Cohere
The enterprise-only foundation lab. No consumer, no drama, no distraction.
cohere.com ↗Cohere is the enterprise-AI foundation model company for buyers who don't want to be anyone's consumer. Gomez made the strategic choice early — no ChatGPT clone, no viral consumer launch, no social-media attention — and it looks prescient as enterprise procurement teams increasingly select vendors based on alignment of focus. The $240M ARR and 50% QoQ growth are real numbers being reported to investors; the IPO narrative is credible. For enterprise AI buyers, Cohere is the safe-choice foundation model. For investors, it's the cleanest pure-play on the enterprise AI market with a non-US base.
What's going for them.
- 01Enterprise-first positioning from day one — no consumer chatbot, no free tier, no distraction. The discipline has become a major advantage as enterprise buyers look for an AI vendor whose priorities align with theirs.
- 02ARR doubled through 2025 to $240M+, with 50%+ quarter-over-quarter growth. Exceeded the $200M target management committed to, now pacing toward IPO-ready scale.
- 03Cap table uniquely positioned for the sovereign-AI buyer: NVIDIA + AMD + Salesforce + major Canadian pension funds (PSP, HOOPP) + Korean partnerships (LG). A non-US institutional anchor OpenAI and Anthropic can't replicate.
- 04The Command and Embed model families are the benchmark for enterprise RAG workloads — retrieval-augmented generation at high accuracy, low hallucination, with deployment options across cloud, on-prem, and air-gapped environments.
- 05Aidan Gomez is a co-author of the original Transformer paper. Co-founder research pedigree matters for both model quality and enterprise trust in a category where 'who built this' is a procurement criterion.
What they built
Cohere is an enterprise-focused foundation model company. The product stack centers on the Command family (reasoning and generation), the Embed family (retrieval and vector search), and the Rerank product (retrieval quality) — a triad that covers the full enterprise RAG stack better than any generalist lab. Deployment options span North, the enterprise AI platform; API access via Cohere, Oracle Cloud, AWS, and Azure; and on-prem or air-gapped installations for regulated customers. No consumer product, by design.
How they got here
Gomez, Frosst, and Zhang — all ex-Google Brain, with Gomez a co-author of the original “Attention Is All You Need” transformer paper — founded Cohere in Toronto in 2019. The early years were quieter than the flashier US labs: no Sam Altman-style media tour, no ChatGPT moment. Instead the team built the enterprise product, the deployment infrastructure, and the specific technical depth (retrieval, embedding, grounding) that enterprise buyers would need.
The slow-burn approach started paying off in 2024, and 2025 was the inflection year. Revenue accelerated: $100M ARR mid-year, $240M+ by end-year, 50%+ QoQ growth reported in investor updates. The August 2025 $500M round at $6.8B (extended to $7B in September with an additional $100M) was led by Radical Ventures and Inovia, with participation from NVIDIA, AMD, Salesforce, and two major Canadian pension funds. The cap table was a deliberate construction: sovereign-adjacent capital from a non-US base, ready to carry the company through an IPO-bound trajectory.
What’s ahead
Three things define the next 12 months. First, product breadth: Cohere has been disciplined in scope (RAG, enterprise search, embeddings), but customers increasingly want agentic capabilities, voice, and multimodal. The question is whether Cohere can expand without losing the focus that’s been its differentiator. Second, IPO path: revenue run-rate and growth rate put Cohere in genuine IPO range. A 2026 or early 2027 filing is plausible, and Cohere would be the first foundation-model company to go public — a category-defining event. Third, the LG/Asia motion: the Korean and broader Asia-Pacific enterprise pipeline (via LG, Fujitsu, Samsung-adjacent work) is underappreciated and could be the piece that distinguishes Cohere from pure-US competitors.
Why it matters
Cohere is the single cleanest study on whether focused enterprise AI can compete with horizontal-consumer AI labs for the same enterprise dollars. For enterprise-AI founders, Cohere’s playbook — deep product focus, no consumer distraction, patient enterprise sales — is the clearest counter-example to the ChatGPT-style distribution strategy. For investors, it’s the most plausible first-IPO candidate in the foundation-model category, and a non-US hedge on a cap table otherwise dominated by Silicon Valley exposure.
Founder interview coming soon.
We'll be sitting down with the founders and operators of the companies we profile — on fundraising, product decisions, and what they're building next. If you're part of the Cohere team and want to share a perspective, get in touch.
Thinking about fundraising or M&A?
Amafi Advisory works with AI companies on strategic, fundraising, M&A, and technical advisory. Even if you're just exploring.