I Survived the Dot-Com Bubble. The Lessons I Learned From That Time Are Keeping Me Out of These 2 Tech Giants, and Here's Why

The article warns that elevated valuations and potential industry downturns make Micron and Nvidia vulnerable despite strong recent results. Micron’s latest-quarter revenue rose 379% year over year, with net income of nearly $37.7 billion versus $3.2 billion a year earlier; its shares once fell 93% from their 2000 peak, while the article notes long-term customer contracts may cushion a downturn. Nvidia’s revenue grew 106% and net income 126% to $59.7 billion, but the author compares its customer-financing exposure to Cisco’s dot-com-era risk and recommends current shareholders hold while warning of a growth disruption if financed customers fail.

Analysis

The key asymmetry is not “AI demand versus no AI demand,” but where the cycle’s pricing power sits. HBM contracts may improve visibility, yet they do not necessarily protect Micron if customers renegotiate, supply expands, or conventional DRAM/NAND weakens; verify contract duration, price-reset terms, and HBM mix before treating current earnings as durable. Samsung and SK hynix adding capacity could eventually shift bargaining power back toward buyers, while Nvidia’s accelerator demand is ultimately supported by customer returns on data-center spending—not build-out announcements alone.

For Nvidia, the Cisco analogy is most useful as a balance-sheet and utilization checklist, not a forecast. The article provides no evidence that Nvidia is financing customers. Monitor customer funding, receivables, inventory, and capex concentration for signs that demand is being pulled forward or supported by circular financing. A slowdown in hyperscaler spending could hit Nvidia first through orders and expectations, then indirectly pressure HBM pricing and Micron earnings.

Near term, momentum can persist while supply remains tight; over 1–3 months, earnings commentary on HBM pricing, delivery commitments, and customer capex is more informative than headline AI demand. Over 6–18 months, capacity additions and accelerator substitution/custom silicon are the principal risks. The contrarian point: a low cyclical P/E is not a floor when peak earnings may be the denominator; equally, the Cisco comparison can overstate Nvidia’s vulnerability absent evidence of customer credit stress or weakening returns on AI investment.

Key Decisions for Investors

• Avoid treating Micron’s reported earnings multiple as normalized valuation. Consider a defined-risk MU put spread only if supplier capacity plans rise while HBM pricing/contract terms weaken; exit the bearish thesis if pricing and customer commitments remain firm through subsequent guidance.
• Do not short NVDA solely on the Cisco analogy. Track hyperscaler capex revisions, customer concentration, receivables/inventory, and evidence of customer financing or deteriorating AI returns; a broad spending slowdown would be the stronger catalyst.
• Avoid a simple long-NVDA/short-MU pair as a hedge: both are exposed to the same data-center capex cycle, and a demand shock could pressure both. Reassess relative exposure when HBM supply and pricing disclosures clarify which company retains more bargaining power.
• Falsification / alert: if Micron reports resilient HBM pricing and firm commitments despite industry capacity growth, defer the downside trade; if major cloud customers cut capex or Nvidia reports weakening orders alongside rising receivables or inventory, the risk case becomes more actionable.