Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More
Even as questions and criticism swirl around the return-on-investment of generative AI tools for enterprises, one specifically enterprise-focused AI startup, Cohere, from Toronto, Canada, has just today announced a fresh $500 million Series D fundraising round.
That brings the company’s total valuation to $5.5 billion, nowhere near the reported $90 billion of OpenAI, but a still significant sum that shows investor enthusiasm in the sector is not waning quite yet.
Cohere’s new fundraising, first reported by Bloomberg and confirmed by the company, was led by the Canadian pension investment firm PSP Investments and a group of new investors, including Cisco Systems, Fujitsu, AMD’s venture arm AMD Ventures, Magnetar, and Export Development Canada.
Cohere head of communications Josh Gartner told VentureBeat that the Series D fundraising also had returning investors in Oracle, Salesforce Ventures and Nvidia.
The news comes following additional investment in the sector for rivals. Anthropic raised $450 million in Series C financing this year, French AI company Mistral got a $640 million investment in June, and the search chatbot Perplexity became a unicorn after a $65 million investment.
Continuous development of enterprise AI models, with focus on data privacy and security
Gartner said the additional capital will support the continuous development of Cohere’s models with a focus on data privacy and security, multilingual accuracy and capabilities like retrieval augmented generation (RAG), one of the leading themes to emerge from the VB Transform event in San Francisco earlier this month.
Cohere has released two enterprise-focused large language models (LLMs), Command R and Command R+.
At the time of the release of Command R+ earlier this year, it boasted metrics matching and in some cases outperforming other, more consumer-facing LLMs from rivals including Anthropic’s Claude 3, OpenAI’s GPT-4 Turbo and Mistral’s Mistral-Large.
But since then, the commercial providers have all released new, more powerful LLMs or in the case of OpenAI, a more powerful GPT-4o and less powerful but far cheaper GPT-4o mini model, which explicitly courts enterprises such as the kind of customers Cohere caters to, showing the intensifying competition in the sector.
Cohere also partnered with Fujitsu to build language models capable of understanding Japanese.
A tight focus on the enterprise market
While companies like OpenAI have released enterprise plans for their chatbots such as ChatGPT Enterprise while maintaining free versions for individual consumer use (ChatGPT), Cohere eschews the consumer market entirely in favor of enterprises.
Last summer, it launched the enterprise AI assistant Coral, which aims to mitigate hallucinations and trains on internal data while offering data security.
A preview version of Coral is available to consumers on Cohere’s website, as well as a Chat API for third-party developers to build their own apps atop it.