Cohere president Martin Cohn says many of the hot AI startups on the market today are building the equivalent of luxury sports cars. His product, he says, is more like a heavy-duty truck.
“If you’re looking for vehicles for your field service department and I take you for a test drive in a Bugatti, you’ll be impressed by how fast and how well it performs,” Cohn said in an interview with CNBC. However, he said, the price, combined with space limitations and the lack of a trunk, will be an issue.
“What you really need is a fleet of F-150 pickup trucks,” Cohn said. “We make the F-150.”
Founded by exGoogle AI researchers and supported by NvidiaCohere is betting on generative AI for the enterprise rather than consumer chatbots, which has been the talk of the tech industry since OpenAI launched ChatGPT in late 2022.
In June, Cohere raised $270 million at a $2.2 billion valuation, with Salesforce and Oracle participating in the funding round. Company executives have attended White House AI forums. Cohere is reportedly in talks for a raise up to $1 billion in additional capital.
“We don’t comment on rumours,” Cohn told CNBC. “But someone once told me that startups always raise.”
The generative AI field exploded last year with a record $29.1 billion invested in nearly 700 deals in 2023, a more than 260% increase in deal value from a year earlier, according to PitchBook. It’s become the buzz phrase in corporate earnings conversations quarter after quarter, and some form of the technology is automating tasks in nearly every industry, from financial services and biomedical research to logistics, online travel and utilities.
Although Cohere is often mentioned alongside AI heavyweights such as OpenAI, Anthropic, Google and Microsoftthe startup’s focus on enterprise-only chatbots sets it apart.
Competitors offer AI products for both consumers and businesses. OpenAI, for example, launched ChatGPT Enterprise in August, and Anthropic opened consumer access to its previously business-only chatbot Claude in July.
Cohn, who is also the company’s chief operating officer, said that by remaining solely focused on the enterprise, Cohere has been able to operate efficiently and keep costs under control even amid chip shortages, rising graphics processing unit (GPU) costs and ongoing changing license fees for AI models.Â
“I’ve rarely seen in my career many companies that can successfully be both consumer and enterprise at the same time, let alone a startup,” Cohn said. He added: “We don’t need to raise billions of dollars to offer a free consumer service.”
Current customers include Notion, Oracle and Bamboo HR, according to Cohere’s website. Many clients fall into the banking, financial services and insurance categories, Cohn said. In November, Cohere told CNBC that it had seen an uptick in customer interest following OpenAI’s sudden and temporary ouster of CEO Sam Altman.
Cohn acknowledges that the changing dynamics of the hardware industry present constant challenges. The company has a supply of Google chips for more than two years, Cohn said, provided in Cohere’s early days to help it train its models.
Cohere is now moving towards using more of Nvidia’s H100 GPUs, which power most of today’s major language models.
Cohere’s relationship with strategic investors is another area where it differs from generative AI competitors, Cohn said. Many companies were created by the likes of Nvidia and Microsoft with some terms that are related to the use of their software or chips.
Cohn is adamant that Cohere has never accepted a conditional investment, and that every check it has cashed — including from Nvidia — has had no strings attached.
“In our last round, we had multiple checks of the same size; we had no conditions attached to any of them,” Cohn said. “We specifically made this decision so we could say we don’t owe anyone.”
Cohere’s decision to focus solely on enterprise chatbots could help the company stay out of the murky territory of misinformation concerns, especially as election season approaches.
In January, the Federal Trade Commission announced an investigation into AI Amazon, Alphabet, Microsoft, OpenAI and Anthropic. FTC Chair Lina Hahn described it as “a market study of the investments and partnerships formed between AI developers and major cloud service providers.” Cohere was not named
Cohn says the company’s growth so far has largely been around areas like search and retrieval, which require their own separate AI models. He calls it “tool usage,” and it involves training models on where, when, and how to look for information the enterprise customer needs, even if the model wasn’t originally trained on that data.Â
Search, Cohn said, is a key part of generative AI that has received less attention than other fields.
“That’s certainly going to be the real unlock for the enterprise,” he said.
In discussing the expansion schedule, Cohn called 2023 the “proof-of-concept year.”
“We think 2024 is becoming the year of deployment at scale,” he said.Â
Don’t miss these stories from CNBC PRO:
WATCHING: Generative AI will democratize access to enterprise data.
https://www.cnbc.com/2024/03/01/cohere-says-rivals-are-building-bugatti-sports-cars-we-make-f150s.html