Internet vs. AI: Lessons Hidden in a 1998–2000 Nasdaq–S&P

AI will become a universal layer across tools, workflows, and industries, in the same way electricity became a foundational layer of all machinery. But because data from the electrification era is sparse, the internet revolution offers a clearer and more measurable model for comparison.

1. The Data: Nasdaq vs S&P 500, Jan 1998 – Mar 2000

 

1. 1998-01-09 → 1998-04-22 — Up

Duration: 103 days
Nasdaq: 27.6% (leg), 27.6% (cumulative)
S&P 500: 21.9% (leg), 21.9% (cumulative)
Δ vs S&P: 5.7% (leg), 5.7% (cumulative)

Narrative:

Relief rally post-Asian crisis as U.S. growth stayed firm; Greenspan’s Fed signaled benign inflation. PC/Internet adoption accelerated; early web portals and e‑commerce names (AMZN, YHOO) drew strong buy‑side interest; analysts highlighted Y2K-related IT outlays aiding Dell/Cisco supply chains. | Details: Post-Asian crisis stabilization; Greenspan signaled benign inflation. AMZN and YHOO gained attention as web traffic soared; Dell/Cisco cited as Y2K/enterprise beneficiaries. Buy-side flows favored growth; multiple expansion resumed.

Sources:

·        FOMC historical statements (1998): https://www.federalreserve.gov/monetarypolicy/fomc_historical.htm

·        Amazon (history & 1990s growth): https://en.wikipedia.org/wiki/Amazon_(company)

·        Yahoo! in the 1990s: https://en.wikipedia.org/wiki/Yahoo!

·        Cisco 1998 annual report (archival): https://investor.cisco.com/annual-reports/default.aspx

 

Mirror Period:  2025-04-08 → 2025-11-03 — Up

Duration: 209 days
Nasdaq: 51.9%
S&P 500: 37.6%
Δ vs S&P: 14.2% 
NVDA: 92%

Here’s a quick “what moved the market” timeline for Apr 8 → Nov 3, 2025 (your start date to last completed close):

• Tariff shock & the bottom (Apr 2–10): A sweeping U.S. tariff program sparked a sharp global selloff; China’s retaliation on Apr 4 deepened losses. The S&P 500 set its lowest close on Apr 8, which we’re using as your anchor. Reuters+1

• Tariff whipsaws & partial walk-backs (mid-Apr): Headlines about pauses/adjustments to the new tariffs fed big intraday swings, then markets stabilized. Reuters

• AI earnings drive the rebound (May→summer): Mega-cap tech led a powerful rally. Nvidia posted record revenue for the quarter ended Apr 27, 2025 (data-center strength), and shares ripped higher; Big Tech’s AI capex and backlogs (e.g., Microsoft) reinforced the theme. Investopedia+2NVIDIA Newsroom+2

• Rates tailwind builds (Sep–Oct): The Fed cut 25 bps in September and again in October, easing financial conditions; inflation ran ~3% YoY into early fall. The 10-year U.S. yield hovered near ~4.1% by early Nov. Forbes+2bls.gov+2

• Tariff flare-up & wobble (Oct 10): A fresh round of China-related tariff moves hit risk appetite and tech briefly, but the uptrend held into early November. Reuters

• Mega-cap momentum into Nov: Headlines around Nvidia’s $5T milestone and continued AI demand kept QQQ leadership intact (even as investors debated AI spend levels). Bloomberg

Net effect: from the tariff-panic low (Apr 8) to Nov 3, QQQ outperformed SPY thanks to AI-led mega-caps, while two Fed cuts plus moderating inflation underpinned the broader tape; intermittent tariff headlines injected volatility but didn’t derail the rebound.

2. 1998-04-22 → 1998-06-15 — Down

Duration: 54 days
Nasdaq: −10.5 (leg), 14.0% (cumulative)
S&P 500: −4.7 (leg), 16.4% (cumulative)
Δ vs S&P: −5.8 (leg), −2.4 (cumulative)

Narrative:

Post‑run digestion: valuation pushback and headline risk. U.S. DoJ’s Microsoft antitrust case (filed May 18, 1998) raised platform uncertainty; hardware inventories built at some PC makers, prompting cautious notes from sell‑side tech analysts. | Details: U.S. DOJ filed its Microsoft antitrust suit on 1998-05-18; states joined (focus on browser tying). Analysts flagged rich PC valuations and inventory build; cautious notes on hardware margins.

Sources:

·        DOJ v. Microsoft filed May 18, 1998: https://www.justice.gov/atr/case/us-v-microsoft

·        United States v. Microsoft (overview): https://en.wikipedia.org/wiki/United_States_v._Microsoft_Corp.

 

Mirror Period:  2025-11-04 → 2025-11-20 — Down

Duration: 12 days
Nasdaq: -8%
S&P 500: -4.5%
Δ vs S&P: 3.5% 
NVDA: -18%

⚠️ AI bubble

Despite the strong numbers, Nvidia’s stock went from being up ~5% (or more) earlier in the session to down a few % later. Some key reasons:

The broader market sentiment turned cautious: investors are wary that the valuations of high-flying tech/AI stocks like Nvidia may be stretched. 

Even though Nvidia beat, the question “can all this AI investment pay off?” remains hanging. As one Reuters piece puts it: “its upbeat results did not dispel concern … whether AI spending will pay off.” 

Macroeconomic/interest-rate concerns: A stronger-than-expected jobs report (119,000 jobs added) revived fears that the Federal Reserve may delay rate cuts, which tends to hurt growth/tech stocks. 

It appears the reversal was intraday: early gains + then selling pressure, possibly profit taking.

3. 1998-06-15 → 1998-07-20 — Up

Duration: 35 days
Nasdaq: 17.4% (leg), 33.6% (cumulative)
S&P 500: 9.9% (leg), 27.7% (cumulative)
Δ vs S&P: 7.5% (leg), 5.9% (cumulative)

Narrative:

Earnings beats from network/PC bellwethers (CSCO, DELL) and solid online traffic data rekindled risk appetite. Buy‑the‑dip flows returned as funds leaned back into growth/tech. | Details: Upbeat guidance from CSCO/DELL/INTC; improving online adoption metrics. Momentum funds rotated back into tech; volatility compressed into earnings season.

Sources:

·        Cisco Systems (history): https://en.wikipedia.org/wiki/Cisco

·        Dell (history): https://en.wikipedia.org/wiki/Dell

·        Intel (history): https://en.wikipedia.org/wiki/Intel

 

4. 1998-07-20 → 1998-10-08 — Down

Duration: 80 days
Nasdaq: −29.6 (leg), 0.0% (cumulative)
S&P 500: −18.9 (leg), 5.2% (cumulative)
Δ vs S&P: −10.7 (leg), −5.2 (cumulative)

Narrative:

Global shock: Russia’s Aug‑17 default and the LTCM near‑collapse (late Sep) drove forced deleveraging. Banks and prime brokers reduced risk; tech/growth bore the brunt. Fed later engineered a private rescue for LTCM, but fear dominated. | Details: 1998-08-17 Russia default (GKO/OFZ) and ruble devaluation; 1998-09-23 LTCM private recap ($3.6B) led by a 14-bank consortium under NY Fed auspices. Forced deleveraging hit high-beta tech hardest; funding markets stressed until Fed eased.

Sources:

·        1998 Russian financial crisis: https://en.wikipedia.org/wiki/1998_Russian_financial_crisis

·        NY Fed LTCM statement (Sep 23, 1998): https://www.newyorkfed.org/newsevents/statements/1998/ps980923

·        Long-Term Capital Management: https://en.wikipedia.org/wiki/Long-Term_Capital_Management

 

5. 1998-10-08 → 1999-01-29 — Up

Duration: 113 days
Nasdaq: 76.6% (leg), 76.6% (cumulative)
S&P 500: 33.4% (leg), 40.4% (cumulative)
Δ vs S&P: 43.2% (leg), 36.2% (cumulative)

Narrative:

Powerful policy‑backed rebound: the Fed delivered three rate cuts (Sep–Nov ’98). Dot‑com IPO window reopened; CSCO, INTC, MSFT, and networking plays surged; analysts talked ‘networked economy’ and Web scale effects. | Details: Fed cut rates 25bp on 9/29, emergency 25bp on 10/15, and 25bp on 11/17. AOL-Netscape deal announced (1998-11-24) boosted Web platform narratives; networking/infrastructure leaders (CSCO, JNPR nascent) surged; IPO window reopened.

Sources:

·        FOMC 1998 rate cuts (Sep/Oct/Nov): https://www.federalreserve.gov/monetarypolicy/fomc_historical_year.htm#1998

·        AOL–Netscape acquisition (Nov 24, 1998): https://en.wikipedia.org/wiki/AOL#Acquisitions

 

6. 1999-01-29 → 1999-02-17 — Down

Duration: 19 days
Nasdaq: −10.3 (leg), 58.2% (cumulative)
S&P 500: −3.0 (leg), 36.6% (cumulative)
Δ vs S&P: −7.3 (leg), 21.6% (cumulative)

Narrative:

Quick reset after a torrid Q4/Q1 run. Select earnings wobbles and stretched multiples triggered profit‑taking; notes from skeptics warned on revenue quality at some dot‑coms. | Details: Hot-money consolidation after Q4/Q1 melt-up; sell-side warned about revenue quality/marketing spend at select dot-coms; spread-widening and valuation checks triggered a brief risk-off.

Sources:

·        Dot-com bubble (valuation context): https://en.wikipedia.org/wiki/Dot-com_bubble

 

7. 1999-02-17 → 1999-04-12 — Up

Duration: 54 days
Nasdaq: 15.6% (leg), 83.2% (cumulative)
S&P 500: 9.4% (leg), 49.2% (cumulative)
Δ vs S&P: 6.2% (leg), 34.0% (cumulative)

Narrative:

Momentum re‑accelerated: the late‑’98 AOL–Netscape deal signaled platform consolidation; Priceline’s March 31, 1999 IPO fed animal spirits. Sell‑side raised targets on traffic/user growth over GAAP earnings. | Details: Priceline IPO (1999-03-31) exploded higher; AOL-Netscape integration and portal land-grab fed TAM narratives; analysts raised targets on user/traffic growth over GAAP EPS.

Sources:

·        Priceline IPO (Mar 31, 1999): https://en.wikipedia.org/wiki/Priceline

·        AOL–Netscape: https://en.wikipedia.org/wiki/Netscape#Acquisition_by_AOL

 

8. 1999-04-12 → 1999-04-19 — Down

Duration: 7 days
Nasdaq: −9.7 (leg), 65.1% (cumulative)
S&P 500: −5.1 (leg), 41.7% (cumulative)
Δ vs S&P: −4.6 (leg), 23.4% (cumulative)

Narrative:

Brief shakeout around Tax Day and pre‑FOMC nerves. A few analysts flagged excessive price/sales ratios across portals and e‑commerce names. | Details: Tax-day selling and pre-Fed caution; notes flagged extreme price/sales for portals/e-commerce; quick shakeout with shallow breadth damage.

Sources:

·        Price-to-sales discussion (dot-com era): https://en.wikipedia.org/wiki/Price-to-sales_ratio

·        1999 FOMC meetings: https://www.federalreserve.gov/monetarypolicy/fomc_historical_year.htm#1999

 

9. 1999-04-19 → 1999-07-16 — Up

Duration: 88 days
Nasdaq: 22.1% (leg), 102.5% (cumulative)
S&P 500: 10.0% (leg), 56.3% (cumulative)
Δ vs S&P: 12.1% (leg), 46.2% (cumulative)

Narrative:

Spring‑summer melt‑up: bandwidth/fiber build‑out stories (WCOM, Qwest) and enterprise networking demand lifted the complex. Upbeat calls on routers/switches and hosting drove multiple expansion. | Details: Bandwidth/fiber build-out (WCOM, Qwest), hosting, and routers/switches optimism; enterprise capex tailwinds; aggressive analyst target hikes on CSCO/JDSU/NTAP.

Sources:

·        WorldCom (telecom boom): https://en.wikipedia.org/wiki/WorldCom

·        Qwest (fiber buildout): https://en.wikipedia.org/wiki/Qwest

·        JDS Uniphase: https://en.wikipedia.org/wiki/JDSU

·        Network Appliance (NetApp): https://en.wikipedia.org/wiki/NetApp

 

10. 1999-07-16 → 1999-08-10 — Down

Duration: 25 days
Nasdaq: −13.1 (leg), 76.4% (cumulative)
S&P 500: −9.7 (leg), 41.1% (cumulative)
Δ vs S&P: −3.4 (leg), 35.3% (cumulative)

Narrative:

Rate and earnings jitters: Fed tightening path (June & Aug hikes) and guidance haircuts in pockets of tech sparked a pullback; funds rotated defensively while volatility rose. | Details: Fed hikes underway (6/30 and 8/24) raised discount rates on cashflows; earnings/guidance haircuts in pockets of tech; risk pared into summer as VIX rose.

Sources:

·        FOMC rate hikes (Jun 30 & Aug 24, 1999): https://www.federalreserve.gov/monetarypolicy/fomc_historical_year.htm#1999

·        VIX basics: https://en.wikipedia.org/wiki/VIX

 

11. 1999-08-10 → 2000-01-03 — Up

Duration: 146 days
Nasdaq: 65.9% (leg), 201.3% (cumulative)
S&P 500: 12.5% (leg), 59.8% (cumulative)
Δ vs S&P: 53.4% (leg), 141.5% (cumulative)

Narrative:

Blow‑off advance into Y2K: record dot‑com IPOs/secondaries; hyperscale narratives around web infrastructure; mega‑caps (INTC, MSFT, CSCO) led flows; performance‑chasing into year‑end by momentum funds. | Details: Y2K capex sprint; blockbuster IPOs/secondaries; mega-caps MSFT/INTC/CSCO led; analysts framed 'first-mover/network effects' stories; performance-chasing into year-end by momentum funds.

Sources:

·        Year 2000 problem (Y2K): https://en.wikipedia.org/wiki/Year_2000_problem

·        IPO mania (1999 tech IPOs): https://en.wikipedia.org/wiki/Dot-com_bubble#Initial_public_offerings

 

12. 2000-01-03 → 2000-01-06 — Down

Duration: 3 days
Nasdaq: −9.8 (leg), 171.6% (cumulative)
S&P 500: −2.8 (leg), 55.5% (cumulative)
Δ vs S&P: −7.0 (leg), 116.1% (cumulative)

Narrative:

Post‑Y2K profit‑taking and de‑risking after a parabolic Q4. Early earnings previews prompted trims in crowded winners. | Details: Profit-taking after parabolic Q4; early warnings on inventory/demand normalization; high gross exposures trimmed.

Sources:

·        Dot-com bubble peak dynamics: https://en.wikipedia.org/wiki/Dot-com_bubble

 

13. 2000-01-06 → 2000-01-21 — Up

Duration: 15 days
Nasdaq: 13.6% (leg), 203.0% (cumulative)
S&P 500: 2.8% (leg), 59.0% (cumulative)
Δ vs S&P: 10.8% (leg), 144.0% (cumulative)

Narrative:

Dip‑buying drove a swift rebound; upbeat pre‑announcements and analyst reiterations on mega‑cap tech restored confidence. | Details: Beats/reiterations from mega-cap tech; buy-the-dip reflex strong; CTA/momentum models re-engaged.

Sources:

·        Mega-cap tech (MSFT, INTC, CSCO): https://en.wikipedia.org/wiki/Microsoft

 

14. 2000-01-21 → 2000-01-28 — Down

Duration: 7 days
Nasdaq: −8.2 (leg), 177.0% (cumulative)
S&P 500: −5.6 (leg), 50.3% (cumulative)
Δ vs S&P: −2.6 (leg), 126.7% (cumulative)

Narrative:

Pre‑FOMC (Feb 2, 2000) rate‑hike anxiety; valuation concerns resurfaced. Some hedge‑funds reduced gross exposure. | Details: Pre-FOMC (2000-02-02) hike anxiety; valuation concerns and redemptions at some funds; liquidity thinner into month-end.

Sources:

·        FOMC Feb 2, 2000 meeting context: https://www.federalreserve.gov/monetarypolicy/fomc_historical_year.htm#2000

 

15. 2000-01-28 → 2000-03-10 — Up

Duration: 42 days
Nasdaq: 29.9% (leg), 259.0% (cumulative)
S&P 500: 10.8% (leg), 67.0% (cumulative)
Δ vs S&P: 19.1% (leg), 192.0% (cumulative)

Narrative:

Final euphoric leg: telecom + Internet synergy, capacity‑expansion plans, unprecedented new issues; target hikes leaned on TAM narratives more than earnings—classic late‑cycle behavior. | Details: Final blow-off leg: telecom/fiber mania, record new issues, upgrades leaning on TAM. Speculative turnover spiked; breadth narrowed to leaders; classic late-cycle markers appeared.

Sources:

·        Dot-com bubble (late-stage signs): https://en.wikipedia.org/wiki/Dot-com_bubble#Market_pinnacle

 

16. 2000-03-10 → 2000-03-31 — Down

Duration: 21 days
Nasdaq: −9.4 (leg), 227.0% (cumulative)
S&P 500: −0.6 (leg), 66.3% (cumulative)
Δ vs S&P: −8.8 (leg), 160.7% (cumulative)

Narrative:

Post‑peak unwind: valuation gravity reasserted; global growth jitters and reports of fund stress (e.g., Tiger Management closing) hit sentiment; analyst tone turned defensive as misses/guidance cuts appeared. | Details: Japan recession headlines hit risk (mid-March); reports of hedge-fund stress (e.g., Tiger Management closure news) dented sentiment; first guidance cuts in high-flyers met with aggressive selling.

Sources:

·        Tiger Management closure (Mar 2000): https://en.wikipedia.org/wiki/Tiger_Management

·        Dot-com peak (Mar 10, 2000): https://en.wikipedia.org/wiki/Dot-com_bubble#Bursting

 

2. What the Table Teaches

a. Volatility is the Toll

From one leg to the next, Nasdaq’s moves averaged ±15 – 25%. The largest single drop (−29.6%) erased nearly all prior gains—yet those who stayed survived to see +76% rebounds.
📈 Lesson: successful investors plan for volatility instead of predicting it.

b. Beta Cuts Both Ways

Every up-leg shows Nasdaq beating the S&P by +5–50 points (Δ Change %), and every correction overshoots on the downside.
📉 Lesson: higher-beta assets demand smaller size and tighter risk discipline.

c. Emotion Drives Regime Shifts

Each cycle of the table maps to crowd psychology:

1. Hope → early recovery (1998 Q1)

2. Greed → “new paradigm” story (1999 Q1 – Q4)

3. Euphoria → blow-off (Jan 2000)

4. Denial → Fear → Collapse (Mar 2000 onward)
   🧠 Lesson: sentiment extremes are timing indicators, not forecasts—trim when stories replace earnings; add when fear is loud but credit is calm.

d. Macro Liquidity Dominates

The biggest drawdowns weren’t about earnings—they were liquidity events (Russia/LTCM ’98, Fed hikes ’99, Japan ’00).
💡 Lesson: watch credit spreads, dollar funding, and central-bank tone—AI stocks won’t defy tightening any more than dot-coms did.

e. Compounding Reality

Despite massive gains, the cumulative ROR whipsaws show that two bad legs can erase a year’s return.
⚙️ Lesson: protect principal first; compounding resumes only after capital survives.

3. Applying the Lessons to the AI Cycle

Dot-Com Era Insight	AI-Cycle Application	Actionable Rule
Narratives inflate before fundamentals arrive.	Early-AI firms will tout “platform dominance” before profits.	Separate enablers (chips, infrastructure) from story stocks; size accordingly.
Late-cycle parabolas end abruptly.	AI indices could jump +30% in 6 weeks, then retrace −15%.	Pre-commit to trim 10–25% after any +25% surge in < 2 months.
Liquidity shocks cause real crashes.	Watch real yields + credit spreads; AI won’t be immune to funding stress.	Hedge or cut beta when 10-yr > 5% and credit > 150 bps.
Rotation to quality precedes new bull legs.	Profitable AI infrastructure will bottom first.	Re-enter via leaders with cash flow > capex.
Crowd euphoria peaks with “this time is different.”	“AI replaces everything” headlines = signal, not truth.	Move from growth to value sleeve; hold cash/T-bills for redeploy.

Practical Playbook for 2025–2030 AI Wave

1. Core / Explore: 70% diversified index (S&P + global); 30% AI sleeve (semis, infra, platforms).

2. Trend filter: Stay fully invested while AI ETF > 200-day; cut half when broken.

3. Buy-the-dip tiers: Add ¼ position after −10%; ½ after −20% if credit stable.

4. Take-profit rule: Trim 10–20% after +25–30% vertical moves.

5. Sentiment watch: Track fund inflows, IPO count, analyst language (“paradigm,” “new economy”). When hyperbolic → harvest.

6. Re-risking signal: Post-panic Fed easing + earnings trough = time to add.

7. Measure AI vs S&P deltas: widening positive deltas = overheating; narrowing = relative safety returning.

---

4. The Emotional Compass

Emotion	Market Behavior	Investor Advantage
Fear	Underpricing of durable assets	Accumulate methodically
Relief	Policy support, rate cuts	Add on confirmation
Greed	Easy gains, analyst euphoria	Tighten risk, trail stops
Euphoria	Blow-off, parabolic rise	Trim aggressively
Denial → Pain	Valuation cracks	Preserve capital; wait
Despair → Hope	Capitulation; fundamentals improve	Begin next accumulation

---

5. Final Takeaway

The table from 1998–2000 is not history—it’s a mirror.
Each cycle of innovation (railways, electricity, Internet, AI) follows the same rhythm: narrative → euphoria → collapse → utility.
Smart investors use history not to time peaks but to design systems that survive them.

In the AI era, the winners will be those who treat volatility as information, not trauma.