newsfilter
← Interviews
Baron Capital59 min

The Next Wave of AI Investing: Why Active Management Matters

Michael Lippert, Ashim Mehra, Matt Camuso, Mike Lippert

Webinar, Conference Presentation
  • The AI market is transitioning from a capital-light "build-out" phase to a nuanced "Act 2" characterized by agentic AI, longer-form inference, and a shift in focus from training models to agents acting as co-workers by 2026.
  • Revenue growth is accelerating following historical technology adoption curves, with forecasts consistently revised upward, though high growth rates in the infrastructure layer are expected to normalize into a sustainable, profitable environment.
  • Companies that consistently increase R&D budgets are projected to outperform peers by nearly 2x over the next decade, while the duration of competitive advantage is shortening, requiring constant retesting of strategies due to rapid disruption.
  • Innovation will diversify beyond software and semiconductors into sectors including music streaming, drones, autonomous driving, healthcare, and space, with the greatest companies of the next 20 to 25 years expected to be innovators or support entities.
  • The industry is shifting from a capital-light era to a capital-intensive model requiring significant investment in factories, compute, networking, and power, with energy identified as the most critical scarce resource and a potential bottleneck for U.S. data center growth.
  • Infrastructure companies are expected to move toward high profitability and earnings per share growth via stock buybacks even as growth rates slow, while the top 10 companies in every decade change with six or seven new names appearing regularly.
  • Investment strategy will require a research-driven approach to identify winners across a multi-model world of open, frontier, and specialized models, with value potentially captured by a "harness" or model router layer optimizing for latency and cost.
  • Market participants are expected to prioritize the cost to complete tasks accurately over "cost per token," with demand for memory and data storage rising significantly alongside the shift toward agentic inference and context-aware systems.
  • Private sector engagement is expected to intensify to identify disruptive companies that may eventually be invested in upon public listing, with a focus on diversifying portfolios beyond infrastructure to include application sciences and companies generating unique value.
  • Economic benefits are predicted to diffuse across the entire economy rather than remaining confined to physical infrastructure, with the AI trend comparable to historical investments in canals, railroads, or electrical systems as a percentage of global GDP.
  • Specific market risks include the potential tapering of AI capital spend, the U.S. ability to generate sufficient power, and complex compliance issues for entities like CoreWeave that may hinder workload migration.
  • Leadership in the 2030s remains unknown with the top companies of the decade currently undefined, implying that active management is necessary to distinguish signal from noise as the market evolves from simple ownership of infrastructure to sophisticated application of AI.