Reuters reports that Cohere has secured $450 million in new funding from NVIDIA, Salesforce, and Cisco, raising its valuation to $5 billion. This figure represents more than double the $2.2 billion valuation the company held in June 2023.
While Cohere and its major investors have not officially confirmed the figures, sources indicate the company reached an annualized revenue run rate of $35 million in March. That is roughly $3 million per month, nearly three times the $13 million run rate recorded at this time last year.
Strategy
Established in 2019, Cohere is one of the earliest developers of commercial large language models. It has trailed behind OpenAI, Anthropic, and Google in the general-purpose AI segment. Instead of chasing foundation models, Cohere has focused on customising technology for specific use cases.
The company offers several tools designed for niche applications:
- Command R focuses on retrieving augmented generation at scale. This improves response accuracy by grounding answer retrieval in known datasets.
- Embed is designed to improve LLM response accuracy.
- Rerank is a foundation model intended to enhance semantic search and deliver cost and latency optimisation for RAG architectures.
Focusing on niche solutions is a reasonable strategy when primary competitors are tech giants and funded startups with valuations four to twenty times higher. Niche segments represent smaller addressable markets than general-purpose solutions. However, they are attractive because it is easier to align solutions around specific value propositions.
Outlook
Reaching an annualized run rate of $36 million over the next twelve months shows progress. Another tripling of revenue over the next year would indicate strong momentum. This is well behind OpenAI’s multi-billion dollar revenue engine, but Cohere competes in the solution optimisation market.
The company faces two key risks:
- General purpose frontier and open-source models continue to improve in response quality, latency, and cost.
- Adoption of advanced generative AI solutions is delayed.
The first point would result in Cohere’s solution optimisations becoming less differentiated. The second reflects the fact that Cohere is focused on system optimisation when much of the enterprise market is oriented around deploying solutions that work. Many clients are not yet thinking about optimisation.
Another issue Cohere could face is the rise of data-centric RAG optimisation solutions. Databricks and Snowflake now have LLMs combined with their data products. While their data products are general-purpose, their entry into generative AI is focused on a mix of convenience, performance, and optimisation for RAG-related solutions.
Still, the valuation and revenue progress are promising. Generative AI represents a very large and rapidly growing market. There is room for many winners and some of the niches will be sizeable in their own right.
The funding round also shows that investor enthusiasm in generative AI foundation models remains strong.
