Artificial Intelligence for Alternative Investment: Facing the speed of AI

Back

Market watch

Artificial Intelligence for Alternative Investment: Facing the speed of AI

  • 13 November 2026

  • Infrastructure

Reading time: 6 minutes

    At the third edition of Ardian’s Artificial Intelligence for Alternative Investment (Ai × Ai) conference in Paris on November 3, leading thinkers and business leaders confronted the uncertainty created by AI’s extraordinary pace of change. From the future of work to Europe’s place in the global AI race, here are the key ideas and debates that emerged from the morning’s discussions.

    What exactly are AI systems? What will they do to our economies and jobs? And are institutions built before AI equipped to cope with it? For Pauline Thomson, Head of Data Science and Managing Director at Ardian, these three uncertainties share a single root: time. Technology is moving faster than our understanding and our rules.

    This gap was the theme of Ai x Ai: Artificial Intelligence for Alternative Investment, Ardian's annual AI conference, which held its third edition in Paris.  

    François Jerphagnon, Executive Vice-President of Ardian*, Chairman of the Executive Board of Ardian France, Head of Expansion, and Member of the Executive Committee, opened the conference with a reminder: after the introduction of electricity, factories took decades to benefit from the new technology because most just switched their steam engines for electric motors. To really add value, they needed to rethink the whole factory and its processes. AI raises the same question at a faster pace. But companies do have a choice about how they respond. “Speed is a choice that managers make. It is not something technology decides for us,” he said.

    Speed is a choice that managers make. It is not something technology decides for us.

    François Jerphagnon, Executive Vice-President of Ardian*, Chairman of the Executive Board of Ardian France, Head of Expansion, and Member of the Executive Committee

    An economist and a philosopher discuss uncertainty

    An economist and a philosopher discuss uncertainty 

    The event's keynote was a conversation between Daniel Andler, professor emeritus of philosophy at Sorbonne Université, and Jean Tirole, 2014 Nobel laureate in economics and professor at Toulouse School of Economics, introduced and moderated by Sophie Pedder, Paris Bureau Chief of The Economist.

    No one, including the people who build generative AI, can fully explain why it works, Andler claimed. He called for more clarity on what makes these systems work and what they can do, plus fresh ways to describe their kind of "understanding." Tirole argued that AI labs and countries locked in a race give safety too little weight. The two parted ways on jobs: Tirole warned that AI could displace jobs much faster than previous technologies, while Andler said most jobs will remain, because work means dealing with real situations as well as solving problems. Both called for public oversight beyond industry self-regulation. On Europe, Tirole was blunt: "As long as we don't have sufficient AI, then we won't have a say in negotiations." 

    Should organizations move at the pace of AI?

    Should organizations move at the pace of AI?

    The first panel brought together Marco Di Dio Roccazzella, Global Managing Director at Jakala; Anne-Laure Giret, Head of Google Cloud AI GTM; Fanjuan Shi, Chief Information Officer and Chief Data & AI Officer at Prosol (Grand Frais); and Helen Lee Bouygues, Head of Embedded Value Creation at Ardian Buyout. Their shared answer? It depends.

    Di Dio Roccazzella argued for making decisions on security and compliance slowly, then learning and scaling fast, while redesigning business processes rather than simply adding AI to existing ones. Giret advised companies not to commit to a single model provider, since the best LLM changes from week to week, and to invest instead in the platform around the models, with its tooling, guardrails and cost controls.  

    Shi proposed four tests before chasing speed: value, governance, security and engineering capacity. The exception, he said, is cybersecurity, where attackers now move at machine speed and security teams must do the same. Lee Bouygues said the right pace depends on the use case. She also highlighted how private equity investors can help portfolio companies adopt AI by sharing proven applications and lessons learned, while avoiding common pitfalls. “That's the power of private equity that we bring as partners to our portfolio companies,” she said. 

    What we try to do is identify the pitfalls and share lessons learned across our portfolio companies. That's the power of private equity that we bring as partners.

    Helen Lee Bouygues, Head of Embedded Value Creation, Ardian Buyout

    What is slowing AI down in Europe?

    What is slowing AI down in Europe?

    The second panel gathered Olivier Darmouni, Associate Professor at HEC Paris; Alexandre Défossez, co-founder of the open-science AI lab Kyutai; Khaled Maalej, CEO and co-founder of the French fabless semiconductor company VSORA; and, from Ardian, Pauline Thomson and Daniela Ordonez, Senior Economist.

    Ordonez set out the gap: more than $280 billion of private AI investment in the US in 2025, with Europe's figure 14 times smaller.1 But she warned against writing Europe off: “We are in the very, very early stages of this technological revolution.” The eventual economic winners are far from clear. Défossez was equally determined that Europe should remain in the race: “We're not going to lose it until we give up.” He urged Europe to pick its battles, citing opportunities in on-device AI for robots, cars and medicine. Maalej, meanwhile, identified inference – running AI models rather than training them – as a promising entry point for European semiconductor companies.  

    Asked how Europe could move faster, the panelists called for greater political courage and risk-taking, alongside closer alignment of interests and better education about AI. 

    We are in the very, very early stages of this technological revolution.

    Daniela Ordonez, Senior Economist at Ardian

    Workshops: operating models and software due diligence

    Workshops: operating models and software due diligence 

    Alongside the panels, two workshops moderated by Jonas Corné, Operating Partner at Ardian Infrastructure, took a practical turn. In Workshop A, Alberto Calcagno, CEO of Dedalus, and Nadine Zariffa, Director at Ardian Buyout, asked whether accelerating AI adoption means rethinking the operating model. In Workshop B, Oxana Karpenko Pickeral, CEO of Ennov, Skander Kamoun, Lead Data Scientist at Ardian Infrastructure, and Olivier Roy, Director at Ardian Growth, examined what software in the age of AI means for due diligence.

    Closing the conference, Thomson reflected on the uncertainty surrounding AI and what it means for investors. “Our whole craft is to price uncertainty,” she said – but some of the unknowns surrounding AI cannot yet be priced. For organizations facing the speed of AI, meanwhile, “there is no gap in ambition, there is a gap in tempo.” AI models change every few weeks, while companies operate on slower planning cycles.  

    Yet Thomson suggested that time need not be an enemy. With their long-term investment horizons, private equity funds have the opportunity to develop value creation plans that make sense for the companies they support. “Time is actually one thing that private equity funds have,” she concluded. 

    There is no gap in ambition, there is a gap in tempo.

    Pauline Thomson, Head of Data Science and Managing Director, Ardian

    Watch a replay of the event

     

    *Subject to confirmation at Ardian’s next Annual General Meeting, to be held in May 2027.

     

    1. Stanford University, AI Index Report 2026, p. 183. https://hai.stanford.edu/assets/files/ai_index_report_2026.pdf