Ep14. Public Market Volatility, AI Air Pocket, $GOOG Ruling | BG2 w/ Bill Gurley & Brad Gerstner

Episode Summary

Executive Summary: The discussion argues that recent market volatility is a healthy reset after a huge six-quarter rally, not a 1987-style crash. The hosts focus on five risks shaping public markets: possible tax increases after the election, signs of slowing growth/recession, an AI CapEx/revenue mismatch, the Japan carry trade unwind, and Middle East geopolitical risk. They also debate Google’s antitrust ruling, AI engagement metrics, and take-under deals like Character AI.

Main Topics: Public-market volatility as a healthy consolidation (Priority: 5/5): The speakers frame the recent selloff as a normal correction after strong gains since early 2023, emphasizing active risk management and reduced exposure when risk/reward deteriorates. Election and tax risk (Priority: 5/5): They discuss the potential reversion of individual tax cuts in 2025 and argue that higher personal taxes could be a meaningful headwind to consumer spending and markets. Recession/soft-landing debate (Priority: 5/5): The conversation focuses on weakening labor and consumer data, rising delinquencies, and whether the Fed is already behind the curve as the economy slows. AI CapEx versus revenue timing (Priority: 5/5): They debate whether hyperscalers’ massive AI spending will be justified by end-user demand, and whether current enthusiasm resembles earlier infrastructure cycles where revenue lagged investment. Japan carry trade unwind (Priority: 4/5): The hosts explain how Japan’s rate hike triggered forced unwinding of leveraged yen-funded trades, adding global selling pressure and reminding markets of hidden fragility. Google antitrust ruling and search competition (Priority: 4/5): They analyze the court ruling against Google’s Apple search deal, likely remedies, and whether consumer choice or new AI search tools could meaningfully change search market share. Take-under deals and AI industry structure (Priority: 4/5): They discuss unusual transactions involving Character AI and Inflection as a way for startups to access incumbents’ capital and distribution while avoiding standard M&A scrutiny.

Key Arguments: The recent drawdown is a consolidation after extraordinary gains, not evidence of a broken market regime. Risk should be reduced when skew worsens; the speaker moved from roughly nine units of risk to three units as conditions changed. The election matters because automatic reversion of individual tax cuts could remove about $150B of annual consumer support. The economy is slowing, as seen in weak company commentary, softer travel/home-furnishings demand, weak jobs data, and a rising unemployment rate. The Fed will likely cut rates, but cuts may not rescue markets if they only confirm weakening growth. AI spending is still being defended by big tech because firms see capacity constraints and fear underspending more than overspending. The key unresolved AI question is not whether CapEx continues, but whether demand/revenue on the other side arrives fast enough. The Japan carry trade was a major hidden source of leverage; a BOJ rate hike forced rapid deleveraging and global asset sales. Google’s search dominance may persist even after the ruling because consumer behavior and default economics still strongly favor Google. Generative AI apps currently show lower engagement and retention than top consumer apps, suggesting search chatbots alone may not create a billion-user breakout. Take-under deals may reflect startups’ inability to scale independently against hyperscalers’ capital, compute, and distribution advantages. The broader policy implication is that antitrust should stop clear monopoly abuse but not make normal, small-company acquisitions impossible.

Data Points: S&P 500 move: Down 8% - Recent market selloff discussed as part of a broad risk-off move Market value wiped out: $4 trillion - Approximate value erased in the S&P 500 decline QQQ decline: Down 11% - NASDAQ-heavy ETF fell sharply during the volatility Magnitude of rally since Q1 2023: Up 60% - Stocks’ gains from early 2023 to the discussion date QQQ peak gain since early 2023: Nearly 80% - At its July peak before the correction Example stock performance: NVIDIA more than doubled - Used to illustrate how much the market had run up Public-market risk units: 9 units to 3 units - Portfolio exposure was cut as risk/reward deteriorated Illustrative portfolio size: $5 billion - Example used to explain what “units of risk” means Individual tax cut reversion impact: About $150 billion a year - Estimated annual consumer income at risk if tax cuts expire GDP context for tax impact: About 60 basis points of U.S. GDP - Converted from the annual tax amount Election polling/markets: Kamala favored by a slight amount - Used to explain why the market is pricing tax risk differently Soft-landing probability: About 80% in markets vs. 30-35% by Jamie Dimon - Illustrates disagreement about recession odds Positive revenue beat rate: 48% - Among S&P companies reporting; lowest since Q3 2019 Companies reported: 80% - Share of S&P companies already reported at the time Unemployment rate: 4.3% - Used with the Sahm Rule discussion as a recession signal Fed funds rate: 5.5% - Evidence that policy is highly restrictive 10-year Treasury yield: 4.0% - Market-implied policy path already below Fed funds 10-year TIPS yield: 1.8% - Compared to a neutral level around 0.5% to show restrictiveness Expected inflation: 2.9% CPI next print; last print 3.0% - Supports the case for Fed rate cuts Expected rate cuts: 125-150 basis points over 4-5 months - Market pricing for Fed easing Big tech AI CapEx: $220 billion this year - Estimated combined spending by the big four hyperscalers NVIDIA data center revenue expectations: $136 billion this year; $166 billion next year - Used to show continued growth but slower growth rates Japan carry trade size: Up to $20 trillion yen borrowed - Estimated scale of leverage funding global yield-seeking trades Carry trade unwind: About half unwound in first two days - JP Morgan estimate after BOJ rate hike Nikkei drop: 13% one-day decline - Described as its biggest one-day drop since 1987 Google Apple deal: Over $25 billion per year - Default search-engine payments cited in the antitrust ruling Search behavior: Google still gets about 90% of searches - Used to argue consumer choice may preserve Google’s dominance AI app engagement: ~40% WAU/MAU for ChatGPT, Claude, Gemini - Compared with much higher engagement for top consumer apps Top app engagement: Over 80% WAU/MAU - Examples included Spotify, Instagram, WhatsApp ChatGPT month-one retention: About 65% - Best of the Gen AI apps in the cited dataset Best-in-class app retention: Over 90% - WhatsApp, Instagram, Chrome-level retention cited as benchmark Character AI engagement: About 64% WAU/MAU - Used as an outlier among generative AI apps Character AI 12-month retention: 35% - Additional long-term retention figure referenced by the hosts

Pivotal Quotes: "This is not 1987." — Speaker: Framing the recent selloff as a normal correction rather than a historic crash "When uncertainty goes up, the discount rate, the margin of safety you need on your portfolio goes up." — Speaker: Explaining why multiple compression follows geopolitical, economic, and policy uncertainty "It would be the first major phase shift we went through where that's the case." — Speaker: On the idea that AI CapEx and revenues will be perfectly aligned, unlike prior tech cycles

Implications: Investors should expect continued volatility, slower growth, and higher selectivity. The biggest near-term risks are recession, policy shifts, and AI revenue lag, while AI search and take-under deals may reshape competition but not necessarily dethrone incumbents.