In 2020, Cathay Innovation backed a company that watched people work. Skan used computer vision to follow tasks across business applications, exposing the distance between a company’s official process and what employees actually did. Six years later, Cathay co-led Skan’s $63 million Series C. The observation tool had become a foundation for AI agents. Before a machine could do the job, somebody had to understand the job.
- The bet: AI built around specific industries and their working habits.
- The mechanism: equity capital plus access to corporate expertise and potential partners.
- The scale: a $1 billion third flagship fund, closed in May 2025.
That small gap between a process chart and a working day is a useful place to start with Cathay. Venture capital is good at finding promising technology. Getting that technology into a bank, a hospital or a manufacturing business requires another kind of knowledge. Cathay has organized its firm around the people who possess it.
A physicist meets a border crosser
Denis Barrier began in Orange’s laboratories, studying semiconductor quantum dots. He holds a doctorate in solid-state physics. His subsequent career took him through corporate venture investing, including Orange and the Orange/Publicis fund at Iris Capital. He had experience on both sides of the awkward journey from a scientific result to a useful business.
Mingpo Cai brought a different education. Arriving in France from China at 20, he learned French, built businesses and developed commercial relationships across countries. One early venture, Stonest, worked in the gravestone industry. He also helped establish a Shanghai representative office for Groupe SEB, the appliance manufacturer. International business, in this telling, began with learning whom to trust.
The pair launched Cathay Innovation in Paris in 2015, within the Cathay Capital family. Their shared proposition was practical: founders could benefit from access to established companies, while those companies needed help finding technology outside their own walls. The biographies explain the emphasis. This was a firm built by people accustomed to translating between different working worlds.
Three parties, one difficult introduction
Cathay’s business has three constituencies. Institutions and corporations supply capital to its funds as limited partners. Founders receive investment and support. Corporate participants also contribute industry knowledge and may become commercial partners. Keeping those roles straight matters: a corporation investing in a fund has not automatically agreed to buy every startup’s product.
The corporate names make the strategy tangible. Fund III’s announced backers include Sanofi, TotalEnergies, Valeo, BNP Paribas Cardif, Groupe SEB and Groupe ADP. These businesses understand drug development, energy systems, automotive supply chains, insurance, appliances and airports. For a founder entering one of those markets, relevant expertise can be more useful than another general conversation about disruption.
An introduction creates an opportunity. A commercial agreement still has to be earned.
For founders, the offering is equity capital, business development, strategic introductions and help entering international markets. For corporate partners, it is a route into emerging technology. Cathay occupies ground between a conventional venture investor and a corporate innovation network. Other venture firms offer introductions; Cathay makes those relationships central to its proposition.
What the billion actually buys
The financial progression is substantial: a €287 million first flagship fund, closed in 2017; €650 million for Fund II in 2020; and $1 billion for Fund III in May 2025. The currencies deserve to remain attached to the numbers. So does the word “fund”: these are investor commitments to investment vehicles, rather than Cathay’s annual revenue.
At Fund III’s final close, Cathay reported €235 million already deployed across 14 investments, half based in Europe. Its stated emphasis was vertical AI: technology tailored to healthcare, financial services, consumer businesses, energy and mobility. Barrier called AI a “general-purpose technology.” The investing question was where that general capability could become a specific business.
For a startup, accepting venture investment means negotiating ownership and governance along with the cheque. For Cathay’s investors, the economic ambition is appreciation in portfolio companies and eventual exits. Its portfolio includes Chime, Pinduoduo and Glovo. Those names establish experience across markets; they do not, by themselves, reveal what any individual fund returned.
The molecule test
Aqemia makes the industrial argument concrete. The Cathay portfolio company combines generative AI with physics-based calculations to design drug candidates. In December 2023, it announced a Sanofi collaboration carrying up to $140 million in upfront and milestone payments across programs. Aqemia would design molecules; Sanofi would handle laboratory research, development and commercialization.
That figure was a conditional ceiling, not a cheque delivered on announcement day. And the collaboration already had a history before Cathay’s December 2024 investment. It illustrates the kind of relationship Cathay values; it does not establish that Cathay created the deal. The division of work is the interesting detail: computational invention needs an organization capable of testing and developing it.
Borrow the question before the cheque
Skan supplies the corresponding lesson in enterprise software. In Cathay’s August 2026 founder interview, its leaders described customers moving from diagnosing workflows to asking for deployed agents. The limitation they identified was operational context: manuals and system logs did not fully explain how employees handled exceptions. Better models alone could not fill that gap.
Who owns the problem - and what would make them change how they work?
The question is worth borrowing before a fundraising meeting. Identify the workflow, the budget owner and the evidence needed for adoption. Cathay’s August 2026 essay on enterprise AI economics makes a related argument: measure the cost of completing useful work. A cheap model that needs repeated correction can become an expensive system.
The model depends on willing participants. An introduction helps when an industry partner has a relevant problem, authority to act and patience for implementation. It helps less when the product cannot survive procurement, regulatory requirements or the customer’s economics. Cathay’s wager is that investors can improve those conditions. Skan’s trajectory shows why knowing the working day belongs in the investment discussion.