Supermicro - the hardware underneath somebody else's brand

Supermicro was founded in San Jose in September 1993 by Charles Liang with Sara Liu and Wally Liaw. Liang was born in Chiayi, Taiwan, left a teaching post at his own university in the 1980s for Silicon Valley, and had spent the two previous years running a motherboard design company. By his own account the new business was profitable within six months, and it signed its first European distribution partnership in its founding year.

For most of its life it was the unglamorous half of the industry - motherboards, chassis and subsystems, sold on engineering rather than brand, to buyers specifying machines rather than buying them. That position looked like a ceiling for nearly three decades. It became the opposite when demand shifted to dense, power-hungry accelerated computing, and the company reported revenue of nearly 22 billion dollars in 2025 with about 6,200 employees.

One decision inside that history is worth separating out. It began developing energy-efficient servers in 2004, well before power was a purchasing criterion for most buyers, and manufacturing in Taiwan from 2012. Building for efficiency ahead of the market is a bet that only pays if the market eventually arrives, and in datacentre hardware it did - power and cooling became the constraint that decides what can be installed at all.

The lesson for anyone reading a technology timeline is about where value sits over time. A components company competing on engineering looked strategically inferior to a brand for about thirty years, and then the layer it occupied became the scarce one. Position in a stack is not permanent, and the layer that is boring during one cycle is frequently the constraint in the next.