Any offering would land in an increasingly crowded field. Rivals OpenAI and Anthropic are also evaluating potential listings, people briefed on their deliberations said, though neither has publicly set a timetable. Both competitors have made sizeable inroads with corporate users: OpenAI reported in November that more than 1 million businesses have adopted its technology, while Anthropic said in September it serves upward of 300,000 businesses. Those penetration rates underscore the scale advantages that Cohere will confront as all three companies race to lock in long-term platform commitments from global enterprises.
Cohere’s investor memo argues that the startup’s business model is more capital-efficient than those of many AI peers. Roughly 70% of revenue comes from software subscriptions rather than direct compute services, the document said, allowing customers to run the firm’s large language models through popular managed-cloud offerings or on their own hardware. By shifting infrastructure spending to clients and cloud partners, Cohere says it avoids the sizeable data-center outlays that have weighed on other AI vendors and can instead “invest more aggressively” in research, product development and customer acquisition.
Gross margin averaged around 70% in 2025, improving by 25 basis points from the previous year, according to the memo. Management attributed the steady margin expansion to a policy of scaling compute resources only when demand materializes, a tactic it claims shields the company “from speculative excesses” in the wider AI market and supports sustainable growth.
The company plans to maintain that approach while enlarging its geographic footprint. In 2026, Cohere intends to deepen operations in continental Europe and accelerate development of its AI agent platform, North. The memo projects another year of “rapid growth,” though specific revenue targets were not disclosed.
Imagem: Internet
Investor interest in enterprise AI has intensified as organizations look to deploy generative models that meet stringent security and compliance requirements. Research published by McKinsey & Company estimates that generative AI could add as much as $4.4 trillion in annual economic value, with the bulk stemming from business-specific applications rather than consumer tools. Cohere’s management contends the firm is well-placed to capture a share of that spending by offering flexible deployment options and limiting data that leave a customer’s environment.
Despite a favorable revenue trajectory, Cohere will need to balance expansion against a fast-evolving competitive landscape. Cloud hyperscalers such as Microsoft, Google and Amazon continue to integrate their own foundation models into enterprise software suites, while independent providers race to differentiate on cost, performance or policy controls. For Cohere, success in the public markets could hinge on demonstrating that its capital-light infrastructure strategy can translate into sustained margin performance even as compute requirements grow.
Management has not publicly commented on a prospective listing structure, but advisors expect Toronto and New York to feature prominently among venue options, reflecting the company’s Canadian roots and the deeper liquidity of U.S. exchanges. The timing may depend on broader equity-market sentiment toward high-growth technology issuers and on whether larger incumbents decide to pursue offerings of their own.
For now, Cohere is emphasizing operational metrics rather than a calendar date. “Our thesis is clearly resonating in the market,” the memo said, pointing to an expanding customer base in industries with strict data regulations. If the company maintains its current growth rate and cost profile, analysts say it could enter the public arena with one of the sector’s stronger combinations of revenue scale and gross margin.
Crédito da imagem: Bloomberg via Getty Images
