The Great AI Unwind — Which Companies Survive the Adoption Phase? | Kai Wu

Lead-Lag Live33mApril 1, 2026

Get the full intelligence

Search transcripts, export clips, track mentions, and explore all topics from “The Great AI Unwind — Which Companies Survive the Adoption Phase? | Kai Wu” inside PodZeus.

AI-Generated Summary

In this episode of Lead-Lag Live, host Michael Guyad and guest Kai Wu of Sparkline Capital explore the pivotal shift from AI infrastructure buildout to the adoption phase, arguing that the market is at a critical inflection point. Wu contends that while the past few years saw over a trillion dollars poured into AI hardware, data centers, and chip manufacturing—benefiting companies like NVIDIA, Microsoft, and Meta—the next phase will favor early adopters who leverage AI to enhance productivity and competitive advantage. Drawing parallels to the dot-com boom and railroad era, Wu warns that infrastructure-heavy 'Mag-7' firms may be sowing the seeds of their own decline due to unsustainable capital expenditures and overcapacity, while the true long-term winners will be companies that integrate AI into their core operations. The discussion also covers the risks of circular financing among AI players, the deflationary potential of AI, and the strategic importance of intangible value—such as intellectual property, brand equity, and human capital—in portfolio construction, especially in a world of rising trade tensions and regionalized economies. Wu advocates for active stock selection focused on intangible-intensive, multinational firms as a hedge against geopolitical and economic volatility. Key takeaways include: 1) The AI cycle is transitioning from infrastructure to adoption, favoring users over builders; 2) Overinvestment in AI infrastructure could lead to overcapacity and margin compression, undermining the business quality of Mag-7 firms; 3) The most resilient companies will be those that use AI to drive innovation and efficiency, not just build it; 4) Intangible value investing—focusing on IP, brand, and human capital—offers a powerful framework for outperformance in a fragmented, trade-sensitive world; 5) International markets aren’t inherently weak; underperformance stems from underinvestment in intangibles, not geography. The episode concludes with a call for thoughtful, active investing over passive exposure to capital-weighted indices, especially in the AI era.

Key Takeaways
1

The AI cycle is shifting from infrastructure buildout to adoption—early adopters will outperform infrastructure builders.

2

Mag-7 companies may be undermining their own business quality by overinvesting in AI data centers and capital-intensive models.

3

Intangible value—IP, brand, human capital—is the new frontier for long-term outperformance, especially in a fragmented global economy.

4

Companies that serve large enterprises are more resilient to AI disruption due to high switching costs and liability shields.

5

Active stock picking focused on intangible-intensive, multinational firms offers a superior alternative to passive exposure to the S&P 500.

…and 3 more takeaways available in PodZeus

Chapters
0:00
10 min

The AI Transition: From Infrastructure to Adoption

I've probably automated 70, 80% of my business now with agentic AI. And I myself also believe with that will probably come a massive surge in unemployment as people catch on to how powerful these tools are.

Highlight
10:00
10 min

Historical Parallels: Dot-Com, Railroads, and the AI Cycle

The telecom index was down over 90%. And from that, of course, that did not spell the end of the internet. We use the internet every day today.

Highlight
20:00
10 min

The Mag-7 Dilemma: Winners or Self-Destructors?

These are the companies building AI, yet they are potentially sowing the seeds of their own destruction or their own demise.

Highlight
30:00
10 min

Intangible Value Investing: The New Alpha

Wu introduces his investment framework based on intangible assets—IP, brand, human capital, network effects—as a superior lens for stock selection. He argues that these assets transcend borders and are less vulnerable to trade wars.

40:00
15 min

Globalization, Trade, and the Future of International Investing

The discussion shifts to trade policy and the underperformance of non-U.S. equities. Wu reveals that the root cause isn’t geography but underinvestment in intangibles, and that active selection of intangible-intensive global firms can unlock hidden value.

High-Impact Quotes
The telecom index was down over 90%. And from that, of course, that did not spell the end of the internet. We use the internet every day today.
Kai Wu4:20
Viral: 90.0
These are the companies building AI, yet they are potentially sowing the seeds of their own destruction or their own demise.
Kai Wu16:17
Viral: 88.0
I've probably automated 70, 80% of my business now with agentic AI. And I myself also believe with that will probably come a massive surge in unemployment as people catch on to how powerful these tools are.
Michael Guyad1:49
Viral: 85.0
Speakers

Host

Michael Guyad

Guest

Kai Wu
Topics Discussed
AI Adoption Phase95%Intangible Value Investing92%AI Infrastructure vs. Adoption90%Mag-7 Business Quality88%Active vs Passive Investing85%Corporate Innovation and Disruption80%Global Trade and Geopolitics75%Deflationary Impact of AI70%
People & Brands

Kai Wu

person

25xPositive

Michael Guyad

person

18xPositive

Sparkline Capital

organization

12xPositive

Lead Lag Report

organization

8xPositive

Microsoft

organization

7xNeutral

S&P 500

other

6xNeutral

NVIDIA

organization

6xNeutral

OpenAI

organization

6xNeutral

Meta

organization

5xNeutral

X (formerly Twitter)

other

4xNeutral

Get the full intelligence

Search transcripts, export clips, track mentions, and explore all topics from “The Great AI Unwind — Which Companies Survive the Adoption Phase? | Kai Wu” inside PodZeus.

Start discovering podcast insights today

Start with a 7-day trial and explore a growing catalog of popular podcasts. No credit card required.

No credit card required • 7-day trial • Cancel anytime