Databricks to use AWS Trainium AI chips to accelerate development of custom models built with Mosaic AI
Databricks, the Data and AI company, announced a strategic collaboration agreement (SCA) with Amazon Web Services (AWS) to accelerate the development of custom models built with Databricks Mosaic AI on AWS. Databricks will leverage AWS Trainium chips as the preferred AI chip to power Mosaic AI model training and serving capabilities on AWS. Joint customers can leverage Mosaic AI to pretrain, fine-tune, augment and serve large language models (LLMs) on their private data, backed by the scale, performance and security of AWS. The expanded partnership will also include new integrations for Databricks on AWS Marketplace.
Databricks and AWS power customized, production-quality generative AI applications
The collaboration builds on Databricks’ newly expanded Mosaic AI capabilities, which includes Mosaic AI Model Serving. Mosaic AI Model Serving is designed to support a variety of model providers, including the broad selection of models available via Amazon Bedrock. Customers can also scale model training on Mosaic AI at a low cost using high-performing, cost-effective AWS Trainium chips. Together, Databricks and AWS will make it possible for thousands of joint customers to build, deploy and monitor high-quality, customized AI applications without sacrificing control over their data or intellectual property.
“Generative AI is disrupting every industry and a company’s data will ultimately be its competitive advantage. Strengthening our collaboration with AWS allows us to provide customers with unmatched scale and price-performance so they can bring their own generative AI applications to market more rapidly,” said Ali Ghodsi, Co-founder and CEO at Databricks.
“Our collaboration with Databricks unlocks the potential for customers to drive real value from their data using generative AI,” said Matt Garman, CEO of AWS. “By using AWS Trainium to power Mosaic AI, Databricks will make it cost-effective for customers to build and deploy generative AI applications on top of their analytics workflows, regardless of their industry or use case. This powerful combination will allow businesses of all sizes to gain valuable insights from their data, helping customers innovate faster by focusing on what truly matters most for their business.”
Also Read: Axiado and AMI Partner for NVIDIA MGX Server Management Stack
Expanded Databricks and AWS go-to-market partnership and joint offerings
This agreement builds on the long-standing relationship between Databricks and AWS, its first cloud partner. Databricks’ AWS business has already surpassed a $1 billion run rate, fueled in part by AWS Marketplace, where total contract value has doubled on an annualized basis over the past two years.
As part of this relationship, Databricks and AWS will continue to invest in product capabilities and programs to support customers, including:
- Custom Model Optimization & Enhanced Security: Databricks and AWS will collaborate on the development of custom models built with Mosaic AI using AWS Trainium chips to improve cost and performance for training and running LLMs. AWS Trainium benefits from the AWS Nitro System, a collection of hardware and software components designed to provide high performance, availability, privacy and security for cloud-based workloads.
- Migration and Modernization for Generative AI: Databricks and AWS will work with systems integrator (SI) partners to migrate data estates running in on-premises data centers onto AWS and accelerate the modernization of their data platforms.
- Tailored Industry Solutions and Accelerators: As the demand for generative AI grows across industries like Media and Entertainment and Financial Services, Databricks and AWS will continue to develop joint industry solutions that help customers solve challenges within their organization and drive innovation.
- Enhanced Ease of Use: The two companies are making it easier for customers to adopt and run Databricks on AWS through new integrations on AWS Marketplace, including simplified onboarding and configuration and serverless compute powered by AWS.
- Joint Generative AI Go-to-Market Programs: Alongside its SI partners, Databricks will build technical solutions and implementation resources to help customers identify generative AI workloads and on-premises migrations that can be optimized with Databricks on AWS.
- Expanding Co-Marketing Initiatives: Databricks and AWS will continue to invest in strategic joint marketing programs, including each company’s flagship industry events. This work will also include co-branded technical resources and developer-focused events designed to allow customers to prototype generative AI solutions.
Continued commitment to joint customer success
The SCA underscores Databricks’ commitment to AWS and both companies’ shared focus on delivering business value to joint customers, including Block, Rivian and SEGA.
“At SEGA, our goal is to deliver a personalized gaming experience that keeps our millions of customers coming back for more. To do this, we knew we needed a unified data platform that could both handle and harness our data to help us turn insights into new features our customers will love,” said Stuart Wright, CTO of SEGA Europe. “The rapid integration of generative AI has immense potential to accelerate innovation in the video game development, publishing and creative process. The collaboration between AWS and Databricks gives us the model choice, scalability, security and price-performance we need to build a better gaming experience, improve brand loyalty and grow our community.”
The Databricks Data Intelligence Platform is available on AWS and has achieved FedRAMP® High Agency Authority to Operate for its cloud services on AWS GovCloud.
Source: PRNewswire