Yazılar

AI Firm Cohere Doubles Annualized Revenue to $100M by Targeting Enterprise Sector

/in /tarafından

Cohere, the Toronto-based AI startup, has doubled its annualized revenue to $100 million as of May 2025, according to a source familiar with the matter. The company’s enterprise-first strategy—focused on private, secure deployments in regulated industries—is fueling its rapid growth.

Although a Cohere spokesperson declined to confirm the financials, the company told Reuters that 85% of its business now comes from long-term enterprise contracts, with profit margins reaching 80%.

Strategic Shift: Enterprise Over Scale

Cohere’s revenue surge follows a strategic pivot in Q3 2024, when CEO Aidan Gomez announced a move away from building general-purpose, massive foundation models in favor of smaller, customized AI systems tailored to individual sectors like:

  • Finance

  • Healthcare

  • Government

This reflects a growing industry trend: domain-specific AI is now seen as more scalable, secure, and immediately useful for enterprise workflows.

“The era of scaling models for raw power is giving way to delivering domain-specific intelligence,” said Gomez in a year-end internal memo.

New Product Launch: North

In January 2025, Cohere launched North, a ChatGPT-style assistant designed to help knowledge workers with tasks like document summarization and data analysis. The product is currently in limited trials with early customers including:

  • Royal Bank of Canada

  • LG

Market Position and Backing:

  • Founded in 2019, Cohere has raised over $900 million from investors including Nvidia, Cisco, and Inovia Capital.

  • The company was last valued at $5.5 billion.

  • Current enterprise clients include Fujitsu, Oracle, and Notion.

Industry Context:

Cohere’s enterprise-focused model aligns with broader AI sector dynamics, where efficiency, security, and customization are increasingly favored over monolithic, general-purpose AI systems. This comes as major AI labs face diminishing returns from increasing model size, a strategy that once drove breakthrough performance.