ENYO™SEMICONDUCTOR ← All insights
Insights · Own the Customer, Own the Code · 1980–1999 · 10 Nov 2026 · 2 min read

Part 2: Rent the Factory, Own the Design

In 1987 the Taiwanese government and Philips backed a new company called TSMC. Morris Chang built it to manufacture chips designed by other companies, and never to compete with them. That single decision changed what a chip startup could be. We still live with its consequences today.

Nov 10, 2026

Part 2 · 1980–1999 · Rent the Factory, Own the Design

In 1987 the Taiwanese government and Philips backed a new company called TSMC. Morris Chang built it to manufacture chips designed by other companies, and never to compete with them. That single decision changed what a chip startup could be. We still live with its consequences today.

Thanks for reading! Subscribe for free to receive new posts and support my work.

Fabless companies had already begun to appear, with Xilinx founded in 1984 and Qualcomm in 1985, and Nvidia followed in 1993 with a graphics chip that would later define a new market. None of them needed to build a factory. They spent their capital on engineers, design tools and customers, and rented production capacity when they needed it.

The venture model changed with it. A chip startup now needed tens of millions of dollars rather than hundreds of millions, and investors could back more teams and accept more risk. Success also became easier to measure. A company won design slots with major customers, or it did not.

One point here is easy to miss, and it matters for us. TSMC itself was not a capital-light venture. It was a large capital bet, made by a state with a long-term plan, and Korea made a similar bet through Samsung. So the fabless era did not prove that capital intensity always loses. It proved that two different roles exist, with states and very large groups carrying the factory and startups carrying the design, the IP and the customer.

The fabless winners owned two things that no foundry could take from them. They owned their intellectual property. They also owned the customer. Qualcomm licensed its wireless patents across the handset industry, and Nvidia built a community of developers on its graphics architecture.

For Bangladesh this era carries the most direct lesson, because our role is not the role TSMC played. A design-first strategy is the right starting point for the country. But design only creates lasting value when the design belongs to us. Engineering hours sold to a foreign client build our skills, and they do not by themselves build a company that owns anything.

That difference will decide the outcome of our strategy. Selling hours and owning IP are different businesses.

Thanks for reading! Subscribe for free to receive new posts and support my work.

Get new Enyo analysis by email. Subscribe on Substack