Aidan Gomez
The Architect of Enterprise AI and the Case for a More Independent Future
By Michelle Clark
Aidan Gomez belongs to a generation of artificial intelligence pioneers whose influence began long before AI became a household phrase. As co founder and chief executive of Cohere, he has helped shape a company focused on bringing advanced artificial intelligence into the practical world of business, government and critical infrastructure. His story is rooted not only in entrepreneurship, but in fundamental AI research, giving his approach to the industry a distinctive combination of scientific curiosity, commercial ambition and concern about who ultimately controls the technology.
Before becoming a technology chief executive, Gomez was a researcher. His academic journey included the University of Toronto and Oxford, while his professional career brought him into the orbit of some of the most influential researchers in modern artificial intelligence. At Google Brain, he worked alongside figures including Geoffrey Hinton and Jeff Dean. Most significantly, Gomez was one of the co authors of the influential 2017 research paper Attention Is All You Need, which introduced the Transformer architecture that became foundational to the development of modern large language models.
That background matters when considering the company he subsequently helped build. Cohere was founded with the ambition of developing powerful language models for businesses rather than focusing primarily on consumer chatbots. Under Gomez’s leadership, the company has concentrated on enterprise applications where privacy, security, customisation and control are central considerations. Its technology has increasingly been positioned for organisations operating in highly regulated and sensitive environments, including financial institutions, telecommunications companies, governments and other critical sectors.
Gomez’s philosophy has increasingly become associated with the idea that artificial intelligence should not simply be rented from a small number of technology companies. He has argued that businesses and governments need meaningful control over the systems on which important parts of their operations depend. This has become particularly important as AI has moved from an experimental technology into infrastructure capable of influencing finance, healthcare, energy, communications and national security.
The argument has also placed Gomez directly inside the growing debate over AI safety. As some of the world’s most prominent AI leaders warn about increasingly powerful systems and potentially severe long term risks, Gomez has taken a position that accepts the seriousness of AI risks while questioning who should be responsible for defining the rules designed to manage them. He has argued that the risks associated with AI are real, but has rejected the idea that a small group of Silicon Valley companies should effectively determine how the world keeps the technology safe.
For Gomez, the central issue is therefore not whether AI needs safeguards. He has made clear that it does. The question is who writes those safeguards and what evidence is used to establish them. He has argued for an evidence based framework in which governments, researchers and other stakeholders participate in defining specific risks, the circumstances in which they emerge and the appropriate responses. His position challenges the idea that companies developing frontier AI should simultaneously become the creators, interpreters and enforcers of the rules governing their own technology.
This perspective has become particularly relevant as the industry debates whether powerful AI companies should be allowed to cooperate more closely in the name of safety. Gomez has questioned whether such arrangements could unintentionally strengthen the market position of a small group of dominant companies. His concern is that safety regulation, if designed without sufficient competition and independent oversight, could have consequences extending beyond safety itself by determining which organisations are able to develop and deploy advanced AI.
His argument for broader participation is closely connected to another theme that has become central to his leadership: sovereign AI.
Gomez has warned that countries relying heavily on foreign AI providers could find themselves exposed to decisions made elsewhere. In his view, control over AI increasingly resembles control over other forms of strategic infrastructure. If access can be restricted, withdrawn or altered by an external provider, governments and institutions may not possess the independence required to protect their most sensitive systems.
That concern became particularly visible as governments confronted the strategic implications of dependence on foreign AI providers. Gomez has argued that simply renting AI capabilities from another country or company could create structural risks because access, data and operational control ultimately remain dependent on a third party. His preferred model is one in which organisations can deploy AI within their own environments, including infrastructure where sensitive information can remain under their direct control.
The philosophy is reflected in Cohere’s growing focus on sovereign and secure AI. The company has developed deployment options that allow organisations to operate its models within their own infrastructure, including environments where data does not need to be transferred to Cohere’s own servers. This approach is particularly relevant to governments, financial institutions and other organisations for which data sovereignty and operational independence are major concerns.
Gomez’s vision is also becoming increasingly international. Cohere has expanded its presence in Europe, including a major expansion of its London operation, as demand grows for enterprise AI that can operate within local regulatory and security requirements. The company has positioned Europe as an important market for sovereign AI, while Gomez has highlighted the importance of combining technological capability with regulatory frameworks and local expertise.
His approach does not place technological openness and safety in simple opposition. Instead, Gomez argues for a system in which safety is measurable, evidence based and subject to wider participation. His broader philosophy reflects an understanding that responsible AI cannot be reduced to a single technical solution. It requires researchers, governments, businesses and institutions to participate in determining how powerful systems are developed and deployed.
The result is a distinctive profile in the global AI landscape. Gomez is both a researcher who contributed to one of the foundational developments behind contemporary AI and an entrepreneur now responsible for translating that technology into practical systems used by major organisations. His career illustrates how quickly the boundaries between academic research, business strategy, national policy and technological governance have disappeared.
Aidan Gomez’s influence ultimately rests on a question that is becoming increasingly difficult for governments and businesses to avoid. As artificial intelligence becomes embedded in the systems that keep modern economies functioning, control over AI may become as important as access to it. Gomez has built his career around developing the technology, but his increasingly prominent role in the safety and sovereignty debate suggests that his larger concern is what happens after the technology has been built. For him, the future of AI cannot be separated from questions of accountability, competition, security and independence. The challenge is no longer simply creating intelligent machines. It is deciding how society can use them without surrendering control over the infrastructure, information and institutions that depend upon them.
