Amazon Web Services (AWS), a subsidiary of Amazon.com, has announced a collaboration with Nvidia to make available graphics processing units based on the new Blackwell architecture in Amazon’s cloud services.
Amazon web services will utilize the GB200 Grace chips and B100 graphics processing units, expanding the long-term strategic collaboration between the two companies to provide more secure and advanced infrastructure and software to help customers develop new capabilities for cutting-edge artificial intelligence.
Nvidia and AWS have continued to integrate their best technologies, such as Nvidia’s Blackwell-based system and enhanced AI programs, along with advanced security services like AWS KMS and Amazon Web Services’ Nitro system, in addition to petabit EFA networks and Amazon EC2 UltraCluster for flexible cloud computing.
Both companies provide the infrastructure and tools that enable customers to create and run large language models in real-time, consisting of trillions of parameters.
These parameters are numerical values that define how the neural network processes data, at a lower cost than previous-generation graphics processing units in the Amazon EC2 service.
Adam Selipski, CEO of Amazon Web Services, stated that the long-standing collaboration between the two companies began over 13 years ago when they launched the world’s first cloud graphics processing unit.
Today, they offer a wide range of Nvidia graphics processing unit solutions to customers. Nvidia’s next-generation Grace Blackwell processor is a significant step in the field of cutting-edge artificial intelligence and graphics processing unit computing.
When integrated into the high-performance computing of Amazon EC2 UltraClusters and the unique Nitro system, they enable customers to quickly and securely create and run large language models in the cloud. They continue to innovate to make Amazon Web Services the best place to run Nvidia graphics processing units in the cloud.
Jensen Huang, founder and CEO of Nvidia, stated that artificial intelligence is experiencing unprecedented momentum, leading to the emergence of new applications and business models in various industries. He emphasized that collaborating with Amazon Web Services accelerates the capabilities of new generative technology, providing customers with massive computational power at the same time.
Building on the success of the EC2 P5 systems powered by the H100 graphics processing unit, Amazon Web Services plans to introduce EC2 systems with the new B100 processors that are being deployed in EC2 UltraCluster services to accelerate training and deliver insights for wide-scale artificial intelligence generative processes.
The GB200 graphics processing units will be available on the Nvidia DGX Cloud platform, designed in collaboration with Amazon Web Services, to provide direct access for developers in the professional sector to the necessary infrastructure and software for building and deploying advanced artificial intelligence models.
