Alpha and AI Investing

Key Concepts

• AI as a Tech Cycle: AI is viewed as the next major technological revolution, comparable to the PC, internet, mobile, and cloud cycles.
• Intelligence on Demand: AI is characterized as a fundamental new technology providing "intelligence on demand," akin to electricity or the internet.
• Incumbents' Edge: Large tech giants (hyperscalers) currently have an advantage in AI development due to capital, engineering, and data intensity.
• "Mag 7" / "Mag 10": Refers to the largest publicly traded technology companies, with the number indicating a changing index of trillion-dollar market cap companies.
• Private vs. Public Markets: Discussion on where investment opportunities lie, with a focus on the interplay and differing skillsets required for each.
• Broken IPO Market: The current IPO process is considered by some to be fundamentally flawed, leading to companies staying private longer.
• Tokenization: The potential for tokenizing private assets to create liquidity and effectively make them public.
• Global AI Landscape: Comparison of innovation and investment potential in the US, China, and Europe.
• Cost of Compute: The decreasing cost of AI computation (measured by tokens) and its implications for application development and ROI.
• Circular Economy in AI: The interconnectedness of AI companies and semiconductor manufacturers, and the health of this ecosystem.

AI: A Super Cycle, Not a Bubble

The discussion begins by addressing the question of whether Artificial Intelligence (AI) is a bubble. Bill argues that it is not a bubble, but rather a real tech cycle, comparable to previous major technological shifts like the PC in the 80s, the internet in the 90s, mobile in the 2000s, and cloud in the 2010s. He anticipates significant long-term productivity gains. However, he acknowledges moments of overvaluation and misallocation of capital are likely.

Philipe adds a layer of concern, suggesting that while AI is a real phenomenon, the long-term winners and losers are not yet clear. He raises the possibility of significant job destruction as a consequence of AI's success, even as the economy grows rapidly, viewing this as a 10-15 year concern.

Both speakers agree that this AI wave feels more dramatic and fundamental than previous cycles, potentially reshaping the world significantly in 5-15 years. Bill describes AI as "intelligence on demand," a foundational technology that will create huge markets, similar to railroads or electricity.

Incumbents' Advantage in the AI Race

A key point of differentiation from the internet cycle is that incumbents (big tech giants) currently hold an edge in AI. This is attributed to the capital-intensive, engineering-intensive, and data-intensive nature of AI development. Companies like OpenAI and Anthropic are mentioned as potential new entrants, but the established players have significant advantages.

The Shifting Landscape of Dominant Companies

The conversation touches upon the longevity of current tech giants, often referred to as the "Mag 7" (or "Mag 10" given the number of companies exceeding a trillion-dollar market cap). Philipe, with a European perspective, is skeptical of corporate immortality, drawing an analogy to personal mortality. He notes that acronyms like "Fang" and "Mag 7" often signify a mature phase, and the index of dominant companies will likely change over time, potentially including international players.

Investment Strategies and Opportunities

Following Entrepreneurs and New Applications

Bill emphasizes that their investment strategy involves following entrepreneurs and is excited about a new generation of application companies that will harness "intelligence on demand" to build unimagined solutions. He uses the analogy of railroads, where the full potential of on-demand transportation was not initially foreseen.

Transforming Industries: Healthcare and Education

Philipe provides a compelling example of robotic surgery by Intuitive Surgical. He envisions AI enabling robots to learn from every surgery, potentially making complex procedures accessible globally without requiring a human surgeon to operate the controls. This could democratize access to high-quality healthcare, especially in underserved regions.

Bill agrees that healthcare and education, industries less impacted by previous tech cycles, are poised for significant transformation by AI. He foresees AI enhancing medical intelligence for doctors and improving access to quality education and teacher talent.

The AI High School Experiment

Philipe shares an anecdote about visiting an "AI high school" where students learn independently with AI programs tailored to their strengths and weaknesses, with a mentor providing motivation rather than traditional teaching. While acknowledging the importance of social interaction, he sees this as an example of evolving educational models.

Private vs. Public Market Opportunities

The discussion delves into investment opportunities in private versus public markets. General Atlantic focuses primarily on private markets, while CO2 invests in both. Bill stresses the importance of understanding the public market, particularly the actions of large public tech companies, to make informed private market decisions.

Philipe highlights the different skillsets required for public and private investing. Public market investing requires not only belief in the future but also an assessment of whether that future is already priced in. Private market investing demands more patience, active involvement with founders, and a longer-term horizon (5-7 years), which aligns well with the extended timeline of the current tech cycle.

The Broken IPO Market and Tokenization

Philipe expresses a strong belief that the IPO market is "totally broken," citing the significant decline in IPOs compared to previous decades. He argues this is unfair to retail investors who have easier access to public markets post-IPO. He suggests that tokenization of private assets will be a disruptive force, creating liquidity and effectively making these assets public, potentially leading to a future where all assets are tradable.

Bill, while acknowledging an "exit recession" in private markets, is more optimistic about the IPO market's recovery. He points to "green shoots" observed in the current year and anticipates a resurgence in 2026, driven by demand for high-quality companies from public investors.

Global AI Landscape: US, China, and Europe

US Dominance in AI

Philipe asserts that the US is the leading country for AI, possessing the necessary engineers and innovative companies. He contrasts this with Europe, which he describes as "stuck in reverse" due to bureaucracy and competing investment needs in energy and defense.

China as a Strong Innovation Economy

Both Bill and Philipe view China as an incredible market with strong engineers and a large captive market. Bill notes that China has transitioned from being considered "uninvestable" a year ago to being investable again, with significant gains in its equity and bond markets and a leading position in IPOs. He believes China is a strong contender in innovation, alongside the US.

Europe's Challenges

Philipe's assessment of Europe is critical, highlighting its bureaucratic nature and inability to make decisive actions, likening it to "the United Nations of Europe" rather than a unified entity. He notes that despite strong market performance this year, he remains cautious about its long-term innovation prospects.

The Two-Horse Race: US and China

Bill frames the global innovation landscape as a "two-horse race" between the US and China. He believes both countries have large home markets, significant engineering talent, and leading AI competitors like Alibaba, Tencent, and ByteDance. He also suggests China may be ahead in robotics and industrial automation.

Cost of Compute and Investment Ecosystem

Decreasing Compute Costs and Infinite Elasticity

The discussion addresses the cost of compute, measured by tokens. Philipe uses the analogy of gasoline for an engine, stating that even as the price of tokens decreases dramatically, the "P * Q" (price times quantity) can go to infinity due to the elasticity of what can be done with cheaper compute. This will unlock numerous applications across various sectors, including cars, humanoids, and machines.

Bill shares that his firm has applied AI aggressively to its portfolio of 200 companies, even at current costs, and has seen high paybacks in areas like customer care, coding, and digital marketing. He anticipates further solutions and applications as costs inevitably decline.

Health of the AI Investment Ecosystem

Regarding concerns about the "circular economy" between AI and semiconductor companies, Bill expresses watchfulness but not outright worry. He contrasts the current situation with the dot-com bubble of 2000, where capital was fueled by IPOs and dubious business models. Today, he argues, the capital is largely driven by large public companies with strong free cash flow, real boards, and return on capital requirements. He believes the system is healthy, and investments made by these companies reflect a belief in significant future opportunities, with earnings already following through. He highlights the remarkable earnings growth of the "Mag 7" as evidence of this.