Equinix was one payroll from death in 2002, then became the data-center king

Equinix built carrier-neutral data centers where telecoms could exchange traffic - a model with no customers and no revenue at the start. In the internet boom it went all-in, and when the largest US carriers chose Equinix as their exchange point, the model looked proven.

Then the bubble burst. Peter Van Camp was brought in in 2000 to take Equinix public and pulled off an August 2000 IPO even after its lead investment bank backed out. But by 2002 the stock was down to 31 cents, the company carried $263m of debt it could not cover, and it was, by its own account, one payroll away from closing.

The management team renegotiated the debt and got new terms, buying a second life. Equinix then outlasted the crash, bought the data centers abandoned by telecoms that had quit the business, and used its interconnection network to build the dominance that now underpins the AI era.

Raised cash in an Aug 2000 IPO, renegotiated $263m of debt in 2002, and outlasted the telecoms that sold Equinix their data centers.

The IPO, despite the lost lead bank, bought the cash and time the company needed to keep building.

Renegotiating the debt instead of defaulting preserved the network effect the data centers had already created.

Survivors bought failed rivals' capacity cheaply, turning a crisis into consolidation.

Being first to a network-effect business is worthless if the financing dies. The firms that survive a bust get new terms on their debt, then buy the cheaper assets their failed rivals leave behind.

Equinix survived to become the world's leading data-center and interconnection company, joining the Nasdaq-100 and later riding the cloud and AI boom. Its 2001-02 near-death became a case study in selling a network-effect business through a financing crash - and in buying the assets of the competitors that did not.