亚马逊表示，这项为期五年的协议(条款未披露)可能会让公司花更少的钱来开发自己的人工智能能力。According to reports, amazon has signed a five-year agreement with Databricks, a competitor of Snowflake, allowing this data warehouse company to use amazon's Trainium ai chip.
According to the WiseTongue Finance APP, according to reports,$Amazon (AMZN.US)$has$Snowflake (SNOW.US)$signed a five-year agreement with its competitor Databricks, allowing this data warehouse company to use amazon's Trainium ai chip.
This collaboration mainly revolves around amazon's Trainium ai chip, which provides companies wishing to customize or build ai models.$NVIDIA (NVDA.US)$A cheaper alternative to GPUs.
Amazon stated that this five-year agreement (terms undisclosed) may allow the company to spend less money on developing its own AI capabilities.
Databricks executive Naveen Rao said in an interview that collaborating with Amazon will make AI faster and cheaper for enterprises, as it enables Databricks to pass on the cost savings from using Amazon's AI chips instead of NVIDIA chips, which dominate the AI accelerator market.
Databricks hopes to pass on the cost savings from using Amazon chips to customers in order to break NVIDIA's moat.
As part of the agreement, Databricks will also increase its use of NVIDIA GPUs rented through AWS. According to Rao, AWS has generated over $1 billion in revenue for Databricks and remains the company's fastest-growing cloud partner.
Dave Brown from Amazon Web Services stated that the Trainium chip designed specifically for ai tasks can help enterprises reduce ai development costs by 40%.
It is understood that Amazon released the second version of its Trainium chip in November 2023. The company also has a range of ai chips called Inferencia, which are used to build and leverage ai models.
Editor/ping
