Economics
Why this boom isn’t a repeat of the dot-com bubble
Still 1998, not 1999
The useful answer is still 1998. Real revenue and profits are here. Euphoria is a pricing and cash-timing story, not a crash call yet.
The question people keep asking is whether this boom feels like 1998 or 1999. Anyone who lived through early 2000 carries scar tissue. That question is not the same as an answer. Overfitting the old pattern is its own risk.
What 1999 meant, and what people are afraid of repeating
In the late 1990s, internet companies were often valued on metrics that were not dollars: page views, hits, stories about the future. The crash in early 2000 wiped out long tails of companies. Offices emptied. Paper wealth vanished in months. Every new bull market, that scar tissue asks the same question: is this a bubble again?
Cycle read
Still 1998
Real revenue · durable profit for strongest
Cash quality vs late-90s vanity metrics (page views/hits). Products in market; customers pay. Euphoria can price a real market too aggressively — clock is when repricing comes, not whether revenue is fake today.
This cycle has dollars where the last scare cycle had vanity metrics
The distinction that matters now is cash quality. Leading AI labs and the infrastructure around them are booking real revenue and, for the strongest names, durable profit. Products are in market. Customers pay. Euphoria can still price a real market too aggressively — the clock is when that repricing comes, not whether revenue was fake today.
Edge froth
50–100×
Late private · top-line stretch
Unproven teams at 50–100× look like late-dot-com checks. One-of-one operators can defend high multiples; thin wrappers cannot wear pure-software multiples forever. Froth is a pricing problem — not proof Anthropic/OpenAI-scale revenue is imaginary.
Cash timing is the founder problem inside the euphoria
If the cycle still has years of heat left, the practical question is not deny the boom. It is when to take chips off the table and when to raise. Delaying a raise for a prettier price risks a closed window. San Francisco housing is a local wealth-wave mirror, not by itself the national crash signal.
Cash timing
Raise open
Optionality · when the window is open
Take chips carefully by role (late-stage slice ≠ early founder cash-out signal). Delaying a raise for a prettier price risks a closed window. SF housing is a local wealth-wave mirror, not by itself the national crash signal.
How to read this: Still 1998 = cash-quality cycle. 50–100× = edge froth. Raise open = founder cash timing inside the heat.
Quiet: True 1999→2000 match would be widespread monopoly-only multiples, vanity metrics back, late-stage sellers finding no bids — froth at edges now, not empty-office base case.
Still 1998 means price carefully, raise when the window is open if you need the cash, and do not confuse a real profit cycle with a permission slip for every private beta to clear at a hundred times sales.