The aftershock of DeepSeek continues to be felt in the financial market and the news. While several governments around the world seek to ban the Chinese-made AI model, at least for government employees, companies are introducing the model to their own services. Among them, is American e-commerce giant Amazon.
On 5 February, it was reported that Amazon had introduced the latest R1 model of DeepSeek to its Bedrock Marketplace and Sagemaker Jumpstart. The model is used for inference. Users can build, experiment and scale their own generative AI ideas on AWS (Amazon Web Service) using DeepSeek-R1. Matt Garman, CEO of AWS says that introducing DeepSeek is an example of the company providing the most cutting-edge AI models. AWS also enables users to test DeepSeek AI in various ways.
AI services have been hugely profitable for AWS, Bedrock completed 97% of its annual sales target by July last year, and sales surpassed targets in Europe, the Middle East, Africa and Asia-Pacific. Amazon’s cloud services grew 20% year-on-year (YoY) in the first 3 quarters of 2024, and growth in Amazon AI services reached a 3-digit percentage, boosting the company’s share prices to a historic high at 212.25 USD. Amazon is also reportedly getting ready to unveil its Olympus AI model.
Similarly, Chinese tech companies and platforms Baidu, Huawei, JD.com, Tencent and Alibaba also deployed DeepSeek models on their cloud services. The release of the model boosted stocks at the Hong Kong Stock Exchange after the Chinese New Year (CNY) break on Monday and Tuesday, after causing the Nasdaq to drop on Monday 3 February.
Meanwhile, Governments and watchdogs in Europe, Australia and the United States are banning DeepSeek in their organisations. South Korean state-owned energy corporations also banned the model. With private enterprises open to incorporating the open-sourced model and experimenting with it, we might soon see an evidence-based conclusion on its safety and security.
