Claude AI Creator Anthropic Might Gain Amazon’s Backing Again—If It Embraces AWS Hardware

(cybermagician/Shutterstock)

Amazon is rumored to be negotiating a second multibillion-dollar investment deal with AI startup Anthropic, the maker of the Claude family of LLMs and a rival to OpenAI. 

Amazon previously invested $2.75 billion in the startup in March, adding on to an initial $1.25 billion backing in September 2023. 

According to the latest report, Amazon is asking Anthropic to use servers powered by its own AWS Trainium and AWS Inferentia AI chips hosted on the AWS cloud in order to reduce costs and reliance on Nvidia hardware. The final funding amount may be contingent on the number of Amazon chips Anthropic commits to using. 

There is no official comment from either of the companies and it’s unclear how far along negotiations have progressed. 

The report mentioned Anthropic may be hesitant to go all in with AWS infrastructure due to the possibility for technical issues with Amazon’s software ecosystem that is not currently as robust as Nvidia’s Cuda platform. A shift toward Trainium could also mean less flexibility with other cloud providers since Amazon hardware is not available in non-Amazon facilities. 

The previous agreement in March included Anthropic using AWS as its primary cloud provider, as well as committing to providing AWS customers with access to Anthropic foundation models on its fully managed Amazon Bedrock service. It was also reported that Anthropic could power future versions of Amazon’s Alexa smart assistant. 

“Anthropic’s visionary work with generative AI, most recently the introduction of its state-of-the-art Claude 3 family of models, combined with Amazon’s best-in-class infrastructure like AWS Tranium and managed services like Amazon Bedrock further unlocks exciting opportunities for customers to quickly, securely, and responsibly innovate with generative AI,” Dr. Swami Sivasubramanian, vice president of Data and AI at AWS, said in March at the time of the $2.75 billion investment. 

Anthropic was co-founded by siblings Dario and Daniela Amodei who are former OpenAI executives. Anthropic’s flagship AI is Claude, an AI model based on “constitutional AI,” a proprietary method which the company describes as a way to align general purpose LLMs “to abide by high-level normative principles written into a constitution,” and guide outputs away from toxic responses. The constitution Claude adheres to is written and curated by Anthropic employees and takes inspiration from documents like the UN’s Universal Declaration of Human Rights, according to the company. 

Anthropic has raised $9.7 billion in funding rounds from investors like Google, Microsoft, Salesforce Ventures, Zoom and of course, its largest investor Amazon. It was reported in late September the company was looking to raise more funds at a valuation of nearly $40 billion and will likely reach $1 billion in revenue this year, a number that would represent a 1,000% YoY growth for the company. 

Just yesterday Anthropic announced it is teaming up with Palantir Technologies and AWS to provide access for its Claude 3 and 3.5 models on the AWS GovCloud. The companies said the partnership will allow for an integrated suite of technology to operationalize the use of Claude within Palantir’s AI Platform while leveraging the security and flexibility provided by AWS. Claude can now be used in Palantir’s defense-accredited environment, Palantir Impact Level 6 (IL6), a security level which the DoD reserves for systems with critical national security data only one step below the “top secret” designation. 

Related