Founder Story: Aidan Gomez of Cohere
In 2017, a paper with a slightly cheeky title — "Attention Is All You Need" — introduced the transformer, the architecture that now underpins nearly every large language model on earth. It had eight authors. Seven were established researchers. The eighth was a roughly twenty-year-old University of Toronto undergraduate interning at Google Brain: Aidan Gomez. Most people would have spent the rest of their career dining out on that credit. Gomez went and built a company instead.
That company, Cohere, is now one of the most valuable AI businesses in the world — but it got there by refusing to play the game everyone else was playing. While OpenAI chased consumers and Anthropic wrapped itself in the language of existential safety, Gomez pointed Cohere squarely at the least glamorous, most durable customer of all: the enterprise. It was a contrarian bet, and it has aged remarkably well.
The Intern Who Was in the Room
Gomez grew up in Brighton, Ontario, a small town well outside the usual Silicon Valley pipeline, and studied computer science and mathematics at the University of Toronto. It was there, through Toronto's deep machine-learning community, that he landed an internship at Google Brain — and found himself, barely out of his teens, contributing to what would become the most cited AI paper of the modern era. He was the youngest author and the only student in the group.
What's telling is how he describes that period. Even the people who built the transformer, he has said, didn't grasp what they'd set loose. "There's no one in the field who was around back then working who could have foreseen where we are in terms of technological capability," he told CNBC, adding that today's models are "doing stuff that I personally thought maybe I'd see at the end of my career, maybe in like 40 years." That mix of proximity and humility — being in the room, yet honest about how little anyone understood — would shape how he built.
From Oxford to a Different Bet
After Google Brain, Gomez began a PhD in computer science at the University of Oxford, working under Yarin Gal and Yee Whye Teh. But the pull of building proved stronger than the pull of academia. He paused the doctorate to start a company, eventually being granted the PhD in 2024, years after the research world had already reorganized itself around the architecture he'd helped create.
In 2019, Gomez co-founded Cohere with Nick Frosst — the first employee of Google Brain's Toronto lab and a direct mentee of Geoffrey Hinton — and Ivan Zhang, a collaborator from FOR.ai, a research group Gomez had started. From the beginning, the founding thesis was different. The others were building models for everyone. Cohere would build models for business — deployable privately, securely, and inside the walls of the world's most regulated industries.
Cohere positioned itself as a third path: not the consumer lab, not the safety lab, but the one that would quietly make enterprises more productive.
Selling AI to the Unsexy — and Winning
The genius of Cohere's positioning is that it targeted exactly the customers who could never use a public chatbot: banks, insurers, hospitals, manufacturers, and governments, all of whom need AI that can run on-premise, in a virtual private cloud, or fully air-gapped. Cohere built its product line around that reality — the Command family of generative models, Embed for semantic search, and Rerank for retrieval precision, all engineered to power the retrieval-augmented generation (RAG) pipelines enterprises actually deploy.
Gomez is refreshingly blunt about where the value comes from. "You sit a knowledge worker down next to one of these models," he told CNBC, "and once they do, you see productivity lifts that are like 40%." His favorite proof points aren't flashy demos but unlikely adopters: "I never thought that an insurance company for natural resource projects would be adopting large language models. But they are." In January 2025, Cohere pushed further into the future with North, a secure, privately deployable agentic workspace that bundles its models and search tools with customizable agents — early customers included RBC, Dell, and Bell.
The Anti-Hype Founder
In an industry drunk on grand narratives about building god-like intelligence, Gomez has been conspicuously allergic to the mythology. He has spoken about disliking "the vibes of the whole AGI" framing — the sense that people were "larping a new religion" — and has been explicit about what actually motivates him: "I want to save the world with AI… make healthcare better," he's said, which "inspires me much more than building God."
This isn't just branding; it's strategy. By refusing to compete on spectacle, Cohere freed itself to compete on the things enterprises care about — security, deployability, cost, and measurable productivity. It also attracted a striking roster of backers who saw the same opportunity. Cohere raised through a $500M Series D at a $5.5 billion valuation in July 2024, then a further $500M round at a $6.8 billion valuation in August 2025, with Nvidia, AMD, Oracle, and Salesforce Ventures all doubling down. The company even drew Canada's sovereign-AI ambitions into its orbit and hired former Meta AI research chief Joelle Pineau as Chief AI Officer.
Industry Impact
Cohere's rise has quietly validated an entire category. When Gomez started, the conventional wisdom was that foundation-model companies would live or die by consumer mindshare. Cohere proved there was a large, defensible business in being the AI vendor that a compliance officer could actually say yes to — and much of the enterprise-AI market now organizes itself around exactly that promise. Its partnerships read like a tour of the global economy: Oracle embedding Cohere in its cloud, Fujitsu co-developing a Japanese model, SAP and McKinsey and Notion building on top.
Just as important is what Cohere represents geographically. Headquartered in Toronto, it became living proof that a frontier-scale AI company could be built outside Silicon Valley, rooted in the same academic community — Hinton, the U of T, the Vector Institute — that produced the deep-learning revolution in the first place. Gomez didn't leave that ecosystem to build; he built from it.
Closing Thoughts
Aidan Gomez's career is a study in the power of being early and staying grounded. He was early to the transformer, early to the realization that enterprises would be AI's most durable customers, and early to the unfashionable view that the point of all this is utility, not deity. Each of those bets looked contrarian at the time. Each looks obvious now.
The deeper lesson for founders is about conviction under social pressure. It would have been easy for a transformer co-author to join the race everyone was cheering — the consumer chatbots, the AGI manifestos, the trillion-parameter arms race. Instead, Gomez asked a quieter question: who will actually pay for this, and what do they truly need? "I'm not interested in building God," he has essentially argued; he's interested in the insurance company that just got 40% more productive. In an industry that loves to talk about the future, Aidan Gomez built a very large business by taking the present seriously.
Your Story Could Be Next
Every founder in this series started with an idea and the will to build — what has changed is the leverage. Frederick gives you a team of AI agents that discover, build, and market your company, running the work across your whole business so you can focus on the decisions only you can make.
If Aidan Gomez's journey resonated, there is no better time to begin your own. Start building your company with Frederick.
References
- https://en.wikipedia.org/wiki/Aidan_Gomez
- https://en.wikipedia.org/wiki/Cohere
- https://arxiv.org/abs/1706.03762
- https://oatml.cs.ox.ac.uk/members/aidan_gomez/
- https://techcrunch.com/2025/08/14/cohere-hits-a-6-8b-valuation-as-investors-amd-nvidia-and-salesforce-double-down/
- https://betakit.com/cohere-series-d-tops-out-at-500-million-raised-at-5-5-billion-valuation/
- https://www.nbcsandiego.com/news/business/money-report/aidan-gomez-helped-invent-the-transformer-at-google-at-cohere-hes-now-taking-ai-to-businesses/3559270/