Omar Marquez | Light Rocket | Getty Images
Cohere co-founder Nick Frosst is surrounded by buzz around artificial general intelligence (AGI). He is perfectly happy not to participate in the conversation.
It was founded in 2019 by former employees.google Cohere is valued at billions of dollars, according to AI researchers, and is one of the hottest companies in the world of generative AI, which has grown explosively since OpenAI debuted ChatGPT in late 2022. One.
But the company isn’t well-known among consumers who flock to chatbots and other tools from OpenAI, Google, and Perplexity. Rather, Cohere is all about business.
“I’m in meetings all the time with companies in healthcare, banking, and IT,” Frost said in an interview with CNBC this week. “The questions I get are about securely automating tasks such as HR, payroll, investigations, and fraud detection to improve productivity. No one asked me about achieving AGI, let alone ASI. not here.”
The latter stands for Artificial Super Intelligence, or AI that significantly exceeds human intelligence. Both OpenAI and Anthropic aim to achieve that.
In its latest funding round in July, Cohere raised $500 million at a valuation of $5.5 billion, more than doubling its valuation from a year ago. Investors in the company include: NvidiaAMD, sales force and oracle.
That’s historically a hefty price tag for a company less than five years old, but investors include OpenAI, which was valued at $157 billion in a round announced in October, and Anthropic, which CNBC confirmed this week. It’s only a fraction of what you’re paying for the house. The company is in talks to raise funds at a valuation of $60 billion.
Some of Cohere’s main competitors in the AI arms race offer products to both consumers and businesses. For example, OpenAI launched ChatGPT Enterprise in 2023, and Anthropic announced Claude Enterprise in September.
Frost said Cohere’s preference for enterprise centers around the idea that large language models are great for automating tedious tasks and “becoming colleagues.”
“It’s really an automation tool,” Frost said. “If you think about your personal life, there aren’t a lot of things that you actually want to automate. You don’t want to text your friends faster or respond to emails more efficiently in your own life. But in my professional life, I really want to do that.”
“I want to be able to think freely and creatively without getting stuck,” Frost said.
Shortly after completing its funding round in July, Kohia cut about 20 jobs. A company representative said at the time that this was an “internal reorganization” and that Cohere had a “clear vision for the future.”
That vision includes a full commitment to AI agents.
The term AI agent is not clearly defined, but is generally meant to refer to AI services beyond chatbots. Agents are typically designed for specific business functions rather than general purpose, and can be customized based on large-scale AI models.
It can perform complex, multi-step tasks on your behalf and generate your own to-do list, so you don’t have to walk them through the process.
Maintaining capital efficiency
On Thursday, Cohere debuted an early access program for its AI agent platform called North. The program focuses on allowing users of any level of technical background to “instantly customize and deploy AI agents,” with “just a few clicks.” the company said in a press release. Users can search for information across the organization in multiple languages and in departments with previously unconnected programs.
This includes summarizing questions and answers in HR, reducing the time spent on financial reporting, and automating some of the core business functions in customer support and IT departments.
Frost said the platform can be used in any industry, but the company plans to target finance and healthcare, where data privacy and regulation are paramount.
Martin Kon, Cohere’s executive director, told CNBC in March that a continued focus on enterprise AI has kept the company efficient despite chip shortages and rising costs for Nvidia’s graphics processing units (GPUs). He said he will be able to operate the business and keep expenses down. – Changes to license fees for AI models.
Frost said these dynamics are still at play, allowing Kohia to “improve capital efficiency,” which is “increasingly of interest to investors.” Rivals for consumer AI products use large amounts of computing in “consumer applications and scientific projects,” he said.
Enterprise AI sales cycles can be long, but “the recurring business we were able to build is now resonating with investors,” Frost said.
Competition is fierce and technology is rapidly evolving.
Anthropic announced in October that its AI agents are capable of using computers in the same way humans do to complete complex tasks. This feature, called “computing,” allows the company’s technology to interpret what’s on your computer screen, select buttons, type text, navigate websites, and complete tasks through any software or real-time Internet browsing. Now you can run it.
OpenAI is also reportedly planning to introduce similar functionality soon. And last year, microsoft, Meta And Google has regularly touted its goal of making its AI assistant even more productive.
Even without a consumer business, Cohere still needs to spend a lot of money on Nvidia’s expensive GPUs. GPUs are in high demand for enterprises training models and running large-scale workloads. In Cohere’s early days, the company secured a reserve of Google chips to help pre-train models. Over the past year, Cohere has been moving more towards Nvidia’s H100 GPUs.
“They’re working so well that we’ve increased spending on them,” Frost said.

