Are We in a Bubble Yet (vs. the Dot-Com Era)?

 

• Not yet. Tech leadership today (AI cycle) is strong but far less extreme than the late-1990s blow-off.

• The largest annual spreads of QQQ over SPY in the 2020s (+30.5 in 2020; +28.8 in 2023) are well below 1998–1999 (+56.7, +80.9).

• Drawdowns are milder: one bad year in 2022 (≈−32% for QQQ) vs. three consecutive ~−30% to −40% years in 2000–2002.

• 2025 YTD looks like orderly strength (SPY ≈ +18%, QQQ ≈ +20%), not a frenzy.

• Over long horizons, QQQ still beats SPY: 10-yr CAGR edge ≈ +4.8%/yr; 40-yr edge ≈ +3.0%/yr—evidence of structural leadership, not mania.

---

What’s in the numbers

From your table of calendar-year returns (1985 → 2025 YTD) and the rolling CAGRs:

• Internet bubble surge (late ’90s):

  • 1998: SPY +28.6% vs QQQ +85.3% → spread +56.7

  • 1999: SPY +21.0% vs QQQ +102.0% → spread +80.9

• Internet bust:

  • 2000–2002 (QQQ): −36.1%, −33.3%, −37.4% (three straight, very deep)

• AI-era surges:

  • 2020: +18.4% vs +48.9% → spread +30.5

  • 2023: +26.3% vs +55.1% → spread +28.8

• Calmer recent years:

  • 2024: +25.0% vs +25.9% → spread +0.9

  • 2025 YTD: ~+18% vs ~+20% → spread ~+2

• Rolling CAGRs (as of 2025):

  • 5-yr: 14.4% (SPY) vs 15.1% (QQQ) → +0.7

  • 10-yr: 14.8% vs 19.6% → +4.8

  • 20-yr: 10.7% vs 14.6% → +3.9

  • 40-yr: 11.5% vs 14.5% → +3.0

---

Current vs. Dot-Com: Five Clear Differences

1) Magnitude of the blow-off

• Dot-com: Back-to-back monster spreads (+56.7, +80.9) and a +102% QQQ year (1999).

• Now: Big, but not parabolic (largest 2020s spread is +30.5; best QQQ year since 2019 is +55.1).
  Takeaway: Today’s upside is powerful—but nowhere near the late-’90s verticality.

2) Depth and duration of the unwind

• Dot-com: Three-year tech collapse (~−36/−33/−37%).

• Now: One bad year (2022 ≈ −32%) followed by swift normalization (2023 +55%).
  Takeaway: The modern cycle has corrected quickly, not in a grinding, multi-year bust.

3) Annual fireworks vs. decade compounding

• Dot-com: The story was annual blow-offs.

• Now: The story is decade-scale compounding (QQQ’s 10-yr edge ≈ +4.8%/yr), with moderate annual gaps.
  Takeaway: This looks more like sustained structural leadership than a market-wide melt-up.

4) Position of 2025 on the spectrum

• Data point: 2025 YTD (SPY ~+18, QQQ ~+20) = healthy advance; spread ~+2.
  Takeaway: Momentum remains up—but not euphoric.

5) Where bubble risk would start flashing

• Historically, multi-year spreads > +40–50 and doubling years clustered together were the tell.
  Takeaway: We’re not at those levels; if we get there for multiple years, bubble risk rises sharply.

---

What to do with this (practical playbook)

1. Anchor to the 10-yr trend. The sustained QQQ edge argues for keeping core exposure to the innovation complex (AI compute, platforms, foundries, software).

2. Treat the spread as the canary.

   • Watch QQQ–SPY annual spread.

   • Yellow zone: high-20s to low-30s (powerful, still normal).

   • Red zone: repeat +40–50s (late-’90s territory).

3. Rebalance on extremes. If spreads lurch into the red zone or we see 1999-style single-year outcomes, pre-define trims.

4. Diversify within “tech.” Blend semis, hyperscalers, software, and critical suppliers; pair a SPY core with a QQQ satellite to smooth cycle risk.

5. Have the drawdown plan ready. If a 2022-type reset returns, be prepared to add methodically rather than chase peaks.

---

Bottom line

No—this isn’t the dot-com bubble (yet). The 2020s are defined by compounding leadership rather than blow-off excess. Until we see multi-year, 1999-scale spreads and clustered doubling years, the data supports staying with the trend—with guardrails for the day things do go vertical.