Browser → mass adoption → over-investment → bust. Consumer Internet built on telecom over-capacity. NASDAQ peak Mar-2000; ~78% drawdown by 2002. Survivors became 2010s incumbents.
(during the boom) This is a new economy. Old valuation rules don't apply. The Internet changes everything.
Our read
It was a cost-curve breakthrough creating real durable value, but valuations decoupled from earnings power. The bust was inevitable; the recovery was inevitable; the survivors compounded for 25 years. Both extremes (everything is broken / everything is fine) were wrong.
The wedge
In any cost-curve-breakthrough cycle, expect: (1) over-build phase 3-5 years, (2) bust phase 1-2 years with 70-90% drawdowns, (3) survivor-bias compounding 10-20 years. The wealth-creating phase is BOTH the boom and the recovery. The bust eliminates pretenders but not the underlying technology.
§ Thesis
What's actually shifting.
The Internet Boom (1995-2001) was the first general-purpose-technology cycle of the digital era. Mosaic browser (1993) → Netscape IPO (1995) → mass adoption → over-investment → bust (2000-2002) → consolidation → infrastructure-survivor compounding. Three structurally important takeaways: (1) The bust eliminated ~$5T market cap but the technology continued advancing. (2) Picks-and-shovels (Cisco, Oracle, hardware) suffered hardest in the bust because they had been levered to build-out demand that paused; pure consumer plays (eBay, Amazon, Yahoo) had different bust dynamics. (3) The eventual winners (Google IPO 2004, Facebook 2004, Amazon survival) emerged FROM the wreckage — late-1990s incumbents (Yahoo, AOL, Excite, Lycos) were displaced. The investment lesson: cost-curve breakthroughs produce 10-year boom-bust-recovery cycles; survivor-bias recovery delivers most of the durable returns.
NASDAQ peaked at 5048 (Mar 2000), bottomed at 1108 (Oct 2002 — a 78% drawdown), and didn't recover to its 2000 peak until 2015. The infrastructure build was real; the valuations were not. Survivors compounded for 25 years.
Survived bust by burning capital but staying lean enough; AWS (2006) became second business; today $2T+ market cap.
Google (2004 IPO)Industry-defining late-entrant.
Late-entrant in search; founded 1998 but IPO'd post-bust; captured incumbent (Yahoo) displacement; PageRank moat held for 15+ years.
Apple (post-2003)Cycle-leveraging late-stage compounding.
iPod (2001) → iPhone (2007) was the mobile-internet wave. Capitalized on infrastructure built but unused during bust. Largest market-cap company by 2018.
Equinix / Digital Realty / colocationDistressed-asset roll-up.
Acquired distressed datacenters during bust at <30% replacement cost. Served the cloud era 2010+.
Named trapped
AOL Time Warner merger (Jan-2000)Merger-of-the-decade failure.
Largest-ever merger announced at absolute market top. Combined market cap: $350B. By 2002: <$60B. Eventually unwound at fraction of value.
Worldcom (2002 bankruptcy)Largest US bankruptcy at the time.
IPO'd in 1999, all bankrupt by 2001. Burned ~$1B+ each on subsidized economics.
§ Signal tracking
What would tell you the shift is accelerating — or stalling.
Watch for (acceleration)
(historical context)
NASDAQ valuation (peaked at 5048 Mar-2000)
Internet penetration rates
IPO velocity and post-IPO returns
Ad-spending shifts to online
Broadband subscriber growth
Anti-watch-for (stalling / reversal)
(retrospectively visible)
Capex outpacing revenue at telecoms
IPO-to-profit time stretching beyond 5 years
Cash-burn rates accelerating without revenue ramp
Consumer-internet sectors built on subsidized economics
§ Lessons for analog application
What this shift teaches about active shifts that resemble it.
Cost-curve breakthroughs produce boom-bust-recovery cycles of ~10-15 years. The wealth creation comes from survival into the recovery, not from peak-cycle exposure.
Picks-and-shovels (telecom, hardware, infrastructure) suffer worst in the bust because they were levered to build-out demand that pauses. Application-layer with revenue may have softer landings.
Late entrants (Google 1998, Facebook 2004) often beat incumbents (Yahoo, AOL, Friendster) because they built post-bust without the cost-structure baggage.
Mergers at peak are leading indicators of correction. AOL+Time Warner is the canonical example.
Pure consumer-internet plays without revenue model fail catastrophically; B2B SaaS and infrastructure have softer trajectories. Watch revenue model maturity.
Real-estate exposure to engineering hubs benefits durably, even through bust. Bay Area real estate compounded through 2000-2002 contraction.
Surviving 70-90% drawdowns is a feature of the cycle, not a defect. The "this time it's different" frame at peak is wrong; the "this time it's broken" frame at trough is also wrong.
§ Related Lab findings
Where the mechanism is rigorously tested.
No Lab finding has been authored on this shift yet. The shift is tracked here as macro frame; rigorous mechanism testing comes when a finding is registered against the corpus.