Warren Pies of 3Fourteen Research joins Excess Returns to break down the AI bull market, the macro risks investors should watch, and why the data still supports continued strength in semiconductors and equities. We discuss GPU demand, token usage, open source AI, Fed policy, housing weakness, oil, earnings growth, market valuations and the biggest risks to the current cycle.

Warren Pies on X
https://x.com/WarrenPies

3Fourteen Research
https://www.3fourteenresearch.com/

Caliban
https://www.3fourteenresearch.com/caliban

Main topics covered

• Which bearish AI arguments actually matter for investors

• Why regulatory risk may be the biggest long-term AI concern

• How data center spending is crowding out housing investment

• Why the Fed may struggle to cool AI-driven investment without hurting the labor market

• What GPU availability says about real-time AI compute demand

• Why open source AI is not yet replacing frontier models

• How token pricing and OpenRouter data help measure AI usage

• Why semiconductor stocks may still be in the middle of a major cycle

• How semis are being valued differently than traditional cyclicals

• Why Fed policy, earnings growth and market multiples are key to the second half of 2026

• What oil positioning and refined product inventories say about macro risk

• Why 3Fourteen remains constructive on equities despite rising overheating risk

Timestamps

00:00 Intro
01:04 Which bearish AI arguments have teeth?
04:00 Why AI regulation is the biggest long-term risk
07:03 Technology spending versus housing investment
11:03 How AI CapEx is showing up in inflation data
13:04 Why the labor market is more fragile than headline jobs data suggests
16:24 Why GPU availability is a cleaner signal than CapEx announcements
21:00 What token pricing and OpenRouter data reveal about AI demand
27:36 How 3Fourteen benchmarks frontier models against open source AI
30:00 Why the semiconductor selloff looked like a buyable dip
34:02 Are semiconductors still cyclical businesses?
38:08 Why Fed tightening could be the thing that ends the bull market
42:15 What the oil shock means now
45:47 Refined product inventories, crack spreads and energy stocks
47:18 Are earnings estimates becoming too optimistic?
50:49 Why the debasement regime still supports equities
54:05 Where to find Warren Pies and 3Fourteen Research

Ritavan joins Excess Returns to explain The System Gambit, a new framework for understanding competitive advantage, business strategy, AI disruption and long-term compounding. We discuss why traditional moat checklists can miss the real source of value, how companies can build systems competitors cannot copy, and what investors should look for when AI changes the game.

The System Gambit
https://amzn.to/4b0J32I

Main topics covered

• Why the traditional moat checklist can fail investors

• The three requirements for a true System Gambit

• How investors can evaluate business strategy from the outside

• Why code is not always the moat in the age of AI

• What history can teach investors about asymmetry and leverage

• Why AI adoption is not the same as AI value creation

• The difference between moving fast and understanding the game

• Lessons from Nokia, ASML, Amazon and Walmart

• How intangible investment and J curves can hide long-term value

• Why the best companies build compounding systems competitors cannot copy

• How investors can identify companies changing the game rather than optimizing the old one

Timestamps

00:00 Opening preview and introduction

04:00 The three ingredients of a System Gambit

08:49 Why code is not the moat in AI software

13:00 Skanderbeg and changing the rules of the game

17:00 Good moats, good narratives and asymmetric advantage

22:31 Microscope vs telescope as a lesson for AI

28:35 AI winners, losers and high dispersion markets

32:08 Signal quality, bottlenecks and why AI adoption is not enough

36:00 Nokia, agility and the failure to build a causal model

40:15 Why understanding the game beats speed

44:00 Intangible investment, the J curve and ASML's hidden edge

49:54 The contrarian AI thesis behind The System Gambit

54:00 How to recognize a real System Gambit

58:27 Amazon, Walmart and multi-paradigm compounding

1:03:00 Prime, FBA and platform leverage

1:07:00 Walmart's answer to Amazon

1:11:06 Closing thoughts and where to find Ritavan

On this episode of the 100 Year Thinkers, Chris Mayer and Matt Zeigler discuss long-term investing, 100-baggers, AI stocks, SpaceX valuation, founder-led companies, and why the best investments often come with brutal drawdowns. We also cover his new book The Investor's Odyssey, the danger of letting labels like AI do too much work, how to think about TAM and capital allocation, and why patience may be the biggest edge for investors trying to own great businesses for decades.

⁠Subscribe to the 100 Year Thinkers on Spotify⁠⁠

⁠⁠Subscribe to the 100 Year Thinkers on Apple⁠

The Investor's Odyssey: Resisting the Sirens and Playing the Long Game⁠

https://amzn.to/44BMXeJ⁠

Main topics covered

• Why SpaceX, AI and trillion-dollar IPOs are testing investor discipline

• How Chris Mayer thinks about valuation after watching Google become a huge winner

• Why great businesses can still be terrible investments at the wrong price

• The danger of letting labels like AI, quality and TAM replace real analysis

• Why many AI features may not create real customer value

• What the dot-com bubble can teach investors about AI adoption and shakeouts

• Why investors do not need to be early if a company is truly exceptional

• How to separate AI anecdotes from real financial impact

• Why capital allocation and return on invested capital matter more as companies scale

• How to evaluate founder control, governance, incentives and trust

• Why the best long-term stocks can still fall 50 percent or more along the way

• What rational exuberance might look like for long-term investors

Timestamps

00:00 Intro: Chris Mayer on AI, SpaceX and long-term investing

04:00 SpaceX valuation vs Google and the risk of paying too much

08:01 Why labels like AI and quality can do too much work

12:05 The AI pause, the dot-com analogy and where real value may emerge

16:06 Why investors do not need to be early when a business is real

21:00 Becoming a great company versus already being mature

25:10 Thinking about TAM, market share and realistic growth expectations

29:43 Corporate governance, free float and shareholder rights

34:27 How to judge founder trust, incentives and compensation

38:57 Employee ownership, culture and building enduring companies

43:02 Investor frustration in a lopsided AI-driven market

47:02 Why even a perfect stock picker would face brutal drawdowns

52:17 The rise of trillion-dollar IPOs and the question of rational exuberance

56:29 The Investor's Odyssey and playing the long game