Market bubble can inflate a lot more before the end, says Niles Investment's Dan Niles

Key Concepts

• Agentic AI: An advanced form of artificial intelligence that can perform multi-step tasks autonomously (e.g., browsing websites, extracting data, and formatting spreadsheets) rather than just providing text-based answers.
• Token Generation: A metric used to measure the computational output and demand of AI models; a significant increase in tokens indicates higher compute power requirements.
• Dot-com Bubble Analogy: A comparison between the current AI market surge and the late 1990s internet boom, focusing on growth cycles and market valuation.
• Liquidity/M2 Money Supply: The total amount of money circulating in an economy; its contraction or expansion significantly impacts market valuations.

---

1. The AI Market vs. The Dot-com Bubble

Dan Niles, founder and portfolio manager at Niles Investment Management, argues that while the current market behavior mirrors the dot-com bubble, this does not preclude further growth.

• Historical Comparison: During the internet build-out, the NASDAQ rose 109% in the first three years following the release of Netscape Navigator (1994). In the three years since the release of ChatGPT (late 2022), the NASDAQ has risen 122%.
• Growth Trajectory: Niles notes that the internet boom saw two additional years of significant growth after the initial three-year period. He projects at least one more year of strong performance for the current AI cycle.
• Fundamental Differences: Unlike the dot-com era, which was characterized by speculative companies with little cash flow (e.g., Pets.com), current AI leaders are backed by substantial earnings and tangible technological advancements.

2. The "Agentic AI" Catalyst

Niles identifies the finalization of "Open Claw" (a reference to advanced agentic AI frameworks) on January 30th as a major inflection point.

• Shift in Compute Demand: Traditional "chat-based" AI (like early ChatGPT or Gemini) requires significantly less compute power than "Agentic AI." Agentic AI requires 10 to 100 times more compute power to execute complex, multi-step workflows.
• Data Evidence: The impact is visible in token generation statistics. In the two months prior to the finalization of Open Claw, token growth was approximately 20%. In the two months following, growth surged to over 120%.
• Outlook: Niles expects strong growth to continue through the beginning of next year, at which point the market will face "harder comparisons" (lapping the high growth rates of the previous year).

3. Lessons from the 2000 Market Crash

Niles explains the factors that caused the dot-com bubble to burst in March 2000, serving as a warning for current market participants:

• The Role of Liquidity: In 1998, the Federal Reserve was cutting rates. In 1999, despite the Fed raising rates, the "Y2K scare" prompted the Fed to flood the market with liquidity to prevent system crashes, causing the M2 money supply to rise by over 10%.
• The Turning Point: When the Y2K crisis failed to materialize, the Fed withdrew liquidity. Simultaneously, oil prices spiked, and the rate of internet traffic growth slowed from doubling every three months to doubling every year.
• Valuation Mismatch: The market valuations were built on the assumption of exponential growth; when the growth rate decelerated, the valuations could no longer be sustained.

4. Synthesis and Conclusion

The current AI market is in a high-growth phase driven by a fundamental shift toward Agentic AI, which has drastically increased the demand for compute power. While the market exhibits bubble-like characteristics similar to the late 1990s, the presence of strong earnings and the ongoing surge in token generation suggest that the cycle has room to expand further. However, investors should remain cautious of liquidity shifts and potential decelerations in growth rates, as these were the primary drivers that ended the previous internet-era bull market.