Cohere PM salary

Most published Cohere PM salary data is wrong. Either it mixes in Cohere Health (a completely separate healthcare scheduling company), or it reports a number in Canadian dollars without saying so. The median PM total comp on Levels.fyi is CA$190K, which converts to roughly US$140K at mid-2026 exchange rates. That is not a competitive US AI lab salary. It is a competitive Toronto tech salary, and the distinction matters for every candidate currently comparing Cohere against Anthropic or OpenAI.

The CAD reality

Cohere is headquartered in Toronto. Most PM roles, including senior and staff-level positions, are filled from the Toronto office or hybrid across Toronto and New York. Offers for Toronto-based roles are denominated in Canadian dollars. US-based candidates who receive a US-denominated offer are in a different bracket, typically set against New York or San Francisco market rates, but the majority of Cohere PM headcount earns in CAD.

Before benchmarking any Cohere offer, confirm the currency. A CA$200K offer is approximately US$148K. A US$200K offer is a different conversation.

PM comp by level

These ranges synthesize Levels.fyi (2 PM data points, both Canada), Blind (engineering and research data used as directional anchors), and live job posting signals:

Level	Toronto (CAD total comp)	US remote/NY (USD total comp, estimated)
PM	CA$150K - CA$185K	US$180K - US$230K
Senior PM	CA$185K - CA$230K	US$230K - US$300K
Staff/Principal PM	CA$230K - CA$290K	US$290K - US$380K

The sample size is thin. Treat these as directional ranges, not verified bands. Cohere engineering comp runs significantly higher: a Senior Gen AI Engineer on Blind reported US$370K total comp ($250K base, $120K equity), and a Senior Research Engineer reported US$468K TC. PM comp sits materially below engineering peaks at most AI labs, Cohere included.

Anthropic senior PM total comp is reportedly $460K+, which Cohere does not match on base. The gap is real. The question is whether the equity story closes it.

The equity story: $6.8B pre-merger, $20B combined entity implied

Cohere raised $500M at a $6.8B valuation in August 2025, followed by a further $100M bringing the total to approximately $7B. In April 2026, Cohere announced a merger with Aleph Alpha, the German enterprise AI company, at an implied combined valuation of roughly $20B. The merger is the single most important variable in any current equity conversation.

What this means for PM offers:

Equity granted at the $6.8B-$7B valuation has paper upside if the combined entity trades or IPOs above that mark. A $20B implied merger valuation is a meaningful signal.
Cohere reported $240M ARR (CNBC, February 2026). At 995 employees as of May 2026, that is roughly $240K ARR per employee, which is a healthy unit economics signal for a company at this stage.
The Aleph Alpha merger creates the largest enterprise-focused LLM company outside US hyperscalers. Cohere’s sovereign AI positioning (data residency for finance, government, healthcare) is a structural differentiation that no US-based competitor can credibly replicate for European regulated-industry customers.

Cohere equity is late-stage private. You cannot sell it until an IPO, acquisition close, or secondary tender. Before signing, ask the recruiter: current 409A valuation, strike price if options (vs. RSUs), vesting schedule, and whether any tender offers have run. The merger trajectory makes this a more interesting equity bet than a year ago, but it is still illiquid.

PM tracks and comp variation

Cohere is not one product. The PM roles map to distinct product lines, and the scope and comp signals differ:

North platform (enterprise agentic AI): iOS, Android, macOS, web. This is the highest-profile PM surface at Cohere, with mobile/native experience requirements and growth PM skills explicitly listed in postings. Likely the top of the PM comp range.
Command (multimodal generation): The flagship model product. API and enterprise integrations. PMs here need deep model customization and eval design fluency.
Embed and Rerank (retrieval): Developer-facing, heavy on RAG architecture knowledge. PMs who understand vector search, chunking strategy, and retrieval pipeline tradeoffs are preferred.
Transcribe (speech) and Compass (search/discovery): Newer surfaces, smaller teams, more generalist PM scope.
North Mini Code (software engineering assistant): Closest analog to GitHub Copilot. PMs with developer tooling experience are prioritized.

North platform and Command roles likely sit at the top of the PM band. Embed/Rerank PMs working on enterprise retrieval products are increasingly valued given Cohere’s positioning as the retrieval-layer vendor for regulated industries.

What commands a premium at Cohere specifically

Cohere job postings (Ashby, June 2026) require 5+ years PM experience and meaningful depth in at least one of: mobile/native product management, platform product management, or growth PM. Beyond that baseline, the skills that move offers higher:

Retrieval and RAG depth. Cohere competes on retrieval quality. PMs who can speak to chunking, embedding model selection, reranking tradeoffs, and eval design for retrieval pipelines are not common.
Enterprise compliance fluency. Sovereign AI and data residency deals require PMs who understand regulated-industry procurement, security certifications (SOC 2, ISO 27001), and the sales cycle for government or financial services buyers.
Eval design. Cohere’s enterprise customers need to measure model behavior against their own ground truth. PMs who have built or overseen eval harnesses are directly useful.
A/B instrumentation on non-deterministic outputs. Standard testing frameworks break on generative AI outputs. Knowing what to measure and how to instrument it is a rare PM skill Cohere values.

Should you take the offer

If you are a mid-career PM at a US company currently earning US$250K+ TC, a Toronto-denominated Cohere offer will look like a pay cut unless you are comparing net-of-tax Toronto cost of living or the US-denominated offer is in the $230K+ range. The base and total comp trail Anthropic and OpenAI for equivalent levels.

The case for taking it: Cohere has $240M ARR (real revenue, real enterprise customers), a merger that could create a $20B entity, and a sovereign AI moat that is genuinely defensible. For a PM who wants to own a meaningful product surface at a company at an inflection point, the scope is real. For a PM optimizing purely for near-term cash compensation, the math does not work against top US offers.

The 2026 PM question at any enterprise AI lab is not “will enterprise AI work” but “which platform wins the compliance-first segment.” Cohere has a structural claim on that segment. The equity upside is contingent on whether that claim converts to market share before the next liquidity event. That is the bet you are evaluating, not just the salary.