BG2 with Bill Gurley & Brad Gerstner | MANG VC Gone Wild, Can You Trust AI Valuations, 2024 Set Up for Public Tech Stocks, & the VC Correction Grinds On | E01

Episode Summary

Executive Summary: Two venture capitalists, Brad Gerstner and Bill Gurley, discuss market distortions from big tech's AI investments (Microsoft, Amazon, Nvidia, Google), the current VC correction and stages of grief, public market valuations, and the importance of free-market capitalism. They analyze how tech giants' investments in LLM startups create low-quality revenue and distort valuations, while advocating for honest conversations about valuations, cost-cutting, and the long-term compounding power of technology.

Main Topics: Market Distortions from Big Tech AI Investments (Priority: 5/5): Analysis of how Microsoft, Amazon, Nvidia, and Google's massive investments in AI startups (like OpenAI, Anthropic) use cloud credits rather than cash, creating low-quality revenue and distorting valuations. This 'sport of kings' makes it hard for true startups to compete and may lead to price wars below actual costs. VC Correction and Stages of Grief (Priority: 5/5): Discussion of the current VC downturn, noting that companies entered with more capital than in 2001/2009, extending the correction timeline. The speakers argue that denial is ending, and 2024 will be the 'year of acceptance' as founders recognize real valuations and restructure. They criticize the lack of honest board conversations and the fixation on last-round valuation. Public Market Valuations and Tech Compounding (Priority: 4/5): Review of 2023's multiple expansion, with tech trading at a 36% premium to its 10-year average. The speakers highlight the unusual success of large-cap tech (increasing returns to scale), the 13% earnings compounding in tech vs. 6% in non-tech, and the need for normalized expectations after 2023's 80% return opportunities. Free-Market Capitalism and Innovation (Priority: 4/5): Inspired by Argentina's President Milei's speech at Davos, the speakers argue that free-market capitalism drove the hockey-stick growth in global GDP since 1800. They emphasize the power of idea exchange and commerce, warning against 'attacking the invisible hand' and highlighting the role of risk-takers in societal progress. Open Source vs. Closed AI Models (Priority: 3/5): Debate on the competitive dynamics between open-source and closed LLM models. The speakers express concern that some large AI players are attacking open source, despite benefiting from it. They note that open-source models are catching up fast, and that data limitations may create ceilings for scaling laws. IPO Prospects and Secondary Liquidity (Priority: 3/5): Expectation that IPO windows will open in the second half of 2024 as companies accept public market valuations. The speakers note a 'voracious appetite' for IPOs on the buy side and criticize the silly notion of waiting for a big company to go first. They also discuss how secondaries can distort cultures and delay necessary restructurings.

Key Arguments: The big MANG (Microsoft, Amazon, Nvidia, Google) investments in AI startups are creating market distortions because they are often made with cloud credits rather than cash, creating low-quality revenue that auditors sign off on but which represents taking from one pocket to put in another. This structure can lead to price wars where AI services are priced below the cost of credits, distorting the market for true competitors who must raise capital at arm's-length valuations. The current VC correction is different from 2001/2009 because companies had more capital entering the downturn and cut costs quickly, but the lack of honest board conversations and fixation on last-round valuations has delayed necessary recaps and shutdowns. Tech companies have experienced increasing returns to scale rather than diminishing returns, enabling large companies to accelerate growth rates at unprecedented scales. Free-market capitalism, through idea exchange and commerce, has been the primary driver of the hockey-stick growth in global GDP, lifting hundreds of millions out of poverty. Limitless scaling laws may eventually hit data ceilings, and open-source models are closing the gap with leading closed models, making attacks on open source particularly concerning.

Data Points: VC investing by MANG in 2023: $25 billion - Up from nearly zero six-seven years ago, concentrated in a few AI companies. Tech stock price premium over 10-year average: 36% - For Qs (tech index) as of early 2024. Non-tech stock price premium over 10-year average: 12% - Stripping out tech from S&P 500. Tech earnings compound annual growth over 10 years: 13% - Stock prices compounded at 17% over same period. Non-tech earnings compound annual growth over 10 years: 6% - Stock prices compounded at 8% over same period. Technology's share of global GDP: 15% - Up from 5% over the last 15 years. Quarterly shutdowns in 2023: ~180 per quarter - Run rate ~800 per year, happening quietly. Years to double GDP in different eras: 3500 years (pre-1800), 87 years (1820-1900), 60 years (1900-1950), 20-30 years (by 2000) - From Milei's speech data, showing acceleration. People brought out of poverty by Deng Xiaoping: 500 million - ChatGPT's answer cited as evidence of capitalism's impact.

Pivotal Quotes: "The difficult thing as an investor... set aside the fundamental decision, which is: can these companies actually generate a lot of durable and ongoing revenue from this if we have open source providers who are going to collapse the price of the market down to zero?" — Brad Gerstner: Explaining why Altimeter hasn't invested in OpenAI or Anthropic despite the hype. "I think the most differentiated element of this correction versus 09 versus 01 is the amount of capital that the companies went into the correction with, the speed at which they lowered cost afterwards. So there wasn't denial. Everyone got along pretty quickly." — Bill Gurley: Comparing the current VC downturn to previous ones. "Founder-friendly is doing the right thing and being truthful with founders... saying up front what's too often discussed outside the boardroom and behind their backs, even if it means making hard decisions like layoffs, selling, shutting down, down rounds, etc." — Brad Gerstner (quoting Jam and Ball tweet): Criticizing the lack of honest conversations in VC boards.

Implications: Investors should be skeptical of AI startup valuations inflated by cloud credit deals and low-quality revenue. The VC correction will likely continue through 2024, forcing more restructurings and shutdowns. Long-term tech compounding remains attractive, but entry multiples matter. Free-market capitalism's role in innovation should not be taken for granted.