Signal vs noise & the art of disciplined investing

Mastering the Market: Disciplined Investing in the Age of AI

At Web Summit Lisbon 2025, Mr. Blake Montgomery moderated a panel on disciplined investing. Panelists shared how past investment regrets foster discipline, crucial for navigating the dynamic startup funding landscape and distinguishing signal from noise. General Partner Stefan Klestil of Speedinvest noted poor investments often involve excessive time spent on struggling founders, advising against “one-trick pony” founders lacking adaptability.

Mr. Zach Coelius, Founder of Coelius Capital, highlighted the extreme volatility of startup life, stressing that investors must support companies through both triumphs and tribulations. General Partner Maria Palma of Freestyle Capital added that in early rounds, limited data necessitates strong conviction in the team and market, warning that early data can be a “false friend.”

Mr. Klestil detailed Speedinvest’s conviction-driven European model, targeting 10-15% ownership and leading rounds. Their long-term view means they might miss aggressive deals, but attracts founders seeking committed partners. The advent of AI has significantly altered investment, creating pressure to join “hot” deals, sometimes with smaller stakes as major players enter early.

Mr. Coelius cited Higgsfield’s explosive growth from zero to $200 million in annualized revenue in nine months as an example of AI’s transformative power. Such rapid expansion blurs the line between market hysteria and genuine wealth creation, suggesting broad investment might be warranted. Ms. Palma observed that current high valuations are often backed by real revenue growth, unlike 2021’s speculative surges.

Mr. Klestil asserted that despite AI hype, fundamental business principles endure. Founders still need robust distribution strategies for enterprise sales and must navigate complex regulatory landscapes, especially in FinTech, where licenses remain a critical moat. He identified “physical AI” as a European strength, leveraging its industrial base and machine data.

Ms. Palma emphasized aligning investment decisions with a fund’s model and LP promises. Consistent deviation from a disciplined approach can harm long-term returns. She acknowledged the mental toll on VCs, who face constant self-doubt and the challenge of predicting volatile startup trajectories. LP pressure to invest in popular AI companies adds to this.

Mr. Klestil explained that Speedinvest uses an opportunity fund for later-stage investments in previously missed deals. Robust internal investment committee discussions also help them resist hype. Mr. Coelius’s core philosophy is “buy low, sell high,” with success defined by the exit price, centering conviction on the team. He concluded with his “one-sentence test” for a compelling startup: clear, undeniable value that immediately interests a cold-called customer.