Cohere Struggles to Secure Compute Power
Cohere's CEO says the Canadian AI firm is short on the computing capacity it needs to run its operations.
Updated on Nov. 11, 2026 in Data Centers

Live Poll
Do you think AI companies are spending too much on data centers right now?
Cohere CEO Aidan Gomez said the company is struggling to obtain the computing power required to run its operations. The comments came as the firm announced it is raising $7 billion at a $35.7 billion valuation, including up to $1.6 billion from the Canadian government.
Why it matters
Cohere faces a broader shortage of compute, energy, facilities and chips, and its reliance on debt-backed infrastructure providers could leave it exposed if tighter lending slows their expansion. The situation highlights the capital-intensive demands facing Canadian AI companies.
Cohere is raising equity rather than debt and depends on debt-backed infrastructure providers such as CoreWeave for compute capacity. The source does not specify how much additional capacity the company needs or when it expects to secure it.
The players
Cohere
A Canadian artificial intelligence company that develops large language models for enterprise customers and is among the country's leading AI firms.
Aidan Gomez
The chief executive of Cohere, who co-authored the 2017 Transformer research paper while working at Google.
SoftBank
A Japanese multinational conglomerate with major investments in technology and artificial intelligence infrastructure around the world.
CoreWeave
A U.S.-based cloud infrastructure provider that supplies GPU computing capacity to AI companies and is backed by significant debt financing.
The details
Gomez made the comments from Tokyo, where he is working on expanding Cohere's business. Cohere has focused on raising equity rather than debt, and relies on debt-backed infrastructure providers such as CoreWeave for compute.
Timeline
Last month: Cohere announced its $7 billion raise at a $35.7 billion valuation.
Tuesday, November 10, 2026: SoftBank announced a $3.7 billion writedown tied to its U.S. data-center projects.
Roadmap
Cohere's compute shortage reflects a wider pattern in which AI developers depend on debt-financed infrastructure providers such as CoreWeave to build out capacity. SoftBank's $3.7 billion writedown on U.S. data-center projects shows the financial strain that can follow when those buildouts are repriced.
Shortages of compute capacity can slow the rollout of AI tools that businesses and consumers use, affecting pricing and availability of services built on them. Canadian users and enterprises relying on Cohere's products may see the effects through service limits or product timelines.
The takeaway
Access to compute capacity is becoming a central constraint for AI companies, so funding strategy and infrastructure partnerships matter as much as model development. Readers tracking Canadian technology should watch how Cohere's $7 billion raise and government support shape its ability to scale.
Further reading
For more on the infrastructure behind AI firms like Cohere, see the Data Centers section.
Live Poll
Do you think AI companies are spending too much on data centers right now?










