10:33in productionCh. 1 · The four-transistor test/ 10:33 · ceiling 15 min
Founders
Morris Chang
He didn’t bet on Asia — he measured yield in Texas and Japan, then moved the factory.
Morris Chang built the semiconductor foundry model on direct observation — not theory, not trend, not funding. He saw yield differences between Texas and Japan. He ran a four-transistor pilot with IBM. He led TI’s global semiconductor business. Then he founded TSMC — separating design from fabrication, relocating scale to Asia, and creating the infrastructure that now underpins every major chip company.
Foundry relationships existed before Chang — but only as exceptions, not business models.
2:10
From project to platform
Chang turned an outsourcing experiment into a scalable, standalone business — the first pure-play foundry.
4:26
The view from group VP
His authority over TI’s global semiconductor business gave him visibility into yield, cost, and staffing — not just engineering.
6:31
Yield, not ideology
He concluded Asia would lead advanced manufacturing after comparing hard metrics — not culture, cost, or policy.
Worth your time?
Yes. Study the whole thing.
4.5/ 5
What works
business/ideas
business/management
business/strategy
What does not
business/companies
business/deals-and-ipos
business/finance
business/marketing
Study it if
business/founders
business/strategy
business/management
Skip it if
business/startups-and-venture
business/scandals
business/rise-and-fall
The written brief1 min read
What the company or idea is
TSMC is the world’s first and largest dedicated semiconductor foundry, founded on the principle that chip design and fabrication should be separated.
How it actually makes money
TSMC makes money by fabricating semiconductors for fabless chip designers — a pure-play foundry model where it charges for wafer production, not chip design or branding.
What works
Separating design from fabrication enabled rapid iteration, lower capital barriers for chip startups, and economies of scale across customers — proven by TSMC’s dominance in logic node leadership and market share.
What does not
The model does not insulate against geopolitical risk. It concentrates advanced fabrication in one jurisdiction — Taiwan — without solving sovereignty, export control, or supply chain fragility.
What to take from it
Chang’s insight was empirical, not ideological: yield, staffing stability, and technician quality — not ideology or policy — determined where advanced manufacturing could scale.
Is it worth your time
Yes. The foundry model reshaped global semiconductor economics, relocated advanced manufacturing to Asia, and created a structural dependency that still defines chip geopolitics.