AWS and Marvell Expand Partnership for AI and Data Connectivity
Amazon Web Services (AWS) and Marvell Technology, Inc. (MRVL) have announced an expanded five-year partnership focused on AI and data connectivity products. This multi-generational agreement covers multiple AWS products and cloud infrastructure, strengthening the existing strategic relationship between the two companies.
Custom AI products: This includes the AWS Trainium2.0 custom AI ASIC, designed to accelerate machine learning workloads.
These processors are used for high-speed data transmission in data centers.
These DSPs enable high-performance data transmission over copper cables.
These modules are used for connecting data centers over long distances.
These solutions provide high-performance networking capabilities for data centers.
Additionally, Marvell will utilize AWS compute infrastructure to enhance its engineering teams' resources and flexibility. This will allow them to handle dynamic compute requirements and accelerate the development of new products.
Analysts believe this expanded partnership reflects strong demand for AWS's custom-designed Trainium2.0 chips produced with Marvell. Production is expected to ramp up in the fourth quarter of 2023. This partnership could also help alleviate AWS's dependence on supply-constrained Nvidia chips in the long term.
However, it's important to note that the current mix of Marvell and Trainium chips used by AWS is still small compared to Nvidia. Based on available data, it's estimated that Amazon will spend $1 billion to $2 billion on Marvell chips in 2024, compared to over $20 billion on Nvidia chips.
Limited preview of Oracle Database@AWS: This service will be expanded in the coming months as the two companies continue their partnership.
Overhaul of the cloud contact center solution with new Generative AI solutions.
Lotus choosing AWS as its Preferred Cloud Provider for its self-driving development.
Strategic partnership with Orbital Materials to develop and pilot technology for data center carbon removal, chip cooling, and water utilization.
With the growing demand for AI innovation, analysts expect an eventful Re:Invent conference this week. They anticipate announcements on partnerships across Gen-AI infrastructure, access and deployment of new foundational models, and new AI-powered applications.