Interview, Fireside Chat
The AI Boom Will Create Enormous Roadkill: Who Wins & Loses? | David Frankel
- The speaker anticipates a high probability of another dot-com crash, predicting that 95% of entrepreneurs in the current AI wave will fail to survive long-term, while only the top two or 300 companies may achieve trillion-dollar valuations acting as the "new series A."
- Fundraising dynamics are expected to shift significantly, with $50 to $100 million funds likely becoming the worst performers due to size mismatches, and a potential closure of the IPO market after the remainder of this year, leaving many Series A companies orphaned.
- Sovereign wealth funds and public investment corporations are projected to become primary limited partners, shifting investor focus from fund multiples to internal rate of return (IRR), while large endowments may become negligible relative to the scale of sovereign-backed returns.
- A major disruption in the AI sector is predicted within 10 years driven by photonic computing, which will serve as the "NVIDIA disruptor," alongside a competitive threat from Chinese open-source models and a potential rise of China as a superpower in AI and defense technology.
- In healthcare, the speaker expects AI to deliver solutions for chronic conditions and cancer treatments that are currently viewed as prehistoric, though progress may not be fast enough, while autonomous vehicles are predicted to become ubiquitous enough that children will not need to drive within five to 10 years.
- SaaS companies face a bifurcation where "less embedded" tools face cannibalization and growth slowing, while "more embedded" mission-critical solutions within the last 5% may endure longer, with specific mentions of companies like Airtable and Vivaldi facing challenges.
- The labor market is expected to see massive productivity gains and a shift toward younger founders who are mentally plastic to AI tools, though older workers may struggle to retrain, with AI serving as a time-expansion tool for high-stakes services like litigation and insurance rather than causing mass unemployment.
- Government and defense spending is projected to increase dramatically, with potential expenditures of $100 to $200 million on single model results for critical queries, driven by the strategic necessity to compete against China's access to advanced AI capabilities.
- Specific investment risks include the potential for dilution of fund multiples if strategies rely on raising more capital, the possibility that founders viewing early investors as an "insurance policy" is an over-extrapolation, and the risk that junior venture principals moving to start their own funds will leave startups overlooked.
- The capital intensity required to build photonics companies is expected to be dramatically higher than prior technologies, potentially causing a deficiency in US R&D spending, while 95% of entrepreneurs may fail to benefit from mega-platforms taking call options or failing to reach specific revenue milestones.