Cohere: winning enterprise AI without consumer scale

The client and the question

A Middle Eastern sovereign investor building exposure to artificial intelligence asked Northern Lights to assess Cohere as a potential investment and partner. It wanted to understand whether a company without a consumer product could compete with the largest model developers.

Can an enterprise-first AI company build a durable, profitable position, and how much does government support matter?

Context

Cohere focuses on enterprise use cases, with models and tools designed for secure deployment, retrieval and multilingual applications. It has deliberately avoided the consumer chatbot race.

Governments are increasingly treating AI capability as strategic infrastructure, supporting national champions through compute investment, procurement and funding. This creates new demand for providers able to deploy in sovereign environments.

In enterprise AI, distribution and trust can matter more than scale.

Key questions

1. Differentiation. What lets an enterprise-first model developer win deals against hyperscalers and consumer-scale competitors?
2. Economics. Can enterprise and government contracts generate the revenue needed to fund frontier research?
3. Sovereign angle. How do government partnerships change the risk profile and the investment case?

Our approach

We assessed Cohere through the lens of an investor seeking both financial return and strategic access to AI capability.

1. Market positioning. Benchmark of enterprise AI providers on deployment options, security, multilingual capability and pricing.
2. Revenue quality. Analysis of customer segments, contract types and the path to sustainable margins.
3. Strategic fit. Assessment of partnership options for a sovereign investor, from minority stake to joint deployment programmes.

Key insights

1. Focus is a strategic asset.
   By staying out of the consumer race, an enterprise-first developer can concentrate resources on security, private deployment and integration, the attributes large organisations actually buy.
2. Government demand de-risks the path to scale.
   Sovereign procurement and compute partnerships provide anchor demand and credibility, but they also tie the company’s fortunes to public budgets and policy cycles.
3. Value lies in co-deployment, not just equity.
   For a sovereign investor, the strongest case combines a financial stake with a deployment partnership that builds local capability, data infrastructure and skills.

Outcome

1. An investment thesis with clear conditions.
   The client received a balanced view of the upside and the key risks, with the milestones that would justify a larger commitment.
2. A partnership blueprint.
   We outlined how a stake could be paired with a local deployment programme serving government and strategic industries.
3. A framework for AI investments.
   The approach was reused to screen other model developers and AI infrastructure opportunities.

For discussion

1. Can a model developer without consumer scale keep up with frontier research?
2. What should a sovereign investor ask for beyond financial returns?
3. How much should government support weigh in valuing an AI company?