While giants like Google, Anthropic, and OpenAI dominate headlines, Canada’s Cohere has been quietly building momentum.
In 2025, the Toronto-based AI company surpassed its $200 million annual recurring revenue target, closing the year at $240 million in annual recurring revenue. Even more striking was its pace, which exceeded a 50% quarter-over-quarter growth rate throughout the year, according to a memo shared with investors and industry reports.
In a market crowded with research breakthroughs and product launches, sustained revenue growth of that magnitude stands out. Cohere’s performance suggests enterprises are experimenting with AI and committing budgets.
Built for efficiency
Founded in 2019 by Aidan Gomez, Ivan Zhang, and Nick Frosst, Cohere has attracted backing from enterprise-focused technology leaders including Nvidia, AMD, and Salesforce. The investor base reflects its strategy of building AI that fits into real-world corporate infrastructure rather than requiring massive computing budgets.
At the heart of the company’s offering is its Command family of generative AI models. Unlike some rivals that require extensive GPU resources, Cohere positions its models as efficient enough to run on limited hardware. For enterprises grappling with rising infrastructure costs and limited access to GPUs, that promise is practical and financially compelling.
Efficiency is becoming a differentiator. As AI adoption scales, cost control is no longer secondary but strategic.
From models to workflows
Last summer, Cohere expanded beyond models with the launch of North, a high-level enterprise platform designed for secure, customised AI agents and workflows.
North transforms Cohere’s underlying models into operational tools. It offers businesses a structured workspace to deploy AI agents tailored to their internal processes, all within secure environments.
This shift matters. Companies increasingly want AI that integrates with compliance rules, internal data systems, and specific operational needs. North positions Cohere not just as a model provider, but as an enterprise AI partner.
By moving up the stack, Cohere strengthens customer relationships and expands revenue opportunities beyond model access alone.
Is an IPO on the horizon?
The company’s financial momentum now feeds a bigger question: When does Cohere go public?
CEO Aidan Gomez indicated last October that an IPO could come “soon.” If that timeline points to 2026, Cohere may enter public markets alongside other high-profile AI contenders. OpenAI, Anthropic, and even SpaceX’s xAI are reportedly weighing their own public debuts.
An IPO would test whether investors view Cohere as a durable enterprise infrastructure player rather than a short-term AI beneficiary. Its $240 million revenue milestone provides a strong foundation.
In a sector defined by speed and spectacle, Cohere’s disciplined climb may prove its greatest advantage, especially as it edges closer to the public stage.
