The AI revolution is just beginning, say three venture capital firms, and we’ve barely scratched the surface of everything we can do with it. Jérôme Wittamer, Ben Prade, and David Buller share their perspectives on agentic AI, applied AI, and the search for real value.
Since the launch of ChatGPT in November 2022, it seems like the market has been flooded with AI tools and applications. Is it too late for startups to get in the game?
Certainly not, according to Luxembourg VC firms. “This is a revolution that will unfold over a period of at least 20 years. At least, and probably much more than that,” says Expon Capital founding partner Jérôme Wittamer. “Being in the first innings may provide some advantages, but not necessarily, because it’s still very immature,” he adds. “Some of the companies that are going to be worth a trillion euros or dollars in 2040 are just being born now.” It’s still the “Wild West”, and plenty of gold remains to be discovered.
“Some of the companies that are going to be worth a trillion euros or dollars in 2040 are just being born now”
– Jérôme Wittamer
“This is just the beginning,” agrees David Buller, managing partner at Keles, a VC firm that focuses on digital health investments. “I don’t think we understand even a fraction of the possibilities of how to apply AI to what we’re doing, whether it’s in business, healthcare, life science, or even our personal lives. In some respects, it’s probably a good time to get in now.” While the hype cycle is not yet over, he notes, it’s now possible to “see through the smoke” and come in with a fresh, calmer mindset.
It’s like the construction of the American rail network in the 19th century, says BullhoundCapital partner Ben Prade. “Everything that the US has done subsequently has been aided by the development of that rail network. The whole economy is built on that rail network; it opened the whole of the west of America.” When it comes to AI, the “rails” have been laid down. Now, it’s time to build on that infrastructure and develop into new markets.
Understand and harness the potential of AI
So what exactly are VCs looking for?
“Real value,” replies Buller. With a focus on health and life sciences, Keles looks for firms that can deliver better healthcare and tangible outcomes for patients. Qureight, for instance, uses AI to analyse lung data, create models, diagnose diseases, and track treatment. Having proprietary data sets around which models can be built can be a differentiating factor, especially in the healthcare space.
But beyond using tech to improve care or speed up drug discovery, we’re “looking at how companies can harness the evolving landscape of AI,” says Buller. “We’re looking at people who have really got the insights into where AI can take their company over the long-term and keep integrating the evolving AI infrastructure and tools that become available.”
“Any company today should be an AI-native company,” says Wittamer. “They should leverage the growing capabilities of models, and they should be built on agentic workflows. In other words, they should be much leaner than previous generations of companies because there’s a lot that can be automated.” AI offers a toolbox that can give its users “superpowers,” allowing them to be much more competitive. The Luxembourg fintech Next Gate Tech, for example, has an AI-native platform that centralises data and automates workflows for the investment fund industry. It’s a company in which Expon Capital has invested, adds Wittamer, and is on the “leading edge”.
But Prade notes that he still sees a “fair amount” of companies that still don’t “grasp the potential of AI.” Are senior executives really pushing through the adoption of AI through all levels of a company? Are people adapting their habits, are they changing the way they go about their everyday business, are they reaping the cost and efficiency benefits of AI? “What we’d like to see more of, especially in existing businesses, is senior executives really committing to using AI and pushing it through the business.”
“What we’d like to see more of, especially in existing businesses, is senior executives really committing to using AI and pushing it through the business”
– Ben Prade
Where opportunities lie
AI isn’t one single thing, adds Prade. He sees three layers: the “foundational” layer (models and data centres); the “user interface” layer (translations, visual media created from prompts, or coding assistants); and the “applied” layer (where companies in more traditional industries leverage AI to save on costs, develop new products, or ease customer interactions). “We’re quite excited about the applied AI layer.” What’s particularly interesting for BullhoundCapital, says Prade, is the use of artificial intelligence to accelerate the development of technologies that didn’t originate in the AI space. The Australian company Q-CTRL, for example, is “using AI to speed up the development of quantum software.”
The use of “agentic AI applications to automate workflows is absolutely gigantic,” says Wittamer, and will be part of the next wave of opportunities. But we can go further. “In Luxembourg, there’s an extraordinary treasure trove of expertise in international finance workflows.” Like a beating heart that pumps blood around the body, money flows through the Luxembourg financial centre and around the world. “There’s knowledge here of how that heart works and today, the heart works thanks to people and what’s in their head.”
“But we would expect, in a place like Luxembourg, that people recognise: ‘Okay. There’s something unique in how we understand these international financial workflows. I’ve been doing this [task] for 20 years. It’s frustrating and time-consuming. But now we have this technology that would allow us to simplify it and make it more cost-effective and faster. So let’s create a company and let’s do it,’” says Wittamer. “And I’m not seeing enough of that.”
Look at the long-term
“I don’t think we’re at peak hype cycle yet,” says Wittamer. It might be in a year, or two years, nobody knows. A lot of brainpower is going into exploring the new “legs” of the AI revolution. “So we may have moments like the ‘ChatGPT moment,’ but it might not feel like a ‘ChatGPT moment’ because our expectations are much bigger. But the reality of the revolution, I think, is in the details, in the mundane things, like understanding a specific workflow and applying the technology so that it provides a flawless outcome. It’s not shiny, it’s not sexy, but that’s going to deliver productivity gains.” While it may not make the short-term headlines, it will have a transformative impact in the medium- to long-term.
“New technology is coming on board all the time, and something that can seem exciting today is outdated tomorrow,” concludes Buller. “It’s going to be a case of who can spot the long-term success stories.”
“New technology is coming on board all the time, and something that can seem exciting today is outdated tomorrow”
– David Buller
This article was published in the Silicon Luxembourg magazine.
