Amazon’s AMZN cloud-computing arm, Amazon Web Services (AWS), strengthens its ties with NVIDIA NVDA to invigorate its collection with generative AI technology.
NVIDIA’s Blackwell GPU platform, introduced at GTC 2024, will now be accessible on AWS. At its core is the GB200 NVL72, hosting 72 Blackwell GPUs and 36 Grace CPUs linked by fifth-generation NVIDIA NVLink.
This leap will accelerate inference workloads for resource-heavy, multi-trillion parameter language models.
AWS is dedicated to empowering customers to explore new generative AI capabilities by furnishing secure and advanced infrastructure, software, and services. As part of this initiative, AWS will present NVIDIA’s GB200 Grace Blackwell Superchip and B100 Tensor Core GPUs.
In addition, AWS plans to deliver Amazon Elastic Compute Cloud (EC2) instances supported by NVIDIA Grace Blackwell GPUs. These EC2 instances will showcase the new B100 GPUs, set to be deployed in EC2 UltraClusters to hasten generative AI training and inference.
Moreover, Blackwell will seamlessly integrate with AWS Nitro System, Elastic Fabric Adapter encryption, and AWS Key Management Service, granting comprehensive control of the training data and model weights.
Furthering the collaboration, Amazon SageMaker will intertwine with NVIDIA NIM inference microservices to optimize price performance for foundation models running on GPUs.
The Impressive Strength of AWS Portfolio and Competitive Landscape
The recent move signals Amazon’s unwavering commitment to expanding the AWS portfolio, a driving force behind its cloud customer momentum. AWS’s prowess continues to catalyze Amazon’s growth, with AMZN shares up by 74.9% in the last year.
We anticipate that AWS’s robust customer base, propelled by its expanding portfolio, data centers, and cloud regions, will further solidify Amazon’s competitive advantage against industry rivals such as Microsoft MSFT and Alphabet‘s GOOGL Google.
According to the latest data from the Synergy Research Group, Amazon held 31% of the global cloud provider market in the fourth quarter of 2023. AWS raked in revenues of $24.2 billion (14% of total sales) during the same period, marking a 13% year-over-year increase.
Microsoft, commanding 24% of the market share, is capitalizing on the momentum within its Intelligent Cloud business, which reported $25.9 billion in revenues (41.7% of total revenues) in the second quarter of fiscal 2024, a 20% year-over-year surge.
On the other hand, Google Cloud, contributing significantly to Alphabet’s total revenues, captured an 11% market share and generated revenues of $9.19 billion (10.6% of total revenues) in the fourth quarter of 2023, reflecting a 25.7% year-over-year growth.
Amazon’s Forward March in the Generative AI Space
The extended collaboration with NVIDIA aligns seamlessly with Amazon’s proactive drive to boost its generative AI capabilities.
AWS’s announcement about the general availability of its fully managed service, Amazon Bedrock, offering seamless access to high-performing foundation models from AI companies via an API, is a notable feat.
Furthermore, Amazon made the Amazon Titan Embeddings model generally available and incorporated Meta’s Llama 2 into Amazon Bedrock as a new offering accessible through an API.
Introducing Amazon CodeWhisperer, a unique capability for providing generative AI-powered custom code suggestions by leveraging an organization’s internal codebase, further cements AWS’s leadership in the generative AI arena.
The augmentation of generative AI efforts is expected to position Amazon favorably to harness the vast growth opportunities in the generative AI market.
As per a report from Allied Market Research, the global generative AI market is projected to reach $191.8 billion by 2032, enjoying a robust CAGR of 34.1% from 2023 to 2032.