The “winner’s curse”: why history suggests markets may misjudge AI “losers”

Summary

• Delayed disruption: Historical examples show incumbents often remain profitable long after new technologies emerge.
• Reliability premium: Concerns around AI reliability and data quality are reinforcing the value of trusted, proprietary datasets.
• Value rotation: Some investors are reassessing traditional sectors as AI valuations become more concentrated.

LONDON, April 24, 2026 — As investors rush to identify winners in the artificial intelligence boom, market history suggests caution: the biggest gains may not always come from the most obvious disruptors.

Recent volatility in technology stocks has revived a familiar debate—whether markets are once again underestimating the staying power of incumbents in the face of transformative innovation.

Lessons from past disruptions

Market historians point out that technological shifts rarely eliminate incumbents overnight.

• During the rise of railways in the 19th century, canal companies continued to generate strong returns for years.
• In the early internet era, traditional industries initially written off by investors later delivered stable and, in some cases, superior returns.

These patterns highlight a recurring theme: markets often price in disruption faster than it materializes in reality.

Reliability and data as competitive moats

While AI adoption is accelerating, concerns remain around:

• Accuracy and “hallucinations” in AI-generated outputs
• Dependence on training data quality
• Risks in fully automated decision-making

This has reinforced the importance of sectors that rely on:

• Verified, structured datasets
• Regulatory compliance
• Human oversight

Industries such as financial services, legal information, and healthcare data are increasingly viewed as holding defensive advantages in an AI-driven environment.

Rethinking valuation extremes

The strong rally in AI-linked stocks has led to high market concentration in a handful of companies, prompting some investors to:

• Diversify into under-owned sectors
• Re-evaluate companies with strong cash flows but lower AI exposure
• Focus on long-term earnings resilience rather than near-term hype

Rather than a wholesale rejection of AI, this reflects a broadening of investment narratives beyond pure technology plays.

Why this matters

• Market balance: Overconcentration in a few AI stocks increases systemic risk
• Investment cycles: Historical patterns suggest value opportunities may emerge in overlooked sectors
• Data economics: Proprietary and high-quality data is becoming a key differentiator
• Risk management: Reliability concerns may slow full automation in critical industries

FAQs

Q1. Are “AI losers” actually good investments?

Not necessarily across the board, but some companies may be undervalued if markets overestimate the speed of disruption.

Q2. Why is proprietary data important in AI?

AI systems depend heavily on high-quality data. Exclusive, verified datasets can provide a strong competitive advantage.

Q3. Is this similar to the dot-com era?

There are parallels in terms of valuation concentration and optimism, but the underlying technology adoption today is more mature.