Startups > Cohere

Cohere

Cohere helps enterprises build and deploy LLM-powered applications with flexible, secure infrastructure.

Series D $1.54B total Founded 2019 Toronto, Ontario 700 employees

Cohere develops enterprise AI software providing large language models (Command, Rerank, Embed) and workplace productivity platforms that enable developers and enterprises to build LLM-powered applications. The company offers flexible deployment options—hosted API, cloud platforms (AWS, Google Cloud, Azure, Oracle), or private VPC—with advanced retrieval-augmented generation (RAG) capabilities. In 2025, Cohere launched North, an agentic AI platform designed to automate routine tasks and surface insights grounded in enterprise data. The company is distinguished by its focus on enterprise-grade security, multi-deployment flexibility, and the founding team's foundational contribution to transformer architecture.

Problem solved

Organizations struggle to build production-grade generative AI applications without choosing between vendor lock-in, security/privacy concerns, or complex infrastructure management.

Target customer

Enterprise software companies, SaaS platforms, and large organizations building generative AI applications; technical teams requiring fine-grained control over model deployment and data privacy.

Website LinkedIn Crunchbase Twitter / X

Founders

Aidan Gomez

CEO & Co-Founder

British-Canadian computer scientist and one of eight authors of 'Attention Is All You Need' (2017) while interning at Google Brain at age 20; University of Toronto graduate with PhD from Oxford (awarded 2024); named to Time 100/AI list in 2023.

Nick Frosst

Co-Founder

Former researcher at Google Brain; ranked #1 on 2023 Maclean's AI Trailblazers Power List.

Ivan Zhang

Co-Founder

AI researcher who previously collaborated with Aidan Gomez at FOR.ai before co-founding Cohere.

Funding history

Series A $125M September 2021 Led by Radical Ventures · Unknown

Series C $270M June 2023 Led by Unknown · Inovia Capital, Salesforce Ventures, Oracle

Series D $361.55M June 2024 Led by Unknown · PSP Investments, Export Development Canada, Cisco, AMD, Fujitsu

Series D Extension $500M August 2025 Led by Radical Ventures, Inovia Capital · AMD Ventures, NVIDIA, PSP Investments, Salesforce Ventures, HOOPP

Series D Secondary Close $100M September 2025 Led by Unknown · Business Development Bank of Canada, Nexxus Capital Management

Total raised: $1.54B

Industries

Artificial Intelligence (AI) Generative AI Natural Language Processing Machine Learning

Pricing

Pay-as-you-go token-based pricing for API access, charged on input and output tokens. Enterprise custom pricing available for North platform and private deployments.

Notable customers

Not publicly disclosed by company

Integrations

AWS Bedrock, AWS SageMaker, Google Cloud, Microsoft Azure (AML), Oracle GenAI Service, enterprise data connectors for RAG

Tech stack

React (JavaScript frameworks) Next.js (Web servers) dc.js (JavaScripty graphics) Headless UI (UI frameworks) Webpack PWA Open Graph Sanity (CMS) HSTS (Security) Node.js (Programming languages) Google Workspace (Email) Netlify (PaaS) Cloudflare (CDN) HubSpot (Marketing automation) Priority Hints (Performance) Segment (Customer data platform)

Website

cohere.com/

Competitors

OpenAI

OpenAI focuses primarily on consumer and commercial API access; Cohere emphasizes enterprise deployment flexibility and on-premises options.

Anthropic

Anthropic develops Claude models with emphasis on safety/alignment; Cohere targets enterprise workloads with flexible deployment and retrieval tools.

Google (Gemini/Vertex AI)

Google offers broad cloud ecosystem integration; Cohere provides vendor-agnostic deployment across multiple cloud providers and private clouds.

Meta (Llama)

Meta provides open-source models; Cohere offers managed, fine-tuned enterprise models with dedicated support and platform services.

Why this matters: Cohere stands out as the only major LLM company founded and headquartered in Canada with a founding team that authored the transformer architecture paper. The company has secured $1.54B in funding at a $7B valuation while maintaining a clear enterprise focus—emphasizing deployment flexibility, security, and specialized models (Rerank, Embed) rather than competing on general-purpose chat. The 2025 launch of North and recent $500M funding round signal momentum in the agentic AI space.

Best for: Enterprise teams building production-grade generative AI applications that require deployment flexibility, strong security, and enterprise-grade language models without vendor lock-in.

Use cases

Enterprise Search and RAG

Organizations use Cohere's Embed and Rerank models to enhance search accuracy within internal document repositories and knowledge bases. By combining semantic embeddings with reranking, enterprises improve both relevance and speed of information retrieval grounded in proprietary data.

Conversational AI and Chatbots

Customer-facing and internal teams deploy Command models via the Chat endpoint to build intelligent agents for customer support, employee Q&A, and workflow automation. RAG integration ensures responses are grounded in company-specific data and context.

Multi-Cloud and On-Premises Deployment

Regulated industries (finance, healthcare) deploy Cohere models on private cloud infrastructure or across multiple cloud providers (AWS, Azure, GCP) to meet data sovereignty and compliance requirements without sacrificing model quality.

Workplace Automation with North

Teams use Cohere's North platform to automate routine administrative tasks, accelerate complex workflows, and surface business insights by securely grounding agentic AI in enterprise data sources—reducing manual work and decision latency.

Alternatives

OpenAI API OpenAI prioritizes ease of use and consumer adoption; Cohere emphasizes enterprise deployment flexibility and on-premises options.

Anthropic Claude API Claude focuses on safety and interpretability; Cohere targets enterprise workloads with specialized models (Rerank, Embed) and multi-deployment architecture.

Llama (Meta) Llama is open-source and requires self-hosting and fine-tuning expertise; Cohere provides managed, enterprise-optimized models with platform services and support.

Google Vertex AI Vertex AI locks customers into Google Cloud ecosystem; Cohere supports AWS, Azure, GCP, Oracle, and private cloud deployments.

FAQ

What does Cohere do? +

Cohere develops large language models (Command, Rerank, Embed) and enterprise AI platforms that enable developers and enterprises to build production-grade generative AI applications. The company offers flexible deployment across hosted APIs, cloud platforms (AWS, Azure, GCP, Oracle), and private infrastructure, with advanced retrieval-augmented generation (RAG) and agentic capabilities through its North workplace platform.

How much does Cohere cost? +

Cohere uses a pay-as-you-go token-based pricing model for API access, charging for both input and output tokens. Specific pricing tiers are not publicly available; enterprise customers and North platform users can contact Cohere for custom pricing based on volume and deployment requirements.

What are alternatives to Cohere? +

OpenAI API (general-purpose LLM access), Anthropic Claude (safety-focused models), Meta Llama (open-source), and Google Vertex AI (cloud-native). Each differs in deployment flexibility, model specialization, and ecosystem lock-in.

Who uses Cohere? +

Enterprise software companies, large organizations, and technical teams building production generative AI applications—particularly those requiring deployment flexibility, security/compliance, or vendor-agnostic infrastructure. Notable investors including Cisco, AMD, Salesforce, and Oracle suggest adoption across enterprise software, hardware, and SaaS sectors.

How does Cohere compare to OpenAI? +

Cohere emphasizes enterprise deployment flexibility (multi-cloud, on-premises, VPC), specialized models (Rerank for search, Embed for semantic tasks), and managed platform services. OpenAI prioritizes ease of use and broad consumer adoption. Cohere is better for regulated industries and multi-cloud strategies; OpenAI is better for simplicity and rapid prototyping.

What makes Cohere's founding team notable? +

Co-founder Aidan Gomez was one of eight authors of 'Attention Is All You Need' (2017), the seminal paper that introduced transformer architecture—foundational to all modern LLMs. At age 20, as a Google Brain intern, he helped pioneer the technology underlying ChatGPT and modern AI. The team was ranked #1 on Maclean's 2023 AI Trailblazers list.