Cohere, which builds AI models and workplace software for businesses and governments, is reportedly in talks to raise between $2 billion and $3 billion. PYMNTS reported the discussions on September 11, citing The Globe and Mail. Its North platform helps employees search organizational information, prepare documents and automate work across existing systems.
North combines information retrieval with tools that generate summaries and carry out tasks. Cohere says organizations can connect their own data and applications, then customize agents around a team's processes. The platform can run in a customer's private cloud or on its premises, making control over sensitive information part of the product's appeal to organizations with strict data requirements.
A Cohere case study shows that approach at work inside customer CoreWeave. When a support request arrives in Slack, an engineer triggers North to gather details from internal records and checks the resulting ticket. Separate automation creates a Jira issue and brings in engineering teams using North's recommendations. A further North agent retrieves documentation and previous incident reviews to suggest a resolution. The engineer remains involved while software handles information gathering and handoffs.
Cohere charges custom enterprise prices for North and its Compass search product. It also sells production API access on a pay-as-you-go basis and offers dedicated model hosting through Model Vault, priced by computing instance and performance tier. Those offerings give customers several ways to buy the technology: as workplace software, model access or managed infrastructure.
The proposed financing would value Cohere at $20 billion, according to the Globe report carried by PYMNTS, which said terms could change. Cohere told PYMNTS that its Series E process had attracted strong investor interest but declined to confirm the speculation. The reported amount and valuation therefore remain proposed terms, with no completed round established by that report.
The commercial test is how consistently Cohere can reproduce useful deployments across customers. Its CoreWeave case study attributes shorter resolution times to both North automation and improved internal procedures, illustrating that results depend on implementation as well as models. As Cohere pursues more capital, a specific question is whether it can repeat those integrations across organizations with different records, approval processes and infrastructure.