OpenAI rival Anthropic PBC today announced that it has closed a late-stage $3.5 billion funding round at a $61.5 billion valuation.

The Series E deal comes less than two months after the company raised $1 billion from Google LLC. Two months before that, Amazon.com Inc. doubled its investment in Anthropic to $8 billion. Amazon Web Services became the artificial intelligence provider’s primary cloud provider last September. 

Lightspeed Venture Partners led Anthropic’s latest funding round. It was joined by Cisco Investments, Salesforce Ventures and more than a half-dozen other backers. It’s believed Anthropic had originally sought to raise $2 billion but upped its fundraising target in response to strong investor interest. 

The investment comes a few days after Anthropic debuted Claude 3.7 Sonnet, its latest and most capable large language model. It combines the features of a standard LLM and a reasoning model. Anthropic says the algorithm is significantly better at math, physics and programming than its predecessors.

One of Claude 3.7 Sonnet’s flagship features is that developers can customize the amount of time it spends on a task. If the answer to a prompt must be generated quickly, a software team could limit the model’s thinking time to a few seconds. That cap can be raised to several minutes or more for complicated tasks.

Claude 3.7 Sonnet was introduced alongside a second new offering dubbed Claude Code. It’s described as an agentic coding tool. Anthropic says the software, which is embedded in a command line interface, can automate programming tasks that usually require more than 45 minutes of manual work. 

Today the company said it will use the Series E funding round to further enhance its AI systems. To support the effort, the company will commission more compute capacity. Anthropic uses AWS instances powered by the cloud giant’s internally developed AWS Trainium and Inferentia chips to support its model development efforts. 

Last month, Anthropic revealed that its near-term development roadmap will prioritize enhancing Claude Code. The company plans to enhance the tool’s ability to interact with external systems such as code repositories. Additionally, Anthropic will equip the tool with better “self-knowledge of its capabilities.”

The company will also use some of its new funding to support research into mechanistic interpretability. That’s an approach to uncovering how LLMs make decisions. In a paper published last year, Anthropic detailed that its researchers had used the method to study how an earlier version of Claude represented concepts. The company hopes that gaining a better understanding of how LLMs process data will make it possible to find ways of making them safer.

“This investment fuels our development of more intelligent and capable AI systems that expand what humans can achieve, while deepening our understanding of how these systems work,” said Anthropic Chief Financial Officer Krishna Rao.

In parallel with its research efforts, Anthropic will launch an international expansion initiative. Last month, the company hired a former Google DeepMind research scientist to lead a new office in Zurich. It previously opened a branch location in Dublin, Ireland.

According to Bloomberg, the company passed $1 billion in annualized recurring revenue last year. That number has reportedly grown by 30% in the past two months. Anthropic, which is not yet profitable, generates revenue with an OpenAI-like chatbot and an application programming interface that enables developers to embed Claude into their applications.

Image: Anthropic