The AI startup Anthropic is deepening its collaboration with Amazon Web Services (AWS). Amazon is investing an additional 4 billion US dollars, making it the primary cloud and training partner for Anthropic. This increases Amazon’s total investment in the startup to 8 billion US dollars. However, Amazon remains a minority investor. Other investors include Alphabet (Google), Microsoft, Apple, and Nvidia.
Anthropic engineers will closely work with Annapurna Labs, a subsidiary of AWS, which is behind the Trainium chips. These chips are specialized for training machine learning models. The goal of the partnership is to further optimize hardware efficiency. Anthropic aims to benefit from this in training its foundation models, particularly the Claude family.
Claude has been a core infrastructure at Amazon Bedrock for some time. According to Anthropic, tens of thousands of companies, including Pfizer, use it. The pharmaceutical company aims to accelerate drug research and development with Claude models while saving millions in operational costs.
The AI-powered search engine Perplexity also provides answers using Claude in Amazon Bedrock. The European Parliament uses Claude to make 2.1 million official documents searchable in multiple languages with “Archibot,” reducing research time by 80%.
Anthropic was founded in 2021 by former OpenAI members, including Dario Amodei, previously vice president of research at OpenAI. The company specializes in developing general AI systems and language models and is committed to responsible AI use.
However, Anthropic faces criticism: several rights holders have filed lawsuits against the company, alleging that copyrighted material was used to train the Claude chatbot. Recently, three authors accused Anthropic of unlawfully using their works. The company is also dealing with copyright lawsuits from music publishers.