Most AI secondary demand clusters around the names that dominate headlines — OpenAI, Anthropic, xAI. Cohere rarely makes that list, which is precisely why it deserves a close look. Quieter demand can mean more rational pricing. It can also mean thinner supply and harder price discovery. Understanding which dynamic is at work matters before you submit an indication.
What Cohere actually is, and why it matters for valuation
Cohere is an enterprise-grade large language model (LLM) platform. Unlike OpenAI's consumer and API mix, or Anthropic's safety-research identity, Cohere is almost entirely B2B. Its primary product — Command — is sold to large organizations that need LLMs deployable inside their own cloud environments, including on-premises infrastructure. That means Cohere competes for procurement budgets, not developer sign-ups.
The business model consequence is meaningful. Enterprise contracts tend to be larger, stickier, and slower to ramp than consumption-based API revenue. Cohort retention is high once a deployment is embedded in a client's workflow. But sales cycles are long, which creates a lag between business development activity and recognized revenue. Buyers looking at secondary pricing need to factor that lag into any revenue-multiple analysis.
Cohere raised its Series D at a reported valuation in the $5 billion range in mid-2024. Secondary marks have traded at discounts and premiums relative to that level depending on when a lot changed hands and who the seller was. If no primary round has closed since that mark, the secondary price is floating against a reference point that is now more than a year old — a gap worth acknowledging.
The three questions every Cohere buyer should answer
1. Am I buying common stock, preferred stock, or an SPV interest in one of those?
Cohere, like most venture-backed companies, has issued multiple preferred series as well as common stock typically held by employees and early team members. The economic gap between preferred and common can be significant if the company's exit valuation is modest relative to the liquidation preference stack that preferred holders have accumulated. Before signing anything, confirm which share class underlies the lot you are buying. If you are buying through a special purpose vehicle (SPV), the SPV operating agreement should specify this clearly. If it does not, ask.
Common shares from employees often trade at steeper discounts to the last preferred round price than preferred shares do. That discount may reflect genuine economic subordination, or it may represent an overshoot driven by seller urgency. Modeling the liquidation waterfall — even roughly — helps you decide which it is.
2. What does Cohere's transfer policy actually permit?
Enterprise-stage companies often have transfer restrictions embedded in their stockholder agreements or equity plan documents. Cohere, as a private company, controls consent over most direct transfers. If you are buying through an SPV that already holds the underlying shares, the consent question may already be resolved — the SPV manager should have completed that step. If you are attempting a direct transfer of shares, company consent is required and the timeline is uncertain. Right of first refusal (ROFR) may also apply: the company — and sometimes existing investors — can step in and purchase the shares at the agreed price before you take delivery. Ask your intermediary whether ROFR has been waived or is being cleared concurrently with execution.
3. What is my liquidity path, and over what horizon?
Cohere has not filed publicly for an IPO as of July 2026. There are no announced tender offers. That means secondary buyers are underwriting a private hold of uncertain duration. For a B2B AI infrastructure company, plausible liquidity paths include an IPO, a strategic acquisition by a cloud hyperscaler, or a future tender offer sponsored by the company or a lead investor. None of these is guaranteed, and none carries a predictable timeline. Buyers should size positions accordingly — this is not a stock with a six-month exit window.
How to think about discount to last round
Secondary buyers in the AI space often anchor on discount to last round as their primary valuation heuristic. For Cohere, that anchor is a mid-2024 preferred series price. A lot of has changed in the enterprise AI landscape since then: more competitors have shipped, hyperscalers have deepened their own model offerings, and enterprise procurement committees have grown more sophisticated about vendor selection.
A discount to last round is only useful if the last round price was well-calibrated to begin with. Late-stage venture rounds in 2023 and 2024 were sometimes priced on the high end of defensible, particularly in AI. Buying at a 20% discount to a potentially stretched primary price may not represent the margin of safety it appears to.
A more durable frame is to estimate what you think the business is worth on a forward revenue or ARR multiple basis, compare that to the implied valuation at the secondary price you are considering, and then decide whether the gap is adequate given the liquidity risk and hold period. This requires making assumptions — about Cohere's revenue trajectory, about competitive dynamics, about AI infrastructure spending growth — but those assumptions made explicit are more useful than a single percentage figure anchored to a stale primary round.
Supply dynamics on the Cohere secondary market
Cohere's seller population is narrower than the mega-cap AI names. The company has a smaller headcount than OpenAI or Anthropic, which means fewer employee secondary sellers at any given moment. Early employees and executives hold the largest blocks, but those holders often have longer time horizons and more restricted transfer terms. Supply tends to come in episodic waves — after a fundraise, when a lock-up or holding-period threshold is cleared, or when an individual employee's financial circumstances change.
Thin supply is not inherently bad for buyers — it can mean less competition for a lot. But it also means you may be waiting longer for the right lot at the right size, and that the seller you are transacting with has a specific reason to sell that is worth understanding. In a thin market, confirmed supply at time of indication is more valuable than a broad but unverified order book.
The enterprise AI infrastructure thesis in context
The bear case on Cohere is that hyperscalers — AWS, Azure, Google Cloud — will commoditize the enterprise LLM market by bundling their own models with existing cloud contracts. If CIOs can get adequate LLM capability from a vendor they already pay, the switching cost to a standalone platform like Cohere may prove too high.
The bull case is that enterprises with serious data-privacy requirements and multi-cloud mandates will always prefer a model provider that is neutral relative to the cloud layer. Cohere's ability to deploy on any cloud — or on-premises — is a genuine differentiator for regulated industries like finance, healthcare, and government. Those buyers pay premium pricing and sign multi-year contracts.
Secondary buyers are not obligated to resolve this debate definitively before transacting. They are, however, obligated to understand it — because the spread between the bear case and the bull case implies a very wide range of terminal outcomes. Position sizing should reflect that width.
Next step
If you are ready to explore current Cohere availability, browse confirmed seller supply on the Limen Markets marketplace. Minimum position size is $25,000. Settlement runs one to five business days once consent and ROFR questions are resolved. If you want to understand the mechanics of what you are buying before you indicate, the guide on preferred-vs-common economics and SPV waterfall mechanics are the right starting points.