Brad Gerstner: No AI Bubble, Semis Eat the Nasdaq & AI's Take Off Problem

Brad Gerstner argues that every child in America, totaling 70 million under 18, should have access to specific financial accounts, which he views as a significant opportunity for direct philanthropy in the U.S. He believes that increasing capitalism is the solution to rising socialism, highlighting that the current market expansion is primarily driven by earnings.

The market has experienced a 15% rise this year and a 39% increase since January of the previous year, with earnings up by 26%. Gerstner points out that semiconductors contribute to 70% of the NASDAQ's returns and asserts that the current market conditions do not reflect a bubble similar to that of 2000. He estimates that the top three AI labs have a collective run rate of around $100 billion, predicting they need to reach at least $180 billion by year-end to sustain their growth trajectory.

He warns that if the run rate revenue ends the year around $200 billion, it would be insufficient compared to previous years. Gerstner emphasizes the importance of offtake revenue to support the trade in the coming years. This year, the total compute added is approximately 19 gigawatts, with a forecast of 43 gigawatts for the next year, although he considers 25 gigawatts a more realistic target.

Gerstner describes the total addressable market for knowledge work as the largest in history, stating that capturing just 4% of this market, or $1.2 trillion, would cover capital expenditures. He predicts that inference will grow to 1 billion times, with 47 quadrillion tokens expected to be produced this year. Enterprises have reportedly increased their spending on knowledge work by 17 times over the last 18 months, with companies like Uber and Snowflake managing to grow without increasing their headcount.

Despite the potential for AI to expand profit margins significantly, human labor remains the largest cost input for companies. Gerstner notes that regulation could pose risks to AI development. He mentions that Anthropics' revenue is projected to be between $100 billion and $110 billion this year, and there is a considerable risk that rate hikes could significantly impact stock performance.

Monthly revenues for AI labs nearing $8 billion could indicate a market takeoff, with the period from 2023 to 2025 expected to be a super cycle for AI technology. However, Gerstner anticipates that the market landscape in 2026 will be different, as advancements in AI are likely to be already priced in. He describes the current investment position as medium, allowing for adjustments based on revenue and oil prices.