Morris Chang's Bet: When Nobody Believed the Market Existed

In 1967, Morris Chang was 36 years old and already a General Manager at Texas Instruments.

He wasn't running a marginal operation. He was running TI's most critical semiconductor division — and turning it into the world's largest, most profitable semiconductor business. By 1972, he was promoted to Vice President. In that role, he invented a strategy that would later transform the entire industry.

But there was a problem: Morris Chang was Chinese.

In the American tech industry of the 1970s and '80s, there was a ceiling written in no rulebook. No matter how capable you were, no matter how much money you made for the company, the top seats rarely went to Asians — regardless of performance. This was a common reality in that era's American tech industry, not an isolated incident.

In 1978, Chang was transferred out of the semiconductor division and moved to consumer electronics. This was not a reasonable job rotation — TI's core was semiconductors; consumer electronics was a peripheral business. For the person who had built the world's strongest semiconductor division, this was a de facto demotion.

In 1983, something even more striking happened. His title was changed to "Manager of Quality and Personnel Effectiveness."

You don't need to understand corporate politics to read what that title says. It says: you no longer matter.

TI ultimately gave the CEO seat to someone else. Not Morris Chang.

He left. In 1984 he briefly joined General Instrument, but the culture didn't fit — he left again quickly. A man who had spent 25 years at TI building its semiconductor division to global dominance found himself at 55 unable to find a position in American tech that matched his capabilities.

This is not a story about personal disappointment. It is a structural fact: in that era's America, regardless of competence, certain seats were simply not available to you.

---

25 Years at TI: What Morris Chang Took With Him

Morris Chang was born in Ningbo in 1931. He came to the United States at 18, first attending Harvard before transferring to MIT to study mechanical engineering, then earning his PhD from Stanford in 1964.

But the place that truly shaped his business thinking wasn't school. It was TI.

He joined TI in 1958. His first assignment was demanding: raise the transistor yield for IBM mainframes from 10% to 20%. He did it, and rose quickly. As General Manager, he invented a pricing strategy — Learning Curve Pricing.

The logic: deliberately set prices low at the start, low enough to barely break even. The goal is volume. As volume scales, manufacturing yield and efficiency improve rapidly as accumulated experience compounds — unit costs continuously decline. Once costs fall, you have margin to reinvest in the next-generation process. Better process reduces costs further and improves the product, winning more volume. More volume funds the next reinvestment cycle.

This was not a price war. It was a flywheel trading time for structural advantage.

The point isn't what grievances Chang suffered at TI. The point is: in those 25 years, he mastered every fundamental principle of semiconductor manufacturing — yield, pricing, economies of scale, customer relationships. When he left, he didn't take away bitterness. He took away a complete playbook that others would need twenty years to learn.

Learning Curve Pricing wasn't a textbook concept. He had hammered it out at TI with real money and real stakes. Later, he installed the exact same strategy inside TSMC.

---

Taiwan Caught Him

In 1985, the Taiwan government invited Morris Chang to return and lead the Industrial Technology Research Institute (ITRI).

This invitation was not coincidental. Taiwan at the time was the world's contract manufacturing king — but making garments, umbrellas, and electronic components, with margins of 4–5%, stuck at the bottom of the value chain. Technocrats of the Sun Yun-suan and K.T. Li generation understood clearly: if Taiwan didn't move into semiconductors, it would always be working for others. But to make that leap, you needed someone who truly understood semiconductors from the inside — not an academic, but someone who had fought in the trenches at a place like TI for 25 years.

For Chang, the invitation carried a double meaning. First, someone was finally offering him a position where he could actually do something, rather than being warehoused in a peripheral division to be consumed. Second — the American system had rejected him; the Taiwanese system had chosen to catch him.

In 1987, K.T. Li told him: come start a semiconductor company.

That was the true starting point of TSMC. Not a genius having a flash of inspiration in a laboratory — but a man who had been trapped for ten years finally stepping onto a stage that matched his capabilities.

---

"Real Men Have Fabs"

The business model Morris Chang proposed was called pure-play foundry.

At the time, the industry standard was IDM — Integrated Device Manufacturer — design and manufacturing under one roof. Intel did it this way, TI did it, AMD did it. You designed chips, you built your own factory to produce them. You controlled everything end to end.

Morris Chang said: I will only do manufacturing. I will never design my own chips. I will never compete with my customers.

The industry's reaction? Laughter.

AMD's founder Jerry Sanders had a famous line that precisely captured the consensus of that era:

"Real men have fabs."

If you have your own fab, you're a real player. Contract manufacturing? That's for the weak.

This wasn't some lightweight throwaway remark. Jerry Sanders was a heavyweight in the semiconductor world. His words represented the entire industry's belief system: if you don't have your own fab, you're not a real semiconductor company.

Investors weren't buying either. Of TSMC's $220 million in startup capital, 50% came from the Taiwan government, 28% from Philips, and the remaining local investors — who only invested because the government pressured them. Morris Chang himself received no initial equity. He later bought shares on the open market, slowly accumulating what grew to approximately $3 billion.

One thing must be stated honestly: TSMC's birth was not one man's insight alone. Without the Taiwan government's strategic will — 50% of the capital, the Hsinchu Science Park's infrastructure, long-term semiconductor talent development policies — this company could not have gotten off the ground. What Chang received was not a blank check, but this was also not pure market behavior. It was a national-level strategic bet, executed by a person who happened to possess all the necessary capabilities.

A founder who couldn't secure equity in his own company at founding — what does that signal? It signals that even the people writing the checks didn't fully believe this would succeed. What they provided was policy resources, not trust. What Chang received was an opportunity that he still had to prove.

---

The Core Insight: Non-Competition Is the Deepest Moat

This is the most important section of this piece.

What Morris Chang saw was not a technical breakthrough, not a market gap. He saw a structural trust problem.

In the IDM world, every company did both design and manufacturing. If Intel had spare capacity, it could theoretically do foundry work for others — but who would hand their most critical chip designs to Intel? Intel designs chips itself; your design is its competitive intelligence.

This is not paranoia. It is rational behavior.

The same logic applies to every IDM. TI could manufacture for you, but TI is also your competitor. AMD could manufacture for you, but AMD is also competing for your market. Across the entire industry, no one could be trusted — because everyone was simultaneously a manufacturer and a competitor.

Chang's insight was precisely this: if you never design your own chips, you will never compete with your customers. If you never compete, customers will be willing to hand you their most confidential, most mission-critical designs.

Non-competition = trust. Trust = customers willing to give you their crown jewels. This is not a moral commitment — it is structural design.

Chang himself said it clearly in his interview with Acquired: "We should compete on different dimensions — we should not be IDMs."

The dimension he chose to compete on was not better technology, lower prices, or faster speed. It was I will never be your enemy. This is a position that no amount of money or technology can replicate — because any company that is already doing design cannot credibly make this promise.

Intel can't. TI can't. AMD can't. Only a company that declares from day one that it will never do design can occupy this position. The first mover is the permanent occupant.

But it must also be acknowledged: "why didn't others do this" has an answer beyond "they didn't see the trust gap." Wafer foundry carries enormous capital requirements. Without national-level backing, very few parties would stake this scale of bet on a market that didn't yet exist. The IDM model had its own internal logic; incumbents had existing interests to protect and wouldn't easily dismantle their vertical integration. Chang's insight was genuine — but turning insight into reality required not just vision, but capability, timing, and a system willing to cooperate.

When the podcast host Ben asked about competition with Intel, Chang's reply was brief: "I didn't say we would never catch up with Intel." Ben responded: "Just look at where things stand in 2025."

He wasn't boasting. He was stating a fact he spent forty years proving.

---

Flywheel Ignition: From Absorbing Losses to Creating an Ecosystem

TSMC's earliest customers were Intel and Motorola.

But they didn't come because they believed in the pure-play foundry model. They came because they had surplus, money-losing orders they needed to offload. Their own fab capacity needed to be reserved for their most profitable products — the low-margin, small-volume work? Send it to that new company in Taiwan.

What they transferred was not opportunity. It was their losses.

Morris Chang accepted. Then he pulled out the TI playbook.

Learning Curve Pricing, activated.

Take on large volumes of low-margin orders → Manufacturing volume rises → Accumulated process experience → Yield improves, unit cost falls → Profits reinvested in next-generation process R&D → More advanced process attracts more customers → More orders.

Once this flywheel started turning, what followed was something Morris Chang had foreseen — an entirely new industrial structure was born.

Because there was now a manufacturer you could trust, a class of companies that only did design and never built factories suddenly became viable. They were called fabless — semiconductor design companies without fabs.

Nvidia, small at first. Qualcomm. Broadcom. These companies exist because TSMC exists.

Morris Chang didn't "discover" the fabless market. He created the conditions that allowed the fabless market to exist.

And TSMC's operating principle was clear: treat every customer equally, regardless of size. Not prioritize Intel because of large order volumes; not shortchange Nvidia because it was small back then. This isn't idealism — it's business logic. You don't know which small customer today will become your largest customer tomorrow. The cost of damaging any relationship far exceeds any short-term gain.

By 1994, TSMC went public on the Taiwan Stock Exchange at a market cap of $4 billion. In 1997, it listed on the New York Stock Exchange. By 1999, annual revenue exceeded $1.5 billion, with a customer roster including Motorola, Siemens, and TI — several of whom had once mocked the model.

Seven years after being called "a business with no market," twelve years later it was the infrastructure for a global industry.

---

A Framework for CEOs: What Did Morris Chang Actually See?

Many people say Morris Chang was a technical genius, or that he was lucky to catch the semiconductor wave.

Both miss the point.

What Morris Chang invented was not just a company — it was an industrial structure. What he did can be broken into three layers:

Layer One: Identify the trust gap. Across the entire industry, every manufacturer was also a designer, so no one could be trusted. This was not a technology problem. It was a structural problem.

Layer Two: Design a position that cannot be replicated. "Never do design" is not a slogan — it is a structural commitment. Any company already doing design cannot credibly make this promise, so the first mover permanently owns this position.

Layer Three: Use the flywheel to turn position into barrier. Learning Curve Pricing let him trade low-margin orders for volume; volume created cost advantage; cost advantage attracted more customers; more customers funded further reinvestment. Every turn of the flywheel raised the barrier to entry for those who came later.

But there is a Layer Zero that often gets overlooked: he needed a system to receive him. TI's 25 years gave him the playbook, but the structural ceiling in America had sealed off the space for him to execute it. The Taiwan government didn't just provide capital, the science park, and policy — more critically, it provided a position. It let a man who had been misplaced for a decade finally sit where he was meant to sit.

This is not saying Chang's success was the product of "state support." It is saying: even the strongest founders need a system willing to catch them. The insight was his. The execution was his. But the stage did not fall from the sky.

What does this mean for CEOs today?

You don't need the best technology. You don't need the most money. What you need is to find the position in your industry where, for structural reasons, no one can currently be trusted — and then make yourself the first and only person worth trusting.

But remember: seeing the position is not enough. You also need a system that allows you to go occupy it. Morris Chang spent 25 years preparing, then waited another decade for that system to appear.

The bet he made in 1987 was not "pure-play foundry will succeed." His bet was: trust itself is a market. And once that market is created, it will grow so large that no one can ignore it.

He won the bet. But this was a bet made under deep uncertainty — not a foregone conclusion that looks obvious only in retrospect.

---

This is Part 1 of the TSMC Trilogy. Part 2 will focus on how TSMC went from chaser to leader — step by step surpassing Intel in the technology race.

---

Related Reads

• Sony PlayStation: Bet the Company — Not Because They Loved Games, But Because They Could No Longer Be Just a Hardware Company — Another story of structural insight: when you see the market nobody believes in, what do you do?
• Dyson's Succession Play: When a Billionaire Family Faces the Second-Generation Problem — A builder who spent decades building one system, then designed a second system to outlast himself.
• Horace Luke Owes $4.6M and Vanished — The Same Traits That Built Gogoro Destroyed It — The inverse case: a visionary who saw a real infrastructure gap, but picked the wrong capital structure, wrong market, and wrong governance.
• Nintendo: From Cards to Kingdom — Another company that survived by making a bet nobody believed in — in a different era, with a different playbook.