Amazon pushes AI chip development to reduce dependence on Nvidia

Amazon is aggressively developing new artificial intelligence (AI) chips to maximize profits from its investments in the semiconductor industry and reduce its dependence on Nvidia, according to a report from Financial Times. This initiative by Amazon aims to improve data center performance and provide customers with more AI options on the cloud platform, with chips optimized for specific tasks, in contrast to Nvidia's multipurpose tools. Amazon is expected to launch the 'Trainium 2' AI training chip next month. This project is led by Annapurna Labs, which Amazon acquired in 2015 for $350 million. Currently, the Trainium 2 chip is being tested by Anthropic - a company that has received a $4 billion investment from Amazon, as well as partners like Databricks, Deutsche Telekom, Ricoh, and Stockmark. Dave Brown, Vice President of Compute and Networking Services at AWS, emphasized AWS's goal of providing an optimized environment to run Nvidia-based applications, but also recognized the importance of competition and alternative options to maintain a healthy and flexible industry. Amazon claims that its Inferentia AI chips can save up to 40% in operating costs when generating responses for AI models. Brown emphasized the cost impact, stating that while a 40% savings on $1,000 may not be significant, a 40% savings on tens of millions of dollars can make a substantial difference. Amazon plans to spend around $75 billion on technology infrastructure in 2024, up from $48.4 billion in 2023. CEO Andy Jassy predicts that spending may increase further in 2025. Rami Sinno, the Technical Director of Annapurna, emphasized that success comes not only from chip development but also from building the entire system. Amazon's approach includes building everything from silicon to server racks, all supported by proprietary software and architecture. Sinno added that scaling up this process is a major challenge that only a few companies in the world can achieve. Despite these significant efforts, the impact of Amazon on Nvidia's dominant position in the AI chip market remains limited. Nvidia reported revenue of $26.3 billion from AI data center chips in the second quarter of this year, equivalent to Amazon's entire AWS revenue during the same period. Join Telegram: https://t.me/tapchibitcoinvn Twitter (X): https://twitter.com/tapchibtc_io TikTok: https://www.tiktok.com/@tapchibitcoin Mr. Teacher According to CryptoBriefing