Why the Next Tech Crash Won't Look Like the Last One
Every market cycle generates predictions of the next crash. With tech valuations at historic highs and AI investment approaching dot-com-era levels, the comparisons to 2000 are inevitable — and mostly wrong.
The structural differences between today's tech sector and the late 1990s are fundamental enough that any correction would take a very different form.
Revenue vs. Speculation
The dot-com era was defined by companies with no revenue and astronomical valuations. Today's largest tech companies generate genuine cash flows that dwarf most traditional industries. Apple alone generates more quarterly profit than the entire NASDAQ 100 did in 1999.
Even the most speculative AI investments are building on real demand. Enterprise AI spending hit $180 billion in 2025, driven by measurable productivity gains rather than theoretical future applications.
- Dot-com peak (2000): 70% of IPOs had negative earnings
- Current tech: 85% of large-cap tech is profitable
- Enterprise AI spend: $180B (2025), projected $290B (2027)
- Cloud infrastructure: $600B+ annual run rate
What Could Go Wrong
This doesn't mean tech is immune to correction. The risks are different: regulatory intervention, AI capability plateaus, or geopolitical supply chain disruption could all trigger significant repricing.
The most likely scenario isn't a crash — it's a rotation. Capital flows from overvalued AI infrastructure plays into practical implementation companies. The picks-and-shovels trade gives way to the builders.
Positioning for Either Outcome
Whether you're bullish or bearish on tech, diversification across the value chain makes sense. Infrastructure, applications, and services each carry different risk profiles and will perform differently in various scenarios.