a feature [2] - To avoid over-reliance on a single model, AWS is employing a "self-research + investment" strategy, planning to double its investment in Anthropic to $8 billion by November 2024, which requires Anthropic to use only AWS chips for training its LLM Claude [2] Group 3: Hardware and Market Position - AWS is launching its own foundational model series named Nova by the end of 2024, allowing for deep collaboration with Anthropic while maintaining platform openness [3] - AWS is challenging Nvidia's dominance in the hardware space by allowing customers to choose between AWS's self-developed chips and those from Intel, AMD, and Nvidia, addressing industry pain points related to cost and performance [3] - AWS's Q1 revenue grew by 16.9% year-on-year to $29.27 billion, although it has fallen short of expectations for three consecutive quarters, indicating the urgency of strategic transformation [3] Group 4: Competitive Landscape - In the AI infrastructure competition, AWS is demonstrating a new competitive rule: the real victory may lie in providing developers with more freedom in technology combinations and more cost-effective computing solutions, rather than focusing solely on model supremacy [5]