Artificial intelligence startup Cohere has raised $270 million from a mix of venture capital and strategic investors including Oracle Corp. and Nvidia Corp., an executive said in an interview.

Toronto-based Cohere is valued in the round at $2.1 billion to $2.2 billion, according to a person familiar with the matter who asked not to be identified because it was private.

Inovia Capital led the round, with participation, in addition to Oracle and Nvidia, from Salesforce Ventures, an arm of Deutsche Telekom AG, Mirae Asset, Schroders Capital, SentinelOne Inc., Thomvest Ventures and existing investor Index Ventures, said Cohere President and Chief Operating Officer Martin Kon.

Kon, previously a senior executive at YouTube, said the funding round brings the total raised by the company to about $445 million.

There’s been a rush of funding to AI companies after Microsoft Corp. poured $10 billion into ChatGPT-maker OpenAI in January. Kon said it’s a priority for Cohere, which builds and trains AI models for enterprise customers, to not give up too much ownership or control to one company.

“We don’t want to give away a big part of our company to some enterprise and then we need to serve their wishes,” Kon said. “We want make this accessible to as many companies as possible on the cloud platform of their choosing.”

Founded in 2019, Cohere builds large language models — software trained on massive swaths of the internet to analyze and generate text – and customizes them for users. Companies can use its models to do things like summarize customer emails or help write website copy. Co-founder and Chief Executive Officer Aidan Gomez previously worked at Google, where he was a co-author of a 2017 landmark paper in AI research titled Attention is All You Need, which led to advances in the ways computers analyze and generate text.

Nvidia CEO Jensen Huang said in a statement that Cohere has made foundational contributions to generative AI. Nvidia’s market value has surged to more than $900 billion thanks to demand for its chips that power AI.

“Their service will help enterprises around the world harness these capabilities to automate and accelerate,” Huang said of Cohere.

Large language models are expensive to train, and computing power is in high demand as they grow increasingly popular.

Cohere’s approach includes seeking capital from companies looking to deploy the technology to their customers, Kon said. The new funding will help the company meet its computing needs and expand into new technologies, he said.

Author: Katie Roof, Rachel Metz and Liana Baker