---
title: 'Cohere — company profile'
canonical: 'https://feeny.ai/companies/cohere'
type: 'company'
updated: '2026-07-02'
---

# Cohere

> Enterprise AI company building secure, privately deployable foundation models and an agentic workspace (North) for regulated businesses.

- **Website:** https://cohere.com/
- **Type:** Private (Series D)
- **Headquarters:** Toronto, Canada and San Francisco, CA (co-headquartered)
- **Founded:** 2019
- **Founders:** Aidan Gomez, Nick Frosst, Ivan Zhang
- **Valuation:** $7B
- **Total raised:** ~$1.5B
- **Latest round:** Series D extension · $100M · Sept 2025 (following $500M at $6.8B in Aug 2025)
- **Investors:** Radical Ventures, Inovia Capital, AMD Ventures, NVIDIA, PSP Investments, Salesforce Ventures, HOOPP (Healthcare of Ontario Pension Plan), BDC (Business Development Bank of Canada), Nexxus Capital Management
- **Business model:** Enterprise AI: custom enterprise contracts for the North and Compass platforms and private/dedicated deployments, plus pay-as-you-go token billing for direct API access to its models.
- **Industries:** Artificial Intelligence
- **Open roles:** 128
- **Profile:** https://feeny.ai/companies/cohere

## What they do

Cohere builds enterprise-grade AI foundation models (generation, embedding, reranking, and speech) plus North, a turnkey AI workspace, all designed to run privately inside a customer's own cloud or on-prem environment. Its whole differentiator is data sovereignty: enterprises deploy and customize the models without their data ever leaving their control.

## Overview

**The enterprise AI shop that lets you keep the keys**

Most foundation-model companies want your data flowing through their cloud. Cohere built the opposite pitch: run our models inside your own walls, on your VPC or on-prem, and we never see what goes in or comes out. That bet on data sovereignty is the whole company, and it is why banks, hospitals, and governments keep showing up on the customer list.

Founded in 2019 in Toronto by Aidan Gomez, Nick Frosst, and Ivan Zhang, Cohere has raised close to $1.5 billion and now carries a $7 billion valuation. It crossed $240 million in annual recurring revenue in 2025 and is openly positioning for an IPO, which is a rare thing to say out loud in a market crowded with OpenAI and Anthropic.

## What They Do

**Frontier models you deploy, plus the workspace that runs on them**

Cohere sells two things that fit together. One is a family of frontier models: Command for generation and agentic reasoning, Embed and Rerank for retrieval, and Transcribe for speech. The other is North, a turnkey AI workspace that stitches those models into agents, search, and document tools your team actually uses at work.

The common thread is control. Every product can run privately, be trained on your proprietary data, and stay auditable, which is the part enterprises in regulated industries care about more than raw benchmark scores.

## Problems

**Getting real AI into regulated shops without leaking the data**

The hard part of enterprise AI is not the demo, it is deploying something useful without shipping sensitive data to a third party. Cohere's answer is private deployment inside a customer's own cloud, on-prem environment, or a Cohere-managed Model Vault, so the data never leaves. North layers on secure, grounded retrieval across code, docs, tickets, and CRM records, then lets agents actually act: triaging issues, drafting release notes, running due diligence, and automating back-office work.

The payoff Cohere points to is accuracy in places where being wrong is expensive. One banking customer put it bluntly: they do not need a model that talks like Shakespeare, they need financial advice that is correct every time.

### Problems addressed

- Deploying enterprise AI without sending sensitive data to a third party
- Grounding AI answers in an organization's own code, docs, tickets, and CRM data
- Automating routine and complex workplace workflows with trustworthy agents
- Searching across noisy, multilingual, and multimodal business data
- Meeting regulatory and audit requirements for AI outputs

## Who It's For

**Built for the enterprises that cannot let data walk out the door**

Cohere aims squarely at large, regulated organizations: financial services, healthcare, manufacturing, telecom, energy, technology, and the public sector. These are buyers who will trade a little frontier flash for auditable outputs, private deployment, and a compliance story.

The pitch lands with two audiences at once. Business and ops leaders get North's workplace automation, while developers get direct API access to Command, Embed, and Rerank for building their own RAG and agentic pipelines.

### Ideal customer profiles

- **Business and operations leaders** — Automating routine tasks and complex workflows; Getting grounded insights from fragmented internal data
- **Financial services teams** — Accurate, auditable AI for regulated work; Due diligence, risk analysis, and back-office automation
- **Engineering teams** — Context-aware answers across code, docs, and tickets; Automating release and CI workflows
- **Developers** — Building RAG and agentic pipelines with reliable retrieval; Accessing frontier models via a self-serve API

## Products

**A model stack for retrieval, plus the platform on top**

Cohere's lineup splits cleanly into models and platforms. Command handles generation and agentic work, Embed and Rerank do the heavy lifting for search and RAG, and Transcribe covers speech. Rerank is the quiet standout: it sits between retrieval and the model, scoring documents for true relevance, and it is the piece reviewers most often single out.

On top sit North, the AI workspace for agents and automation, and Compass, an enterprise search system that indexes noisy, multilingual, multimodal data so teams can actually find answers across it.

## Business Model

**Enterprise contracts up top, pay-as-you-go for developers**

Cohere makes money two ways. North and Compass are sold on custom enterprise pricing, negotiated per deployment, which is where the big regulated accounts live. The models are more self-serve: a free, rate-limited trial API key to start, then pay-as-you-go token billing once you move to a production key.

For customers who want dedicated capacity, Model Vault runs Cohere models on managed instances priced per instance and performance tier, billed hourly or on longer commitments. It is a classic land-and-expand shape: start on the API, graduate to enterprise contracts and private deployment.

### Plans

Cohere prices in two lanes. The North and Compass platforms are sold on custom enterprise pricing, quoted per deployment. The models are self-serve: a free, rate-limited Trial API key to start, then pay-as-you-go token billing once you get a Production key, invoiced monthly or when the balance hits $250. Customers wanting dedicated capacity use Model Vault, priced per instance and performance tier.

- **North** — Custom · Enterprises
  All-in-one AI workspace to automate tasks and workflows safely
  - Intuitive interface
  - Purpose-built generative models
  - Intelligent search
  - AI agents for routine tasks and complex workflows
- **Compass** — Custom · Enterprises
  Intelligent search and discovery across your business data
  - Pre-built data connectors
  - Intelligent search
  - Document parsing
  - Managed index
- **Model Vault** — From $2,500/mo per instance · Teams needing dedicated, managed model capacity
  Fully managed, dedicated model deployment priced per instance and tier
  - Embed 4 (Small) at $4.00/hr or $2,500/mo
  - Embed 4 (Medium) at $5.00/hr or $3,250/mo
  - Rerank 3.5 / Rerank 4 (Medium) at $5.00/hr or $3,250/mo
  - Rerank 4 Pro (Large) at $10.00/hr or $6,500/mo
  - No shared resources or multi-tenancy overhead
  - Fixed or Flex pricing plans
- **Trial API key** — $0 · Developers evaluating the API
  Free, rate-limited access for testing (not for production)
  - Free API calls
  - Rate limited
  - Not permitted for production or commercial use
- **Production API key** — Pay-as-you-go · Developers running in production
  Token-metered production access at scale
  - Per-token billing
  - Billed monthly or at $250 outstanding balance
  - Requires Owner privileges to enable
  - Rerank billed per search unit (one query, up to 100 documents)

Good to know: Model Vault billing can be hourly, monthly, or annual; Bespoke model customization and private deployment are custom-priced for enterprise; A Rerank search unit is one query with up to 100 documents; documents over 500 tokens are split into chunks

## Competition

**Not the smartest model, the one you are allowed to run**

Cohere competes with OpenAI, Anthropic, and Mistral, and it does not pretend to win the raw-reasoning leaderboard. Reviewers note that GPT and Claude edge out Command on general benchmarks. Cohere's counter is deployment: VPC and on-prem options the big labs largely do not offer, a genuinely differentiated Rerank model for RAG, and cost-efficient retrieval at scale.

The strategic bet now is the agentic enterprise, anchored by North and reinforced by hardware and cloud partnerships including AMD, NVIDIA, Oracle, and the major clouds.

### Their edge

- **Deployment the big labs won't match** — VPC and on-prem options mean data stays inside the customer's environment, a fit OpenAI and Anthropic largely don't offer.
- **Rerank as a RAG moat** — Rerank sits between retrieval and generation and is widely treated as best-in-class for scoring document relevance.
- **Cost-efficient retrieval at scale** — Embed is priced to compete on document-retrieval workloads while matching or beating rival embedding models on benchmarks.

### Where they're betting

- The agentic enterprise, anchored by North
- Sovereign and private AI for regulated industries and governments
- Hardware and cloud partnerships (AMD, NVIDIA, Oracle, major clouds)
- Global expansion
- A potential IPO

## Proof

**The names and numbers Cohere puts on the table**

The traction is real and named. Cohere's technology is embedded across more than 100 generative AI use cases in Oracle Fusion Cloud Applications, and it has piloted or deployed North with RBC, Dell, LG, Palantir, Ensemble Health Partners, and Bell.

The business numbers back the logos: revenue crossed $100 million ARR in May 2025 and hit roughly $240 million by year end, with quarter-over-quarter growth above 50% through 2025. Custom models add technical proof too, like an 83.9% accuracy jump on HumanEval for code-tuned models.

## What People Say

**Loved for deployment and RAG, dinged on raw reasoning**

The recurring praise is consistent: Cohere wins on deployment flexibility that the big labs will not match, Rerank is treated as best-in-class for RAG, and Embed is priced to compete for retrieval at scale. Enterprise buyers also value the security posture, with SOC 2 Type II, ISO 27001, and GDPR coverage.

The recurring knock is just as clear. For pure reasoning on general tasks, GPT and Claude are rated ahead of Command, so Cohere tends to win on fit and control rather than sheer model horsepower.

## Funding

**$7B valuation and an IPO in the crosshairs**

Cohere raised an oversubscribed $500 million round in August 2025 at a $6.8 billion valuation, led by Radical Ventures and Inovia Capital, with AMD Ventures, NVIDIA, PSP Investments, Salesforce Ventures, and HOOPP joining. A month later a $100 million extension nudged the valuation to $7 billion alongside an AMD partnership.

Added to its earlier Series A through D rounds, that brings total funding to nearly $1.5 billion. With ARR climbing past $240 million, Cohere is one of the few in this cohort talking openly about going public.

## Team & Culture

**A Toronto research lab that grew into a global company**

Cohere started as a Toronto research shop and still carries that DNA: it co-headquarters out of Toronto and San Francisco, runs offices in New York, London, Montreal, Paris, and Seoul, and operates remote-friendly across time zones. Its open-science arm, Cohere Labs, has grown to 4,500+ community members and 100+ published papers.

The culture leans research-first and high-agency, valuing autonomy, openness, and momentum. Engineering JDs ask for people who act, tinker, and drive change without waiting for perfect instructions, and who treat correctness and reliability as their own responsibility.

## Compensation

**Global bands in five currencies, with equity for everyone**

Cohere discloses pay across a genuinely global footprint, so its bands span USD, CAD, EUR, and more, reflecting roles from Toronto to London to the Middle East. US engineering roles run to the high six figures, while sales, marketing, and operations bands sit lower and vary by region.

Equity is part of the package across roles, and the company frames employees as co-owners. Benefits are unusually generous, including six weeks of paid vacation and six months of fully paid parental leave.

## Security & Legal

**Certified, and built so Cohere never touches your data**

The registered entity is Cohere Inc., and its whole security story is that private and third-party deployments stay walled off. Cohere states it has no access to data processed in private deployments or on third-party managed AI services like AWS, Azure, or GCP, and it does not train on enterprise customer data.

The certifications back it: SOC 2 Type II, ISO 27001, ISO 42001 for responsible AI, plus GDPR coverage with Standard Contractual Clauses for EU data. A public Trust Center lets customers pull the actual reports.

## In the News

**A raise, a launch, and a path to public markets**

The last year of coverage tracks Cohere's move from lab to IPO candidate. North launched publicly in August 2025 as its flagship agent platform, the funding rounds pushed the valuation to $7 billion, and by early 2026 the story turned to a $240 million revenue year setting the stage for a public listing.

### Coverage

- [Cohere's $240M year sets stage for IPO](https://techcrunch.com/2026/02/13/coheres-240m-year-sets-stage-for-ipo/) — TechCrunch (2026-02-13)
- [Cohere hits $7B valuation a month after its last raise, partners with AMD](https://techcrunch.com/2025/09/24/cohere-hits-7b-valuation-a-month-after-its-last-raise-partners-with-amd/) — TechCrunch (2025-09-24)
- [Cohere raises $500M at $6.8B valuation to accelerate enterprise efficiency with agentic AI](https://cohere.com/blog/august-2025-funding-round) — Cohere (2025-08-14)
- [Cohere's new AI agent platform, North, promises to keep enterprise data secure](https://techcrunch.com/2025/08/06/coheres-new-ai-agent-platform-north-promises-to-keep-enterprise-data-secure/) — TechCrunch (2025-08-06)
- [Cohere achieves ISO 42001 and ISO 27001 certifications](https://cohere.com/blog/iso-42001-and-iso-27001-certifications) — Cohere

## Outlook

**Sovereignty is the moat, if it holds**

Cohere has picked a lane and committed to it: not the flashiest model, but the one a bank or a government is actually allowed to run. With $240 million in ARR, a $7 billion valuation, and North as the centerpiece, the growth is real and the IPO talk is not idle.

The risk is the same as the strategy. If OpenAI and Anthropic close the deployment and privacy gap, Cohere's edge narrows to Rerank and price. For now, sovereignty plus a strong retrieval stack is a defensible place to stand.

## Company details

- **Mission:** Cohere believes AI's highest purpose is to enhance human wellbeing, and aims to scale intelligence to serve humanity by empowering businesses to boost productivity and innovation.
- **Products:** Command, Embed, Rerank, Transcribe, North, Compass
- **Notable customers:** Oracle, RBC, Dell, LG, Palantir, Ensemble Health Partners, Bell, Fujitsu
- **Customer segments:** Financial services, Healthcare, Manufacturing, Telecommunications, Energy, Technology, Public sector / government, Developers
- **Buyers / users:** Business and operations leaders, CX and support leaders, Investment and financial services teams, Developers building RAG and agentic apps, Security and compliance leaders
- **Competitors:** OpenAI, Anthropic, Mistral
- **What sets them apart:** Private deployment (VPC, on-prem, or Model Vault) so customer data never leaves their control; Does not train on enterprise customer data; Differentiated Rerank model for RAG accuracy; Deep model customization on proprietary data and languages; Strong open-science research base via Cohere Labs
- **Tech stack:** VPC / on-premises deployment, Cohere-managed Model Vault, AWS, Microsoft Azure, Google Cloud Platform, Oracle Cloud Infrastructure, IBM Cloud
- **Integrations:** AWS, Microsoft Azure, Google Cloud Platform, Oracle Cloud Infrastructure, IBM Cloud, Oracle Fusion Cloud Applications

## Open roles (128)

- [Head of Investor Relations](https://jobs.ashbyhq.com/cohere/d82b9edf-f2d9-489b-b1f0-b65d976348ab) — Remote
- [Head of Strategic Finance](https://jobs.ashbyhq.com/cohere/39ae8da6-9c3d-4f36-80a3-310159ee1f8d) — Remote
- [Senior Product Designer](https://jobs.ashbyhq.com/cohere/479bed98-46d3-4140-9d5e-96dc193fd845) — New York, NY
- [Infrastructure Security Engineer (Secret + Clearance)](https://jobs.ashbyhq.com/cohere/e48c5456-6ff5-4d4b-a632-df0a3dc497ab) — Remote
- [Machine Learning Intern/Co-op  (Fall, 2026)](https://jobs.ashbyhq.com/cohere/36d1f52f-8270-4652-adf5-5303a0ff341b) — Remote
- [Forward Deployed Engineer, Infrastructure Specialist (North America)](https://jobs.ashbyhq.com/cohere/be48aafc-9610-4ebd-8414-a0722a3cd59a) — Toronto, Canada
- [Senior Technical Program Manager, Machine Learning Infrastructure](https://jobs.ashbyhq.com/cohere/582c4a60-14a2-49d5-b4d3-24c4ec50508d) — Remote
- [Senior Manager of Financial Planning & Analysis](https://jobs.ashbyhq.com/cohere/5adf044a-a601-467a-bb92-807850b4f7ec) — San Francisco, CA
- [Senior Member of Technical Staff, Multimodal AI](https://jobs.ashbyhq.com/cohere/443368a3-6276-4b90-9671-27fed40fd6d2) — Remote
- [Solutions Architect - Public Sector](https://jobs.ashbyhq.com/cohere/6d0f0753-ff22-46fd-90e9-08998914a8e7) — Washington, DC
- [IT Operations Specialist (12 month contract)](https://jobs.ashbyhq.com/cohere/686ef280-eea8-4012-b9e6-9cb23c0c22ab) — London, United Kingdom
- [Full-Stack Software Engineer, Inference](https://jobs.ashbyhq.com/cohere/f5f727bc-9eab-4fcc-9fdd-92a9ce47c37c) — Toronto, Canada
- [Counsel, Privacy and AI Regulation (EU)](https://jobs.ashbyhq.com/cohere/d81c28f5-2945-4c49-bc5b-86e345334cfb) — Remote
- [Manager, Security Engineering](https://jobs.ashbyhq.com/cohere/1599cc22-313e-49c9-ac69-b365f5f8f227) — United States
- [Chief of Staff](https://jobs.ashbyhq.com/cohere/5139acd7-3429-483e-9984-040bd799e8a0) — Remote
- [Product Marketing Manager](https://jobs.ashbyhq.com/cohere/5cf3bd63-cf1d-441a-9db5-2032ca8449b8) — Toronto, Canada
- [HR Business Partner](https://jobs.ashbyhq.com/cohere/8a28e25a-6d8c-403a-8c9a-5155cbe0ce97) — London, United Kingdom
- [Director of Product, North](https://jobs.ashbyhq.com/cohere/44af407e-28a2-4589-bede-6c5e8ee305bc) — Remote
- [Engineering Manager, FDE Infrastructure (NORAM)](https://jobs.ashbyhq.com/cohere/6a6120d5-5e02-4811-99d9-6baf0b910e37) — Remote
- [RevOps Analyst (Analytics)](https://jobs.ashbyhq.com/cohere/cd3eacfe-1169-4df0-8164-93a857d5ddf0) — New York, NY
- [Lead Member of Technical Staff, Inference Infrastructure](https://jobs.ashbyhq.com/cohere/5806a521-8584-4bbe-b60a-49031b919251) — Remote
- [Software Engineer, Security](https://jobs.ashbyhq.com/cohere/b9c8c98e-b0fa-43b6-93b0-fa780d956066) — Toronto, Canada
- [Member of Technical Staff - Sovereign AI](https://jobs.ashbyhq.com/cohere/b9306620-a110-46b5-b4c4-7363e1d03a8d) — Remote
- [Lead Data Scientist](https://jobs.ashbyhq.com/cohere/39cab400-c976-41d9-9002-880ac6985b64) — Remote
- [Member of Technical Staff, Modeling](https://jobs.ashbyhq.com/cohere/3136a5a5-06fd-4c82-8b72-a43467e6b128) — Remote
- [Senior Member of Technical Staff, Web Data](https://jobs.ashbyhq.com/cohere/82ec1512-5bc3-49dd-b995-b6abc2a958d7) — Remote
- [Member of Technical Staff, Search](https://jobs.ashbyhq.com/cohere/d5cc437c-2397-4c1a-9d12-3a2bf64d98ea) — Remote
- [Member of Technical Staff, MLE (UK/EU)](https://jobs.ashbyhq.com/cohere/876a2d31-e6c3-497b-9c8e-0b7d111fb4cd) — London, United Kingdom
- [Member of Technical Staff, Pre-Training Data](https://jobs.ashbyhq.com/cohere/859e2e47-02fb-4afe-bb8a-e83bf4d8c265) — Remote
- [HR Business Partner](https://jobs.ashbyhq.com/cohere/c1543c10-4c71-4bb9-a570-215a31e9e7fc) — Remote
- [Data Engineer, Data Foundations](https://jobs.ashbyhq.com/cohere/9baccd88-c051-474f-bfe8-6867fca54cee) — New York, NY
- [Manager of Technical Staff, Sovereign AI](https://jobs.ashbyhq.com/cohere/013d4545-8343-40f1-9de5-9fdc0f0dd5bb) — Toronto, Canada
- [Member of Technical Staff, MLE](https://jobs.ashbyhq.com/cohere/110ba167-4efd-43b7-85d2-3ff719a28b0f) — Remote
- [Member of Technical Staff, Training Performance Engineer](https://jobs.ashbyhq.com/cohere/d42f5fd4-1ffc-45b9-957c-f09862db6af6) — London, United Kingdom
- [Chief Information Security Officer (CISO)](https://jobs.ashbyhq.com/cohere/e0c86d1a-ee3a-4da6-9244-53f0909ab236) — Toronto, Canada
- [Forward Deployed Engineer, Infrastructure Specialist (Europe/Middle East)](https://jobs.ashbyhq.com/cohere/69a9ab2e-a7eb-4c4f-825a-5c25cf063900) — Remote
- [Data Annotation Specialist, Simplified Chinese / Mandarin](https://jobs.ashbyhq.com/cohere/34533921-d731-4529-89e7-cf4939055a20) — Remote
- [Member of Technical Staff, MLE [Singapore]](https://jobs.ashbyhq.com/cohere/a29a310d-074e-46f2-bdad-308d458d4011) — Remote
- [Senior/Staff Software Engineer, Developer Experience](https://jobs.ashbyhq.com/cohere/6d499dbb-8209-4f02-ae57-a1f1efa9ecbb) — Remote
- [Data Annotation Specialist, Data Science](https://jobs.ashbyhq.com/cohere/9845c80d-95ba-4769-94a7-62408d73d603) — Remote
- [Member of Technical Staff, Integration/RL Team (Research Engineer)](https://jobs.ashbyhq.com/cohere/e9ef9420-88e5-403b-82e8-1c36fa010d4d) — Remote
- [Product Manager, Integrations](https://jobs.ashbyhq.com/cohere/87e9fd4e-49ac-4242-97be-7da6f718020d) — Remote
- [Engineering Manager, FDE Infrastructure (UK)](https://jobs.ashbyhq.com/cohere/9bdbb821-6536-4090-8faa-a2b04a70bbf1) — Remote
- [Staff Software Engineer, Inference Infrastructure](https://jobs.ashbyhq.com/cohere/41f23dad-9da2-451a-bd1e-a1800437cb64) — San Francisco, CA
- [Software Engineer, North for Finance](https://jobs.ashbyhq.com/cohere/f722247c-291b-44ee-af67-5159b8d5d9b9) — Remote
- [Forward Deployed Engineer, Agentic Platform (Korea)](https://jobs.ashbyhq.com/cohere/031c2335-db85-4265-8a88-a54b911be36a) — Remote
- [RevOps GTM Systems Architect](https://jobs.ashbyhq.com/cohere/a03e02e4-76be-4666-a71d-e7b1d51e4a0f) — Remote
- [Member of Technical Staff, Training Infra Engineer](https://jobs.ashbyhq.com/cohere/a13207e7-dc82-473f-8ca4-e832452fe8c3) — Remote
- [Head of Sales Enablement](https://jobs.ashbyhq.com/cohere/53c7d470-711d-4b8f-8781-84ee925f184d) — New York, NY
- [Marketing Chief of Staff](https://jobs.ashbyhq.com/cohere/061d7ee4-91c5-4a13-aff6-20d05bf80abb) — Remote

_…and 78 more at https://feeny.ai/companies/cohere/jobs_

---
_Source: https://feeny.ai/companies/cohere · profile updated 2026-07-02_