Cohere founders become billionaires outside the superintelligence race

Felix Laurent 7 Nov 2026 4 min read

Funding

Cohere's founders are Canada's first AI billionaires. Aidan Gomez, Nick Frosst and Ivan Zhang now hold stakes worth an estimated $1 billion each, according to the Bloomberg Billionaires Index. Their wealth comes from Cohere's latest funding round, which valued the Toronto company at about $25.5 billion.

Bloomberg reported the news on November 6. It argues the three owe their status partly to staying out of the race for superintelligence, a race that spooked financial markets this week. Cohere builds AI for businesses and governments and has never promised a machine smarter than humans.

How Cohere's founders became billionaires

Cohere announced the first close of its Series E on October 29. The round targets up to C$7 billion at a valuation of C$35.7 billion, or about $25.5 billion. Cohere calls it the largest private financing round in Canadian startup history.

The investors include:

lead: companies of Schwarz Group, the German retail group that owns Lidl;
new investors: the Government of Canada, Bell Ventures and Bullhound Capital;
returning investors: Nvidia, Inovia Capital and Prosperity7 Ventures.

The Canadian government agreed to invest up to $1.14 billion, Bloomberg reported. Bloomberg valued Cohere at $20.5 billion before the new money, roughly three times its value a year earlier.

Cohere says its annual recurring revenue grew more than 165% in the year to September. Its customer count rose 20 times, to 300, including RBC, TD Bank, Fujitsu and Bell Canada. The company has not disclosed its revenue, and it declined to comment on plans for a stock listing.

The three founders started Cohere in Toronto in 2019. Gomez co-wrote "Attention Is All You Need," the 2017 Google paper that introduced the transformer behind today's chatbots. Frosst worked at Google Brain under Geoffrey Hinton. Their path to wealth differs from most AI fortunes in one key way.

Why Cohere stays out of the superintelligence race

OpenAI and Meta have both named superintelligence as a goal, and Meta even named its AI lab after it. Cohere has taken the opposite line from the start. It sells language models, search and agents to banks, telecoms and governments, and it pitches return on investment over grand visions.

Frosst has been the most direct about it. "It's not a god inside a computer," he said of large language models at a Vector Institute conference in Toronto in February, BetaKit reported. He has also said he does not expect large language models to destroy humanity or eliminate all jobs. That puts him at odds with his former mentor Hinton, who warns about those risks.

Opting out of the race does not mean opting out of big models. The founders' post on the funding round attacks "frontier fatalism" by name. That is the idea that only two countries and a few companies can build leading AI. Over the next 12 months, they plan to start frontier-scale training with allied partners.

The difference is the goal. Cohere wants competitive models that companies and countries can own and run themselves. That message landed in a week when the superintelligence race turned into a political issue.

What spooked markets this week

On November 5, an AI chip stocks selloff erased trillions of dollars in market value. The S&P 500 fell 4.4% and the Nasdaq Composite 5.5%. The drop came a day after Democrats swept the midterm elections, with several candidates promising to curb data center construction.

Superintelligence became part of that debate. After the vote, Democratic Representative Yassamin Ansari called for a pause on advanced AI. She also wants the new Congress to ban superintelligence. A bill to that effect already exists. The Ban Artificial Superintelligence Act, introduced by Representative Greg Casar in September, includes prison terms of up to 20 years for some violations.

The bill would have to pass both chambers and could face a presidential veto. Still, investors started to price in the risk that the most ambitious AI projects face new limits. Bloomberg's point is that Cohere carries little of that risk, since it never bet on superintelligence. Its other advantage is where it is based.

Sovereign AI is Cohere's other edge

Governments that want AI outside US and Chinese control are Cohere's strongest supporters. Canada's investment makes taxpayers part-owners of the company. In return, Cohere promised to keep its headquarters in Canada. A day after the round, it also expanded a commitment it began in 2024. It now plans to invest more than C$4 billion in research and AI infrastructure there through 2031 and add more than 1,500 jobs.

Europe is the second pillar. This year, Cohere agreed to combine with Germany's Aleph Alpha, with headquarters in Toronto and Berlin. Schwarz Group plans to run Cohere's models on STACKIT, its own cloud. In Canada, Cohere signed a C$309 million GPU deal with Bell to run its models on domestic infrastructure.

This strategy depends on government budgets and enterprise contracts. Those move more slowly than consumer chatbot growth, but they are less exposed to the market swings of this week.

Is staying out of the superintelligence race a winning bet

For now, yes. Cohere tripled its valuation in a year, won a government as a shareholder and made its founders billionaires without promising superintelligence. The midterm backlash against AI gives that position extra appeal, because Cohere sells something regulators find easier to accept.

The bet has limits. At about $25.5 billion, Cohere is worth a small fraction of OpenAI or Anthropic, and it still does not disclose its revenue. Frontier-scale training is expensive, and Cohere has not said which partners or how much compute will power it. The founders' fortunes now depend on whether sovereign AI and enterprise customers can carry a company that chose a narrower race.