Amazon (NASDAQ: AMZN) is partnering with Anthropic to develop safer generative artificial intelligence models and make it accessible for for Amazon Web Services (AWS) customers.
Anthropic announced that it was strengthening its relationship with AWS on Monday by granting millions of developers access to its future foundation models. The AI model developer is offering AWS customers early access to exclusive features for customizing models, leveraging proprietary data, and refining capabilities, all facilitated through Amazon Bedrock.
Anthropic will use AWS Trainium and Inferentia chips to construct, train, and deploy their upcoming foundation models, capitalizing on AWS’s advantages in terms of pricing, performance, scalability and security.
Additionally, Amazon is set to invest up to USD$4 billion in Anthropic, taking a minority ownership stake in the company.
AWS will also become Anthropic’s primary cloud provider for crucial tasks, including safety research and the development of future foundation models with plans to host the majority of their workloads on AWS.
This commitment extends to granting AWS customers global access to forthcoming generations of Anthropic’s foundation models through Amazon Bedrock, AWS’s fully managed service for secure access to top-tier foundation models, along with early access to exclusive features for model customization and refinement.
This partnership will also enable Amazon developers and engineers to incorporate Anthropic models via Amazon Bedrock, empowering them to integrate generative AI capabilities into their projects, enhance existing applications, and create innovative customer experiences across Amazon’s various businesses.
As an AWS customer since 2021, Anthropic has rapidly emerged as a prominent provider of foundation models and a strong advocate for responsible generative AI deployment.
Their foundation model, Claude, demonstrates exceptional capabilities across a wide array of tasks, including sophisticated dialogues, creative content generation, intricate reasoning and detailed instructions, all while maintaining a high level of reliability and predictability.
With an 100,000 token context window, Claude can securely handle extensive volumes of information across various industries, ranging from manufacturing and aerospace to agriculture, consumer goods, as well as technical, domain-specific documents in fields such as finance, legal and healthcare.
Customers have noted Claude’s reduced likelihood of producing harmful outputs, its enhanced conversational abilities, and improved steerability compared to other foundation models, simplifying the process for developers to achieve their desired results.
Anthropic expanding generative AI stack to three layers
Anthropic’s cutting-edge model, Claude 2, consistently scores above the 90th percentile in GRE reading, writing exams and quantitative reasoning assessments.
“Since announcing our support of Amazon Bedrock in April, Claude has seen significant organic adoption from AWS customers,” said Dario Amodei, co-founder and CEO of Anthropic.
“By significantly expanding our partnership, we can unlock new possibilities for organizations of all sizes, as they deploy Anthropic’s safe, state-of-the-art AI systems together with AWS’s leading cloud technology.”
The company has been expanding its generative AI stack to include three layers. At the foundational layer, AWS continues to provide a range of compute instances, including NVIDIA (NASDAQ: NVDA) hardware and their proprietary silicon chips – AWS Trainium for AI training and AWS Inferentia for AI inference.
In the intermediary layer, AWS focuses on delivering a wide selection of foundation models from various leading providers, enabling customers to customize these models, safeguard their data privacy and seamlessly integrate them into their AWS workflows through the newly introduced Amazon Bedrock service.
With this announcement, customers gain early access to features allowing customization of Anthropic models, utilization of their proprietary data for private model creation and the option to employ fine-tuning capabilities through Amazon Bedrock’s self-service feature.
At the top layer, AWS offers generative AI applications and services like Amazon CodeWhisperer, a potent AI-driven coding assistant that provides code suggestions directly within the code editor, thereby enhancing developer productivity during coding tasks.
In their extended partnership, AWS and Anthropic are dedicating substantial resources to facilitate customers in adopting Claude and Claude 2 through Amazon Bedrock.
This support includes the establishment of the AWS Generative AI Innovation Center, staffed by teams of AI experts, aimed at assisting organizations of all scales in creating innovative generative AI-powered applications to drive transformative changes within their businesses.
Amazon and Anthropic joined President Biden at White House in July
Customers using Anthropic’s existing models through Amazon Bedrock are actively developing generative AI-driven applications that streamline various tasks, such as automating market forecasts, generating research reports, facilitating healthcare drug discovery and tailoring educational programs.
A few companies that have already gotten on board with this new tech include travel media pioneer Lonely Planet; Bridgewater Associates, a leading asset management firm catering to global institutional investors; and LexisNexis Legal & Professional, a top-tier global provider of information and analytics, serving customers in over 150 countries.
Amazon and Anthropic actively participate in numerous organizations, including the Organization for Economic Cooperation and Development AI working groups, Global Partnership on AI, the International Organization for Standardization (ISO), the National Institute of Standards and Technology (NIST) and the Responsible AI Institute.
The two companies also demonstrated their commitment by joining President Biden and industry leaders at the White House in July to endorse voluntary commitments aimed at fostering safe, secure, responsible and effective AI technology development. This collaborative effort builds upon their ongoing work to ensure the safety, security, and responsible use of AI, which will persist and expand through their enhanced partnership.