Sequoia Partner, David Cahn on Who Wins in AI, Defence & The New $0–$100M Playbook

Bubble Assessment & Market Dynamics

• The speaker confirms the current AI sector is in a bubble, noting that "fragility" is now a consensus view among major bulls like Sam Altman and Jeff Bezos.
• Unlike previous bubbles, the "killer app" or specific monetization endpoint for the massive compute investment has not yet been proven to the end consumer.
• Investors in "consumers of compute" are predicted to outperform "producers of compute" (commodities) because overproduction drives down prices for buyers while increasing costs for producers.
• Circular financing deals (where chip companies finance data center build-outs by taking equity or revenue stakes) have shifted risk absorption from hyperscalers (Microsoft, Amazon) to smaller entities (Oracle, CoreWeave, chipmakers), increasing systemic fragility.

Physicality of AI & Infrastructure Constraints

• The market has shifted from valuing AI in "bits" (compute models) to "atoms" (physical infrastructure), with power and steel now being the primary constraints.
• Generator capacity is sold out until 2030, and data center construction is facing significant delays due to supply chain complexity and permitting.
• The speaker notes that 800 gigawatts of power build-out represents an $8 trillion question, and 250 gigawatts represents a $20 trillion question, both of which are largely unfunded.
• The $600 billion revenue question (derived from a $150B NVIDIA investment needing 50% gross margins to pay back) has grown to $840 billion by mid-2025 but remains unanswerable regarding end-user health.

Investment Frameworks & Sector Trends

• Vertical Integration: Major AI labs (OpenAI, Anthropic) are becoming vertically integrated "steel servers and power companies" to secure compute, with a trend toward owning their own chips and power infrastructure.
• Defense as the "Next AI": The speaker identifies defense as the upcoming transformational sector, comparable to the post-Transformer/AI moment, driven by a 50-year lag in modernizing military tech.
• National Champions: The defense sector is expected to consolidate into a few "national champions" per country (e.g., Anduril in the US, Celan in Israel, Stark in Europe) rather than supporting a broad ecosystem like SaaS.
• Talent Acquisition: Extreme compensation packages (up to $1B for a single researcher) reflect ecosystem desperation rather than proven marginal impact, with the speaker arguing that macro variables drive progress more than individual genius.

Market Concentration & Risks

• The MAG-7 represents ~40% of the S&P 500, creating a concentration risk similar to Japan's equity market in the 1990s, though the unwinding mechanism is expected to be equity-based rather than credit-based (unlike 2008).
• AGI timelines are being extended by leading researchers (e.g., Richard Sutton, Ilya Sutskever) to 20–30 years, contrasting sharply with the "100 days" sentiment among newer, less experienced participants.
• The speaker rejects the "king-making" ability of venture capital, asserting that capital cannot create success if the founder, product, and market fit are not already present.

Future Outlook & Specific Bets

• Voice Interfaces: The speaker identifies voice as a wildly undervalued AI interface, citing the recent investment in "Sesame" (a conversational AI with high product-market fit) as evidence that the "staring at a screen" era may be ending.
• GDP Impact: While agreeing with the McKinsey/Sony estimate that AI will affect ~5% of GDP, the speaker argues that the profit margin on that GDP will be far lower than the 50% assumed by some, as monopolistic pricing is unlikely in the AI era.
• Long-term Thesis: Despite the bubble, the speaker maintains a 50-year horizon where AI will be a generational change reshaping society, emphasizing that investors should focus on companies that can survive market volatility through product-market fit.

Career & Founder Dynamics

• Hiring Shifts: The speaker advises hiring 23–24-year-old AI generalists over experienced senior engineers, as the "level playing field" of Gen AI negates the value of decades of legacy software experience.
• Memetic Algorithms: Young professionals should not blindly follow the "smartest people one year ahead" algorithm, as the AI disruption renders previous career trajectories obsolete.
• Growth Trajectory: Successful AI companies are now expected to move from zero to $100 million in revenue rapidly, serving as a leading indicator of true product-market fit.

Key Quotes & Forward-Looking Statements

• "If you overproduce compute, prices go down, your COGS goes down, and your gross margin goes up."
• "The lesson that punches you in the stomach in venture is you can't make a company succeed."
• "Defense is the next AI... We're like 1% there on catching up."
• "Anything multiplied by zero is zero" (referring to bankruptcy risk during market volatility).
• "AI is the most important story of our lifetime... it's going to be a once in human history kind of event."