New Delhi: Cohere, a generative AI startup focusing on enterprise applications, has successfully raised $270 million in its Series C funding round. This substantial investment signifies the availability of significant capital for companies involved in the development of AI models. While earlier reports suggested a valuation of over $6 billion for Cohere, a source familiar with the matter now values the company between $2.1 billion and $2.2 billion.
The newly acquired capital will be used to further enhance Cohere’s AI platform, specifically designed for enterprise customers. The platform enables companies to leverage their preferred cloud provider, ensuring data privacy and simplifying implementation. Martin Kon, the President and COO of Cohere, expressed his excitement about the latest funding, emphasizing the company’s plans to invest in computational resources, expand the team, engage with leading enterprises, and advance their AI capabilities. Cohere aims to empower companies to build exceptional products while ensuring data privacy and security.
Cohere was founded in 2019 by Aidan Gomez, Ivan Zhang, and Nick Frosst, who was among the early employees at Google’s Toronto AI lab. Gomez, prior to starting Cohere, co-authored the influential paper “Attention Is All You Need,” which introduced the Transformer architecture used in popular large language models like OpenAI’s GPT-4. Kon joined Cohere in early 2023 after serving as CFO at YouTube.
Differentiating itself from other generative AI startups, Cohere specializes in enterprise use cases. Its AI platform is cloud-agnostic, allowing deployment within public clouds (such as Google Cloud and Amazon Web Services), a customer’s existing cloud infrastructure, virtual private clouds, or on-site. Cohere takes a collaborative approach, working closely with customers to create customized large language models based on their proprietary data.
While Cohere keeps its customer numbers undisclosed, the company boasts collaborations with notable enterprises such as Jasper and HyperWrite for copywriting generation tasks. These tasks involve creating marketing content, drafting emails, and developing product descriptions.
Cohere has also partnered with LivePerson, a conversational marketing company, to improve explainability using fine-tuned large language models. Additionally, the startup collaborates with news outlets and Salesforce Ventures to apply machine learning algorithms for text analysis and summarization.
Kon revealed that Cohere is experiencing high demand from major enterprises, both as customers and partners. With approximately 180 employees, the company has raised an impressive $445 million in capital, surpassing other generative AI startups like Inflection AI and Adept. Only OpenAI and Anthropic have raised more, with funding amounts of $11.3 billion and $450 million, respectively. The significant capital raised by these companies reflects the high costs associated with training AI models. Anthropic, for instance, anticipates raising billions more to train the next generation of its AI systems.
Regarding funding sources, Kon stated that Cohere intentionally seeks a diverse range of global investors and avoids relying on a single company, especially a cloud provider. This strategy ensures their independence, ability to directly serve enterprises, maintain cloud-agnosticism, and deploy data-secure solutions according to customer preferences. This statement possibly alludes to the significant investments received by competitors like OpenAI and Anthropic from Microsoft and Google, respectively. Interestingly, Cohere was reportedly in advanced talks with Google for an investment of around $200 million.
Looking ahead, Kon envisions “search and retrieval” as the next area of growth for Cohere. By enabling models and chatbots to expand their knowledge base and search the web for relevant information, Cohere aims to build more powerful AI systems. This approach aligns with OpenAI’s recent experimentation. Kon believes that incorporating search and retrieval capabilities will provide users with verifiable and fact-checked information, reducing blind trust in AI models. Furthermore, Cohere plans to develop models capable of taking actions on behalf of customers, such as booking flights, scheduling meetings, or managing expenses. This places the company in competition with other startups like Adept, Inflection, and OpenAI, albeit using different approaches to connect AI with third-party applications and services.
Despite the competition, Kon emphasizes Cohere’s strength as an independent, cloud-agnostic AI platform exclusively focused on enabling customers to create proprietary large language model capabilities, leveraging their data for strategic differentiation and business value.
The oversubscribed Series C funding round was led by Inovia Capital, with participation from notable investors such as Nvidia, Oracle, Salesforce Ventures, DTCP, Mirae Asset, Schroders Capital, SentinelOne, Thomvest Ventures, and Index Ventures. The substantial funding positions Cohere to continue its growth and development in the enterprise AI market.