20VC: Sequoia's David Cahn on The Winners and Losers in AI |...

David Cahn states that he believes the AI market is experiencing a bubble and notes its fragility.

Cahn revisits his previous prediction that the year of the data center was coming, emphasizing the underestimated physicality of AI infrastructure.

The discussion links the physicality of AI build-out to GDP growth, noting that AI construction booms are contributing significantly to economic metrics.

Cahn explains his "600 billion question" regarding the revenue needed to recoup AI compute investments, highlighting the importance of customer health.

Cahn discusses his prediction of construction delays in data centers, noting that this is starting to materialize.

The difficulty and complexity of building AI infrastructure are highlighted as a potential moat for companies.

Cahn identifies a miss in his previous predictions: the massive talent acquisition packages in AI, which he did not foresee.

The discussion delves into whether these large pay packages for AI talent are justified, considering the probabilistic nature of success.

Cahn identifies Meta's performance as his second major miss from the previous year, as their AI integration had not progressed as expected.

Cahn discusses his past thesis on vertical integration in AI, questioning if OpenAI and Anthropic's lack of it poses a challenge.

It is noted that OpenAI and Anthropic are becoming more vertically integrated by developing their own chips and securing power.

Cahn reiterates that the AI market is currently in a bubble, a consensus view now compared to a year ago.

The focus shifts to identifying winners and losers post-bubble, drawing parallels to the dot-com era.

Cahn explains how to balance playing in the current market with a long-term AI investment strategy, emphasizing his eight-year experience in AI investing.

Cahn discusses how companies like Databricks have successfully navigated market cycles through strong product-market fit.

Cahn outlines his framework for identifying winners and losers: consumers of compute benefit from bubbles, while producers are in a commodity business.

Cahn emphasizes that companies consuming raw resources like power and producing intelligence are more likely to succeed.

The discussion questions whether cloud providers like AWS, Google Cloud, and Azure are non-commodity businesses.

Cahn argues that we are in an "anomalous monopoly era" but warns against overextrapolating this to all businesses, as monopolies are not hiding in plain sight in AI.

Cahn contrasts the current AI era with the early tech era, where monopolies were less obvious and therefore easier to build.

Cahn concludes that monopolies are unlikely to exist in the AI era due to widespread awareness and competition, which is beneficial for consumers.

Cahn discusses the difficulty of breaking the mimetic chain in career choices, especially with the rise of AI.

Cahn explains that "consumers of compute win" but acknowledges that capital deployment often favors "producers of compute."

Cahn notes the shift from a contrarian idea of an AI bubble to a consensus view, attributing it to circular deals and big tech dynamics.

Cahn highlights the dangerous incentive to invest in companies that consume more capital, even if they are not the most efficient.

Cahn emphasizes the importance of focusing on companies that don't want to raise capital, citing Zoom as an example.

Cahn addresses the question of whether there is a coordinating mechanism for an AI bubble to pop, suggesting it's driven by incentives rather than coordination.

Cahn believes the AI bubble will deflate rather than pop, drawing parallels to Nassim Taleb's concept of fragility.

Cahn identifies the circular deals and shifting dynamics among hyperscalers (Microsoft/Amazon to Oracle/CoreWeave) as key indicators of AI bubble fragility.

The shift from expensive capital from big tech companies to cheaper capital from chip companies is seen as a major change in the AI ecosystem.

Cahn views the numerous AI infrastructure deals as "priming the pump" for future capital raising.

The financing of these AI build-outs is questioned, with a distinction made between debt-funded and equity-funded deals.

Cahn suggests that an AI bubble unwind would likely be an equity unwind, impacting portfolios more directly than a credit crisis.

Concerns are raised about the concentration of value in the "MAG7" tech companies and their reliance on the AI narrative impacting GDP.

Cahn agrees with the 5% GDP impact of AI but questions the assumption of high profit margins on that impact, referencing a McKinsey report.

Cahn emphasizes that GDP accrues to working people and that sustaining economic profit above the cost of capital is difficult.

Cahn identifies overestimation of demand, particularly in sectors like legal, as a problem in the current AI market.

Cahn believes the timeline for AGI is being overestimated, citing commentary from AI leaders pushing back the expected timelines.

Cahn contrasts the optimistic "lunchroom conversation" about AGI timelines with the more measured views of AI pioneers.

The concept of "kingmaking" in venture capital is discussed, with Cahn expressing skepticism about its effectiveness.

Cahn emphasizes that venture capitalists cannot make a company succeed; the company must already be successful, and founders are the primary drivers.

Cahn disagrees with the idea that VC capital directly creates competitive moats, stating that talent and product-market fit are more crucial.

Cahn believes margins matter as an indicator of product development but are not the sole determinant of success, as they can improve over time.

Cahn acknowledges that even companies with zero gross margins can become successful businesses, but his preferred investments have higher margins.

Cahn refers to his "WWDD" (What Would Doug Do) framework, focusing on making money for LPs and founders.

The "0 to 100 million revenue club" is seen as the new benchmark for successful AI companies, indicating rapid product-market fit.

Cahn emphasizes the importance of traction and customer love as indicators of a company's success.

Cahn highlights the "one to 50" growth metric as more critical than the "zero to one" growth for evaluating a company's potential.

Cahn shares examples of companies like Juicebox and Clay that took time to find product-market fit, highlighting the value of founder resilience.

Cahn discusses the tension between abundant capital and the need for disciplined growth, warning against overcapitalization.

Cahn applies Pat Grady's question about what everyone thinks they know but gets wrong to AI, suggesting that people underestimate the impact of momentum loss and "hidden risks."

Cahn believes people are underestimating voice as an AI interface, highlighting the potential of companies like Sesame.

Cahn argues for embracing visible risks in hiring, favoring younger, less experienced talent with clear potential over those with hidden risks.

Cahn critiques the "medic algorithm" of career choices, suggesting that AI's transformative impact requires updating these traditional approaches.

Cahn distinguishes between people focused on "what can I get" from a job versus those asking "what can I contribute."

Cahn notes the slow shift away from traditional career paths like investment banking and consulting, despite the rise of AI.

Cahn believes the defense sector is the "next AI" but is still in its nascent stages of adopting technological transformation.

The Ukraine war is seen as the "Transformer moment" for defense, indicating a shift in the necessity for technological adoption.

Cahn suggests that defense technology adoption is underestimated and that the sector is about 1% of the way to realizing its potential.

Cahn believes that the importance of AI in defense will become universally recognized, similar to the impact of ChatGPT.

Cahn expresses concern about the concentration of buyers in the defense sector and the perceived lack of incentive for government customers.

Cahn's framework suggests fewer, highly concentrated winners in defense, akin to "national champions," due to the single-customer dynamic.

Cahn discusses the digital transformation of defense, comparing it to the early days of cloud adoption in tech.

Cahn agrees that defense is not a broad category but rather a niche with a few key companies likely to succeed.

In a quick-fire round, Cahn shares that he learned to drive in the last 12 months, capitulating just before self-driving cars became widespread.

Cahn reflects on how becoming a father has made him less abstract and more focused on present needs.

Cahn emphasizes the importance of shared values in partner selection, noting his wife's intelligence and their shared values as crucial to their relationship.

Cahn's biggest miss was Datadog, due to a lack of focused effort on a priority deal.

Cahn believes voice as an AI interface is wildly undervalued, highlighting the potential of companies like Sesame.

Cahn's greatest excitement for the next 10 years is AI itself, viewing it as a transformative, once-in-human-history event.