AI视频公司Decart测试显示，Trainium3帮助视频生成帧率提升至其他芯片的四倍。尽管定制芯片崛起，但英伟达市场 地位短期难撼动。Trainium芯片首席架构师Ron Diamant表示："归根结底，主要优势在于性价比。" 又一家大型科技公司进军英伟达领地。 周二，亚马逊云服务(AWS)正式推出第三代定制AI芯片Trainium3，直接瞄准英伟达主导的GPU市场。 这家云计算巨头宣称，新芯片性能较上一代提升四倍，与同等GPU系统相比，可将AI模型训练和运行成本降低最多 50%。 近几个月，越来越多的AI企业正寻求供应商多元化，据报道，Meta Platforms正与谷歌洽谈采购数十亿美元的TPU芯 片，OpenAI则与英伟达竞争对手AMD及定制芯片设计商博通达成合作。 归根结底，主要优势在于性价比。 Diamant强调： 我们并不认为自己在试图取代英伟达。 但AWS周二公布的客户阵容显示其市场渗透力正在增强，除Anthropic外，还包括Karakuri、Metagenomi、 Neto.ai、理光和Splash Music等企业。 AWS早在2015年收购以色列初创公司Annapurna Labs， ...

Core Insights - Amazon Web Services (AWS) has launched its third-generation custom AI chip, Trainium3, targeting the GPU market dominated by Nvidia [1][2] - The new chip boasts a fourfold performance increase over its predecessor and can reduce AI model training and operational costs by up to 50% compared to equivalent GPU systems [2][3] - AWS's strategy focuses on cost-effectiveness rather than directly replacing Nvidia, indicating a shift towards a more diversified supplier landscape in the AI sector [3][6] Group 1: Product Launch and Performance - AWS's Trainium3 chip is designed to enhance performance and reduce costs for AI applications, with significant improvements noted by clients like Decart [2][4] - Decart reported that Trainium3 enabled a video generation frame rate four times higher than other chips, showcasing its potential for real-time rendering applications [4][5] Group 2: Market Dynamics - Despite the introduction of custom chips like Trainium3, Nvidia's market position remains strong, with many AWS clients also purchasing Nvidia products [6][7] - The trend of AI companies seeking supplier diversification is primarily aimed at risk mitigation and cost optimization rather than a complete shift away from Nvidia's technology [7]