NOMINATION HIGHLIGHTS

Over the past year, we have helped large organizations move from proof-of-concept to production with language models that are private, controllable, and multilingual. We are Cohere, founded in 2019 in Toronto, focused on enterprise AI that enhances knowledge work while meeting strict requirements for security and compliance. Our platform provides state-of-the-art models, retrieval-augmented workflows, fine-tuning pipelines, and deployment options that fit public cloud, private cloud, or on-prem environments.

Our innovation is an enterprise-first approach to generative AI. Rather than centering on a single public chatbot, we deliver a secure model platform that plugs into a customer’s data, tools, and governance. Command-class models, including our latest Command A family, are designed for long-context reasoning and RAG use cases, with strong multilingual capabilities. Customers can host via APIs or dedicated clusters and keep data residency under their control, which is essential for regulated industries.

What makes this different is the combination of privacy, portability, and scale. Our customers can deploy where they need to operate, integrate with internal systems, and measure outputs against domain-specific ground truth. We pair this with a growing ecosystem of partners that bring our models to enterprises globally, including work with Fujitsu on Japanese LLMs for private deployments and integrations across major infrastructure providers.

Key strengths and measurable impact:
• Market momentum and scale: in August 2025 we raised 500 million USD at a 6.8 billion USD valuation, followed by an extension reported at 7 billion USD, supporting global expansion and agentic AI development.
• Enterprise readiness: dedicated inferencing options with long context, citations, and secure clusters, enabling production deployments in regulated sectors.
• Product depth: model portfolio optimized for enterprise tasks such as search, summarization, analytics, and automation, with the Command line recommended for current use and Command R+ documented for complex RAG.
• Global team and presence: founded in Canada with offices in North America and Europe; public trackers place headcount in the 500–600 range in 2025.

Why we deserve to win: our work proves that powerful language models can be delivered with the privacy, governance, and flexibility enterprises need. We do not require customers to compromise between innovation and control. By offering cloud-agnostic deployment, multilingual capability, long-context reasoning, and secure integration with internal data, we help institutions improve productivity and decision quality while keeping sensitive information protected.