Venture Capital: Investing in AI | RAISE Summit 2024 | Paris

Panel Context & Moderators

• Arnaud Barthélémy (Alpha Intelligence Capital) moderated the session, noting his firm was founded in 2018 as a pioneer in the space.
• The discussion focused on venture capital strategies, investment timing, and value creation within the AI ecosystem.

Investment Strategy & Value Chain

• Adrian (Merantix) advocates for investing in all layers but prioritizes the application layer, arguing this is where 80–90% of value will be captured over the next 10–20 years.
  • Cited the internet revolution as precedent: infrastructure builders (IBM, Intel) were replaced 20 years later by application builders (Google, Amazon).
  • Focuses on "new category creation" where business models are impossible without machine learning.
  • Portfolio Example 1: Avara, providing population-wide breast cancer screening to governments in emerging markets (India, Egypt), selling the entire program rather than just algorithms.
  • Portfolio Example 2: Cambrium, co-owning IP while designing novel proteins with industry partners, moving up the value chain.
• Anand (Canonical/Lightspeed) emphasizes "founder-market fit" (ikigai) and leans toward foundational/infrastructure investments where the firm has historical expertise.
  • Identifies "full stack" emergence (e.g., OpenAI, Mistral) combining consumer front-ends with backend APIs.
  • Notes that application layer investments will emerge soon, but current focus remains on strengths in foundational AI.
• Kay (SGH Capital) highlights the necessity of data-driven investing to replace reliance on luck, instinct, and serendipity.
  • AI models are used to monitor millions of startups, acting as thousands of analysts.
  • Human role is shifting to "enhanced" decision-making, focusing on cultural fit, founder rapport, and external factor analysis (e.g., post-COVID market shifts).
  • SGH developed a proprietary model via a joint venture with CrunchDAO, leveraging a global community of 5,000 data scientists (including 600 PhDs) across 91 locations.

Target Sectors & Verticals

• High-Potential Verticals: Heavily regulated industries with slow historical adoption are identified as most ripe for disruption, specifically:
  • Medical, Healthcare, Biotech.
  • Legal.
• Challenges: Significant regulatory headwinds and compliance issues must be overcome to cross the "chasm" of mainstream adoption.
• Observation: Early AI applications often appear as "toys" before rapidly maturing into material utilities.

Global Ecosystem Analysis (US vs. Europe)

• US Strengths:
  • Driven by a culture of high risk tolerance and rapid capital velocity.
  • NASDAQ performance and GDP growth reflect a strong "long stock" mentality and willingness to bet on the future.
  • SGH invests 80% of its capital in the US to maximize exit multiples, despite being Paris-based.
• European Landscape:
  • Talent: Significant research output and technical talent pools (e.g., Mistral).
  • Gaps: Lacks generational entrepreneurial capital; fewer operators who have exited successfully to restart ventures.
  • Culture: Higher risk aversion, partly due to government deployment strategies and regulatory environments.
  • Outlook: Adrian views AI as a global solution to major human problems rather than a tool for a new "Cold War" or competitive advantage between regions.

Valuation & Market Dynamics

• Current State: The market is in a "price discovery mode" with no established benchmarks or historical comps.
• Pricing Methodology: Valuations are currently based on the assumption of future monumental value rather than current revenue; many target companies have zero revenue.
• Hype Segmentation:
  • High Valuation/Hype: Infrastructure and model layers.
  • Lower Valuation: Vertical, industry-specific applications (healthcare, biology, robotics, manufacturing) due to lower current adoption.
• Future Trend: As companies mature, traditional metrics (multiples, revenue) will likely become more applicable.

Technological Convergence & Future Outlook

• AI in Investing: Panelists anticipate a future where autonomous agents (VC agents vs. Dev agents) interact to allocate capital and generate code, potentially accelerating the investment lifecycle.
• Adoption Speed: The transition from novel technology to essential utility is occurring faster than in previous innovation waves (e.g., Waymo driverless cars cited as a "living in the future" example).
• Blockchain Integration: AI development increasingly utilizes blockchain infrastructure for primitives, such as community-driven data verification (CrunchDAO model) and automated capital allocation.