Mark Liu: The Berkeley PhD Who Succeeded Morris Chang — Six Years of Decision-Making in the US-China Chip Crossfire

30-second overview: Mark Liu, born in 1954, holds a PhD in electrical engineering from UC Berkeley. He joined TSMC in 1991 and became chairman after Morris Chang's retirement in 2018. During his six years at the helm, TSMC became a central pawn in the US-China tech war: Arizona investment grew from an initial US$12 billion to over US$165 billion, and Liu repeatedly articulated TSMC's indispensability before government officials and the media. In June 2024, he formally retired and handed the reins to C.C. Wei.

The Berkeley Foundation

Mark Liu was born in 1954. After completing his bachelor's degree in electrical engineering at National Taiwan University, he went to the United States for graduate study and earned a PhD in electrical engineering from the University of California, Berkeley.¹ After finishing his studies, he first worked at the Electronics Research and Service Organization (ERSO) of the Industrial Technology Research Institute (ITRI), participating in Taiwan's early semiconductor technology R&D and building a deep grasp of industry dynamics.

This Berkeley training was repeatedly cited during his years at TSMC. The point of the citation was not the prestige of the institution but an epistemology: the engineer's ability to break down complex situations into measurable variables. What Liu later applied this method to was something far harder to measure than circuits — geopolitics.

The conventional wisdom was that succeeding Morris Chang required the aura and charisma of a "Morris Chang 2.0." But Liu offered the opposite answer — he never tried to replicate the founder's symbolic capital. He used numbers, process nodes, and investment commitments as his language of communication, replacing the evangelist's inspiration with the engineer's precision.

This choice happened to be perfectly suited to the era of the chip war. At the negotiating table of chip politics, no one bought the storyteller's narrative, but no one could refuse the numbers of process nodes. Liu's Berkeley foundation ultimately helped him find his own language in the halls of geopolitics.

From ITRI to TSMC

In 1991, Liu joined TSMC, when the company was only four years old.² He started as an R&D assistant manager and was progressively promoted to vice president of R&D, chief operating officer, and other roles, deeply involved in process technology development spanning multiple generations from 0.35-micron to 7-nanometer. During the critical years when TSMC was building its advanced-process moat, Liu was at the core of technical decision-making.

Over these three decades, TSMC's advanced processes moved from the micron era into the nanometer era. Liu participated in several of the most pivotal transitions: from the establishment of the pure-play foundry model as an independent business, to TSMC surpassing Intel and Samsung at successive process nodes. How far the technology could go was the source of his leverage when negotiating geopolitics later.

Dual Leadership Succession

In 2018, Morris Chang announced his retirement, and TSMC activated a dual-leadership structure: Liu became chairman and C.C. Wei became president.² Their division of labor: Liu was responsible for overall strategy and external relations, while Wei led operational execution.

This arrangement was interpreted at the time as a smooth transition for TSMC, but it was soon forced by geopolitical pressure to make decisions far more complex than any design had anticipated. 2018 was the year the US-China trade war fully erupted, and every critical decision TSMC made over the next six years had to simultaneously satisfy Washington, Beijing, and Taipei.

"Nobody Can Control TSMC by Force"

In 2022, as US-China tech competition reached a boiling point, TSMC became the central prize in the global semiconductor landscape. In a CNN interview, when asked about TSMC's fate if China were to attack Taiwan by force, Liu said: "Nobody can control TSMC by force."³

This statement became one of the most frequently quoted remarks of his tenure. Call it bravado — it was also a deterrent. Call it a guarantee — it also hinted at the possibility of destruction. With a single sentence, an engineer drew a geopolitical bottom line.

📝 Curator's note: The conventional narrative is that TSMC's US fab was forced out of the company by Washington pressure. But Liu's logic is closer to another reading: build fabs everywhere they are needed, and you become something no one anywhere can afford to lose — "indispensability" is the moat, not a compromise.

The Arizona Calculus

In 2020, TSMC announced a US$12 billion investment to build a wafer fab in Arizona as part of its global deployment.⁴ This figure was repeatedly revised upward in the years that followed: by March 2025, TSMC's US investment plan had expanded to over US$165 billion (approximately NT$5 trillion), covering multiple sites and the full deployment of advanced processes.⁴

In Liu's final public briefing on the Arizona plan, his framework was: the question TSMC needed to ask was "under what conditions do we go to the US," not "whether to go."

But the execution was not as clean as the strategy. In 2023, the Arizona fab's original mass production timeline was pushed back to 2025. Liu publicly acknowledged that there was a severe shortage of qualified equipment installation technicians in the US, and TSMC had to dispatch personnel from Taiwan for on-site training. This move triggered a strong backlash from Arizona construction trade unions, with worker representatives demanding that Congress block TSMC from bringing in "low-wage foreign labor." A strategic decision denominated in US dollars on TSMC's balance sheet became, on a Phoenix construction site, a real conflict over "whose skills count."⁴

Before retiring, at a shareholders' meeting, he was asked about Huawei's competitiveness. He said: "Huawei is unlikely to surpass TSMC."⁵ Clean, precise, no cracks — the engineer's answer.

"TSMC Is Not a Geopolitical Flashpoint"

At the June 2023 shareholders' meeting, Liu was asked how he viewed TSMC's role in US-China competition. He said: "Many conflicts between China and the United States — they all hope for it, and none can do without TSMC, so they will give considerable thought, even more thought. So I hope that as long as Taiwan's semiconductor industry does well, it will have a stabilizing effect on global geopolitical conflicts."⁶

He further explained: "TSMC is not a geopolitical flashpoint, but many countries know that a lack of chip supply would be a major disaster." The underlying logic was an engineer's precise judgment about "scarcity" — becoming a node that no one can afford to lose is itself the strongest moat.

In May 2024, in his last public interview before retiring, he said TSMC was no longer just "Taiwan's TSMC" but "the world's TSMC," adding: "We must make TSMC's products 'impossible to refuse' for the entire world, including mainland China."⁷

Six Years: What the Engineer Left Behind

When Liu stepped down in 2024, TSMC's market capitalization had surpassed Taiwan's GDP, making it one of the most strategically valuable manufacturing companies in the world. The TSMC he inherited was the semiconductor industry's technology champion; the TSMC he handed off was the most expensive piece on the geopolitical chessboard.

The six-year arc can be summarized by several anchor points: Arizona investment expanded from US$12 billion to US$165 billion; advanced processes advanced from 7nm to 3nm, with 2nm planning completed; the customer landscape expanded from Apple and Qualcomm to Intel. Behind each number was an occasion where he explained geopolitics in an engineer's language before government officials, the media, or shareholders.

He was not the most charismatic CEO, nor the most market-hyped one. But during the most difficult six years, he deepened TSMC's moat again and again — perhaps that is the true color of an engineer.

A person can say "nobody can control TSMC by force" within six years and, upon retirement, casually say "buy TSMC" — these two statements are not contradictory; they are different outputs of the same engineer's logic on different occasions.

"Buy TSMC": The June 2024 Handover

In June 2024, Liu formally stepped down as chairman of TSMC, with C.C. Wei taking sole command.⁸ TSMC's dual-leadership era came to an end.

On the day of his retirement, when asked if he had anything to say, he smiled and said: "Buy TSMC."⁹ TSMC's stock closed at NT$839 that day, and Liu held on to every share of TSMC he owned, with his estimated net worth exceeding NT$10 billion.

In 2025, Liu joined the board of Micron Technology, marking his first international corporate directorship after retirement, and made a substantial purchase of Micron stock worth over NT$246 million.¹⁰

The conventional narrative is that TSMC's most important years were the ones when Morris Chang built the kingdom. But another reading is that the real test of whether TSMC could survive was precisely the six years of Liu's leadership — six years when geopolitical upheaval and semiconductor supply chain restructuring collided, with no script and no precedent.

Engineer, CEO, semiconductor diplomat, and then post-TSMC observer. In 2022 he said nobody can control TSMC by force; in 2024 he said buy TSMC. The distance between these two statements is the entirety of his six-year tenure.

Further reading: TSMC Official Leadership Profiles | C.C. Wei — Current TSMC Chairman

References