The company announced on Wednesday (September 24) that it had achieved its valuation after raising $100 million in new funding.
“We will use this funds to further accelerate the development and global adoption of security-first enterprise AI technology across the public and private sectors,” Cohere said in the announcement.
“This is to rapidly expand its operations across North America, APAC and EMEA to increase the demand for secure, sovereign AI solutions. As organizations prioritize data control and compliance, solutions are being set up on their own to address this important gap in the market.”
The announcement comes weeks after Cohere raised $500 million in new funding at a $6.8 billion valuation in rounds that include investments from NVIDIA, AMD and Salesforce. These investors are back in this latest round, along with new backers Nexxus Capital Management and the Canadian Business Development Bank (BDC).
“The strong investor demand after the initial closing last month is a huge support for the momentum to deploy safe and sovereign AI for businesses,” Cohere's recently hired chief financial officer Francois Chadwick said in a news release.
Advertisement: Scroll to continue
“We believe Cohere's AI solutions meet neglected demand in the technology market by keeping complete control of your data in your own hands while truly improving the efficiency of your data.”
As covered here last month, Cohere positions it as selling only to businesses, using internal company data to train models, and running AI workloads on company premises rather than cloud.
AI startups are also focused on security, with CEO Aidan Gomez arguing that the need is “not met simply by a reused consumer model.”
A study by PYMNTS Intelligence has seen an increase in demand for AI-powered security solutions, and the number of CEOs who said companies were using such solutions last summer.
Cohere faces some of the same legal challenges as many of the AI world's peers. The company was sued by the News Media Alliance in February for copyright infringement. The lawsuit also alleged that it improperly trained the AI model with at least 4,000 copyrighted works. Cohere told Pymnts that the lawsuit was “misplaced and frivolous.”
