Cohere, a Canadian AI firm that was recently valued at $5.5 billion, told Reuters that it will prioritize creating customized models for corporate users over more general foundation models.
Two years after ChatGPT first appeared on the market, many businesses are still attempting to figure out how to integrate huge language models into their everyday operations. This is the reason behind the company’s strategic evolution, which was detailed in a letter to investors on Thursday.
Customers are telling us that they don’t need larger models to be proficient in every area. In an interview with Reuters, Nick Frosst, co-founder of Cohere, stated, “They require models that are truly constructed for their particular use cases.”
Cohere, a rival to AI laboratories like OpenAI and Anthropic, claims that while it will keep creating foundation models, it will put more of an emphasis on alternative training methods to enhance models rather than expanding their sizes. The emphasis is on customized model deployment, even if Cohere will continue to sell Application Programming Interface (API) to its models as a minor component of its business.
An investment boom from startups to big companies has been driven by the competition to create larger and better models. To finance the capital-intensive development of frontier AI models, OpenAI, Anthropic, and xAI have raised billions of dollars.
Cohere, which has its headquarters in Toronto and San Francisco, has raised more than $900 million from investors such as Innovia Capital, Cisco, and Nvidia.
Cohere has positioned itself as a cloud-independent AI firm with an enterprise emphasis. To customize models for particular requirements, it has been collaborating closely with clients including Fujisu and Oracle.
