Cohere CEO Calls for Democratic AI Champions

PARIS – Cohere CEO Aidan Gomez, one of the co-authors of the landmark 2017 Transformer paper that underpins modern generative AI, is pushing a vision of artificial intelligence centered on national sovereignty, critical infrastructure and enterprise adoption rather than a race to build the largest models.

Speaking at the VivaTech conference, Gomez said democratic nations must work together to create viable alternatives to both U.S. and Chinese AI dominance, arguing that concentrated investment in a handful of AI champions is essential if Western democracies want to remain competitive.

His comments came after attending a meeting of G7 leaders and AI executives, where they discussed accelerating AI deployment while managing risks such as cybersecurity threats and child safety.

“It was super special,” Gomez said of the gathering. “There was a very clear sense that they have consensus they need to collaborate much more.”

According to Gomez, government leaders and technology executives shared a common belief that democratic nations must move quickly to capture AI’s economic benefits while coordinating on standards and safeguards.

“The differences were sort of set aside,” he said. “We agreed we need to lead. Democracies need to lead on this technology.”

The remarks reflect a growing geopolitical divide in AI as governments increasingly view advanced models as strategic assets akin to energy, telecommunications and defense infrastructure.

Gomez argued that while the United States currently occupies the leading position in AI, democratic nations need to ensure that the second position is not held by an autocratic rival.

“I think the U.S. is leading in the AI race, and that’s very true,” he said at the Parisian tech conference, which sponsored The AI Innovator‘s trip. “What I said at the meeting was that we need to ensure that a democracy occupies the number two position.”

Building democratic alliances

Central to Cohere’s strategy is a series of international partnerships designed to pool resources among allied countries.

The company recently deepened ties with German AI startup Aleph Alpha, creating what Gomez described as a Canadian-German alliance focused on sovereign AI infrastructure.

Rather than spreading investments across numerous startups, Gomez said governments should concentrate resources behind a small number of promising companies capable of competing globally.

“AI requires concentrated resources,” he said. “We need to direct them behind champions.”

The partnership is aimed at highly regulated sectors and critical infrastructure, including energy grids, telecommunications networks, healthcare systems, financial services and water treatment facilities.

“These systems go down, the economy stops,” Gomez said.

He argued that the growing sophistication of cyberattacks makes sovereign AI capabilities increasingly important because AI is becoming deeply embedded in the software that powers national infrastructure.

The German partnership is only the beginning, Gomez said. Cohere is pursuing similar arrangements elsewhere, including Spain, where he recently signed a memorandum of understanding with Spanish technology company Indra.

He also signaled a deeper commitment to France, where Cohere maintains a significant research presence and plans to expand operations.

“We want to contribute to European sovereignty,” Gomez said. “The talent here is essentially second to none.”

An enterprise-first strategy

Unlike consumer-focused AI companies, Cohere has concentrated on enterprise and government customers since its founding.

That focus has helped differentiate the company from larger rivals with substantially greater computing budgets.

“The strategy is not try to spend the most money,” Gomez said.

Instead, Cohere designs models that can operate efficiently in environments where computing resources are constrained and security requirements are stringent.

According to Gomez, many organizations in critical industries cannot realistically deploy massive frontier models because of cost, infrastructure limitations and data sovereignty concerns.

As a result, Cohere has prioritized efficiency over sheer scale. “We can operate 10, 20, 30 times more efficiently,” he said.

The approach appears increasingly aligned with enterprise demand. Early corporate AI pilots often struggled to demonstrate measurable returns, but Gomez said the market has matured significantly over the past 18 months.

“There was a question a year ago, or 18 months ago” about AI’s ROI, he said. “At this stage, the maturity of market and the maturity of organizations like Cohere to deliver for their needs has increased so much.”

He said many proof-of-concept projects that began during the initial generative AI boom are now moving into production deployments, helping drive rising demand for computing infrastructure across the industry.

The labor question

As AI adoption accelerates, concerns about job displacement have become a growing source of public anxiety.

Gomez acknowledged those concerns but said Cohere’s customers are primarily using AI to drive growth rather than reduce headcount.

“To the best of my knowledge, we don’t see any sort of dismissals happening as a product of AI,” he said.

Instead, he said organizations are using AI to improve employee productivity and expand operations.

Still, Gomez cautioned that governments should prepare for potential disruptions if more advanced AI systems eventually affect labor markets.

He called for policymakers to study options ranging from retraining programs to more ambitious proposals such as universal basic income or job guarantees.

“I do think these need to be well studied and understood in advance of a potential crisis,” he said.

Responding to growing public skepticism

Gomez also addressed mounting public backlash against AI, which has intensified amid concerns over energy consumption, job security and the concentration of power among technology companies.

Rather than dismissing critics, he said many concerns are legitimate.

“I think it’s understandable,” Gomez said. “It feels like a technology that is completely transforming the way that our economy and our society is going to operate.”

He pointed specifically to instances where large data centers have strained local power grids and contributed to rising energy costs.

“There has been examples of places where energy prices have gone up because a data center was built there,” he said.

Developers, he argued, need to be more strategic about locating infrastructure in regions with excess power capacity and clean energy sources.

At the same time, Gomez maintained that AI’s long-term economic benefits could outweigh the risks.

He pointed to sluggish productivity growth across many advanced economies over the past decade and argued that AI may help reverse that trend by boosting efficiency and expanding economic output.

“The promise of AI is restoring growth,” he said.

Defending the Transformer era

Gomez’s comments came a day after a VivaTech appearance by Turing award winner Yann LeCun, who has repeatedly argued that today’s large language models are insufficient for achieving human-level intelligence.

Asked whether he remains optimistic about the future of Transformer-based systems, Gomez offered a nuanced answer.

As a scientist, he said, he hopes the industry eventually discovers architectures that surpass the Transformer, which has dominated AI development since the publication of the seminal 2017 paper, “Attention Is All You Need.”

“It’s absurd that we’re still using that technology,” Gomez said, noting that the Transformer will soon celebrate its 10th anniversary.

At the same time, he challenged critics to demonstrate viable alternatives. “I hope something new comes,” he said.