Anthropic PBC Raises $3.5 Billion to Accelerate AI Development

In a recent announcement, artificial intelligence company Anthropic PBC revealed the successful completion of a $3.5 billion late-stage funding round, pushing its valuation to $61.5 billion. This development comes less than two months after securing $1 billion from Google, and following Amazon's decision to double its investment in Anthropic to $800 million, making it the primary cloud service provider for the company since September of last year.

This Series E financing round was led by Lightspeed Venture Partners, with participation from Cisco Investments, Salesforce Ventures, and more than six other investment firms. Initially, Anthropic aimed to raise $2 billion, but strong investor interest prompted an increase in the fundraising target.

Just days before the funding announcement, Anthropic introduced Claude 3.7 Sonnet, its latest and most advanced large language model. Combining features of traditional large language models with reasoning capabilities, the model reportedly outperforms its predecessors significantly in mathematics, physics, and programming tasks. A standout feature of Claude 3.7 Sonnet is its customizable task-processing time. For quick responses, developers can limit the model’s processing time to just a few seconds, while for complex tasks, processing can be extended to several minutes or longer.

Alongside this, Anthropic also launched Claude Code, an embedded programming tool designed to automate coding tasks that typically require over 45 minutes of manual effort, all through a command-line interface.

Anthropic plans to use the new capital to further enhance its AI systems and expand computational resources. Currently, the company relies on AWS instances powered by Amazon Web Services' Trainium and Inferentia chips to support model development.

Last month, Anthropic shared insights into its near-term roadmap, which focuses on improving Claude Code’s ability to interact with external systems like code repositories and enhancing the tool’s "self-awareness" capabilities.

A portion of the fresh funding will also be allocated toward research into mechanistic interpretability—a method for exploring how large language models make decisions. In a recent paper, Anthropic researchers detailed how they applied this technique to study concept representations in earlier versions of Claude. By gaining a deeper understanding of how these models process data, the company aims to improve their safety protocols.

Additionally, Anthropic is pursuing international expansion. Last month, the company hired a former Google DeepMind research scientist to lead its new office in Zurich, following the establishment of a branch in Dublin, Ireland.

According to Bloomberg, Anthropic’s annualized recurring revenue surpassed $1 billion last year, with a 30% growth recorded in the past two months. Although not yet profitable, Anthropic generates income through chatbot services and application programming interfaces (APIs), similar to OpenAI, allowing developers to integrate Claude into their applications.