Jerry Neumann on the Problem With Investing in AI Right Now

Summary of Jerry Neumann on the Problem With Investing in AI Right Now

by Bloomberg

48m•November 12, 2025
Listen to Episode

Overview of Odd Lots episode — "Jerry Neumann on the Problem With Investing in AI Right Now"

This episode of Bloomberg's Odd Lots features longtime venture capitalist and Columbia Business School professor Jerry Neumann. The conversation centers on Neumann's contrarian take that "AI will not make you rich" for most investors and startups today. He compares AI to past large but uneven technological shifts (notably shipping containerization), explains where value from AI is likely to be created versus captured, and gives practical perspective for VCs, founders and everyday investors about timing, exits, and what kinds of companies are likely to benefit.

Guest background

  • Jerry Neumann: veteran VC (started 1997), angel investor and academic (Columbia Business School).
  • Recent writing: essay "AI Will Not Make You Rich" for Colossus; co-author of Founder vs. Investor: The Honest Truth About Venture Capital from Startup to IPO.

Key takeaways

  • AI is likely revolutionary in capability, but value creation (societal gain) is different from value capture (who gets the profits). Capture may accrue to deep-pocketed incumbents (Microsoft, large cloud providers) rather than many early-stage startups or small investors.
  • Historical analogy: containerization (the shipping box) transformed global trade and productivity but did not create a large class of instant, conspicuous new individual fortunes for outside investors — value often flowed to incumbents or those who sold early.
  • Neumann’s framework (drawing on Carlotta Perez/Kondratiev cycles): AI looks like the tail end/capstone of the information/computer wave rather than the start of a wholly new industrial revolution.
  • Winners will be those who use AI to expand markets (volume, reach, product variety) rather than merely cut costs. Using AI to shrink headcount and squeeze margins is the "wrong move".
  • The current market resembles a picks-and-shovels phase: infrastructure and suppliers (chips, data centers, cooling) are attracting big investment; these plays can be profitable but are not pure AI “winners” in the sense of new consumer-facing monopolies.
  • Timing and exits matter more than ever: abundant late-stage private capital and big corporate investors mean many companies can stay private longer; VCs still must consider fund life and the IPO window.

Main arguments and supporting examples

  • Value creation vs. value capture:
    • AI can create huge productivity gains. Who pockets that value is uncertain — could be consumers, foundational model owners, cloud providers, or larger incumbents.
  • Containerization analogy:
    • Shipping containers solved a systems problem (not a single flashy gadget). The system-level change delivered massive productivity gains and enabled companies like IKEA and Walmart to scale — but that didn’t produce obvious new-forged billionaire investors broadly across the board.
    • Some incumbents (shipping firms, retailers) benefitted more than independent inventors; some early entrants who sold early captured value (e.g., Sealand founder sold to a larger firm).
  • Computers → AI as end vs. new wave:
    • Neumann argues AI is the natural culmination of computing as a knowledge machine; it may be the last big advance in that wave rather than a brand-new industrial revolution.
  • Investor behavior and market structure:
    • Foundation models and large-scale AI systems are capital-intensive. Anyone can "build a model" in principle, but few have the cash to build facility-scale compute and train at top performance.
    • Abundant late-stage private capital and corporate backers (Microsoft, others) reduce pressure to IPO; being public has regulatory/operational downsides that encourage staying private.
    • Power-law VC dynamics remain: VCs back many companies hoping a small number will be massive winners ("lottery ticket" approach).

Practical implications / Actionable advice

  • For early-stage investors:
    • Be selective: look for opportunities that can scale by using AI to expand markets or reduce friction for consumers (not just to cut payroll).
    • Recognize capital limits: building foundational AI platforms typically requires large pools of capital — small funds or angels are better placed to find downstream winners who apply AI in knowledge-intensive industries.
    • Think about exit timing; fund lifecycle still matters. "Sell early" can sometimes be rational if the IPO window closes or competitive dynamics shift.
  • For founders and operators:
    • Use AI to build new value for customers (volume/market expansion, better products), not as a headline-driven cost-cutting PR move.
    • Re-engineer processes around AI rather than retrofitting models into unchanged workflows — big gains often require systems changes, retraining, and new product designs.
  • For public market investors:
    • Distinguish infrastructure plays (chips, data centers) from durable consumer/business franchises that will capture long-term value.
    • Consider that some market froth exists — overvaluation does not always equal systemic bubble risk like 2000 or 2008, because much of the capital is corporate/late-stage.

Notable quotes

  • "There's a difference between value creation and value capture."
  • "If you're firing people because of AI, you're doing it wrong — you should be using AI to say I can use my people to do more."
  • "I think this is the culmination of the computer wave... This is why I compared it to containerization."
  • "VCs invest saying one of these is going to become valuable. It's the lottery ticket theory."

Market risks and structural observations

  • Infrastructure bubble risk: large, simultaneous investments (chips, data centers, power) may overshoot near-term demand; overbuilding could depress returns for suppliers.
  • Leverage and systemic risk: unlike dot-com or housing, much AI-related capital sits with corporates and late-stage funds, which may blunt systemic fallout if valuations fall — but localized damage (companies, employees) would still occur.
  • SPACs and market windows: IPO access still matters; staying private is attractive when late-stage capital is available, but that creates different dynamics (less public price discovery, longer exits).

Recommended further reading / resources mentioned

  • Jerry Neumann’s Colossus essay: "AI Will Not Make You Rich"
  • Carlotta Perez — Technological Revolutions and Financial Capital (on Kondratiev waves)
  • The Box (book) — on containerization and global trade history
  • Michael Malone — The Intel Trinity (on early microprocessors)

Bottom line

Neumann argues investors and founders should temper get-rich-fast hopes around AI today. AI is transformative, but the economic story will be messy: much of the value may accrue to incumbents and infrastructure owners, and the biggest sustainable wins will come from companies that use AI to expand what they sell (volume, variety, reach) rather than merely to cut costs. Think systemically, focus on where AI amplifies knowledge-intensive businesses, and be realistic about timing and exit mechanics.