Core Viewpoint - Amazon has heavily invested in Anthropic, committing $8 billion, and is promoting its AI technology through AWS, but AWS is facing backlash due to limitations on Anthropic's AI capabilities [1] Group 1: AWS Bedrock Issues - AWS executives referred to the capacity issues with Bedrock as a "catastrophic incident" due to arbitrary limits on API usage for Anthropic models, leading dissatisfied customers to purchase directly from Anthropic [2] - Customers are reporting increased frequency and severity of errors when using Bedrock, with some clients limited to five calls per minute for specific models, while Anthropic's own API allows up to 50 calls per minute at the same price point [5][6] - AWS has implemented rate limits on Bedrock to ensure fair access to popular AI model resources, but this has raised concerns about its ability to support future cloud application clients [3][7] Group 2: Competitive Landscape - Google, a competitor of AWS, has also invested in Anthropic and may benefit from AWS's issues as customers shift to using Anthropic's API directly [2][3] - AWS's leadership has recognized the stakes involved, restructuring management to better address Bedrock's challenges under the EC2 cloud server business [9] Group 3: Customer Responses and Alternatives - Some AWS customers are creating multiple accounts to bypass Bedrock's API access limits, while others are using a service called Provisioned Throughput to secure additional resources [10] - Despite the issues with Bedrock, many customers still trust AWS over Anthropic for data security, with AWS's AI business generating over $2 billion annually [7] Group 4: Future Outlook - Anthropic is reportedly investing in more servers alongside AWS and Google, suggesting that current issues with Bedrock may be temporary [6] - AWS has been slow to introduce features that Anthropic has already implemented, which has led some customers to switch to Anthropic's API for better functionality [11]
速递|AWS百亿豪赌或遭反噬:Anthropic API争夺战,与谷歌云的反向收割