Ep11. Coatue Conference Recap, Tesla & ISS Controversy | BG2 with Bill Gurley & Brad Gerstner

Episode Summary

Executive Summary: At CO2’s East Meets West conference, the hosts argued that AI is a durable supercycle, not a bubble, with public and private markets both being reshaped by huge CapEx, rapid product adoption, and shifting labor dynamics. They also debated IPO thresholds, the health of software investing, the risks of overfunding unicorns, and whether public markets, not private ones, should serve as the primary venue for major AI companies.

Main Topics: AI as a supercycle vs. bubble (Priority: 5/5): The discussion centers on whether AI is in a speculative bubble or an early, durable phase shift. Philippe’s slides suggested earnings growth at NVIDIA is supporting valuations and that AI is bigger than the market currently believes. Public market performance and concentration (Priority: 5/5): The speakers noted that a few AI-related names, especially NVIDIA, are driving a disproportionate share of S&P 500 gains, while equal-weighted indices lag, reinforcing concerns about market concentration and the sustainability of returns. CapEx intensity and ROI requirements (Priority: 5/5): A major thread was how much investment in AI infrastructure is being deployed and what economic returns are needed to justify it. The debate contrasted labor savings and GDP growth arguments with concerns that companies are competing away each other’s gains. Software is not dead, but AI is reshaping it (Priority: 4/5): They argued software remains attractive, especially at lower valuation multiples, but different layers of the software stack face different levels of AI disruption. AI may replace workflows and query layers more than repositories. Venture market normalization and the unicorn overhang (Priority: 5/5): The venture market is normalizing after the 2021 funding surge. Many private unicorns have slowed growth, are not raising capital, and may never reach public markets, while AI startups are attracting outsized funding at much higher valuations. IPO access and the need for public markets (Priority: 5/5): The hosts challenged the idea that companies must reach $10 billion market cap to go public, arguing that more companies—including OpenAI and other AI leaders—should consider public listing earlier to improve access, discipline, and liquidity. Governance, ISS/Glass Lewis, and Delaware legal risk (Priority: 4/5): The Tesla compensation vote sparked a broader critique of proxy advisors and Delaware’s legal environment. The speakers argued that ISS and Glass Lewis have strayed from shareholder alignment and that Delaware’s legal predictability is being undermined by activist litigation.

Key Arguments: AI investment is likely a megawave: participants across VCs, big tech, and CIOs are all excited and funding the space heavily. NVIDIA’s valuation looks less like a 1999-style bubble because earnings growth has expanded enough to keep its multiple relatively stable despite a massive stock move. The macro backdrop is no longer the main driver; if inflation remains tame, AI’s impact can dominate returns and investment decisions. The economic justification for AI CapEx is better framed through GDP growth and broad productivity gains than through narrow company-by-company margin defense. Software stocks may be the next beneficiaries of AI because they are trading at discounted forward revenue multiples, though some layers face existential pressure from LLMs. The venture market is normalizing from 2021 excess; many unicorns are overvalued, under-growth, and effectively frozen out of exits. Being private is not inherently better than being public; public markets provide valuation discipline, liquidity, and pressure to improve businesses faster. The idea that a company must be $10 billion to IPO would shrink the set of viable public companies and make venture investing far riskier and more capital-intensive. OpenAI-like businesses could merit public-market access if they have enough revenue scale, growth, trust, and a path to break-even. ISS and Glass Lewis are criticized for imposing governance preferences that conflict with actual shareholder interests, especially in performance-based compensation cases like Tesla. Delaware’s legal framework is viewed as less predictable and potentially hostile to high-performing companies, which could push more firms to reincorporate elsewhere.

Data Points: Technology share of global GDP: 5% to 15% - Over the last 20 years, the hosts said technology’s share of global GDP rose substantially. NVIDIA share price increase: Up nearly 10x - Used as evidence that market enthusiasm has been matched by earnings growth, not just multiple expansion. NVIDIA contribution to S&P 500 returns: 5% of S&P 500 returns YTD - Illustrated how concentrated 2024 market gains have been. Other Mag 6 contribution to S&P 500 returns: 5% of S&P 500 returns YTD - Showed that a small group of mega-cap tech companies is driving most of the index’s performance. Remaining S&P 500 companies contribution: 4% of 14% YTD returns - Highlighted weak breadth beneath the headline index move. Equal-weighted S&P 500 performance: Up 4% YTD - Demonstrated that the rally is not broad-based. AI companies’ share of public market gains: 90% - COTU/Philippe presentation claimed AI companies accounted for most of the public-market gains. Annual AI CapEx discussed: About $200 billion - Referenced as the scale of infrastructure spending going into AI this year. Required economic output to justify CapEx: $1.8 trillion - Philippe’s ‘monkey math’ estimate for 25% return on invested capital. Potential labor-force improvement needed: 5% - Philippe argued a 5% labor productivity gain could justify the investment. Global GDP growth target: 3% to 4% - Satya’s argument that AI could lift world growth enough to justify the spend. Venture investment in 2021: $715 billion - Cited as the peak of venture capital deployment during the ZIRP era. Forecast venture investment in 2024: $250 billion - Used to show normalization to roughly one-third of 2021 levels. AI deals year-to-date: About 200 deals - Illustrated current AI funding activity. AI investment year-to-date: $22 billion - Amount invested in AI deals so far this year. Average AI valuation: $1 billion - Average valuation for AI deals discussed in the venture section. Average AI round size: $120 million - Average round size for AI financing deals. AI valuation premium: 5x to 6x - AI companies were said to raise at 5 to 6 times the valuation of non-AI companies. Private unicorn count: 1,400 - Used repeatedly to describe the overhang of late-stage private companies. Employee growth in the 1,400 unicorns: 75% to 10% - LinkedIn growth slide showing sharp deceleration in hiring. 2024 IPO count: Fewer than during 2008 crisis - Used to emphasize how weak the IPO market has been. OpenAI revenue estimate: $3.5B to $4B - Discussed as a possible candidate for public markets. OpenAI consumer revenue estimate: About $2.1B of $3.4B - Derived from credit card data analysis mentioned in the conversation. OpenAI consumer churn: About 65% annually - Credit-card cohort estimate implying about 35% annual retention. Tesla vote support: Retail 90-10; institutional 70-30 - Described the shareholder support for Elon Musk’s compensation package. ISS/Glass Lewis opposition: Both recommended no - Proxy advisors advised against the Tesla compensation package. Delaware lawsuit ask: $4B to $5B - Referenced as the size of the legal compensation demand in the Tesla case. Plaintiff share ownership: 9 shares - Highlighted to criticize the scale of the derivative lawsuit. OpenAI valuation referenced: $90B private valuation - Used to frame the question of whether a fast-growing company at that scale should go public.

Pivotal Quotes: "If it's 10 billion or bust, that's a bit command." — Bill: Reaction to the idea that companies must reach a $10B market cap before IPOing. "We have to prepare the company that there's going to be a mismatch in timing here between investment and return." — Satya Nadella (as paraphrased in discussion): Used to describe the long lag between AI infrastructure spending and realized economic value. "If you have to be at a billion in revenue and growing, yes, like if you want a 10x multiple, you gotta be growing 30, 35%." — Bill: Argument that an overly high IPO threshold makes the venture business much riskier and more capital intensive.

Implications: AI will likely reshape capital markets, corporate governance, and labor faster than most investors expect. Companies may need to go public earlier, stay disciplined on CapEx, and prepare for valuation resets, while regulators may face pressure to make public markets more accessible.