Lisa Su

On October 8, 2014, AMD announced that Lisa Su would become the company's president and CEO¹. The news landed with little fanfare in the industry — AMD was then worth roughly $3 billion, its stock barely clearing $3 a share, and a quarter of its workforce had already been let go². The company had sold its own Austin campus to raise cash and was leasing it back so engineers had somewhere to sit³. Server market share stood at 2 percent³. Observers inside and outside the semiconductor world had largely written AMD off as an also-ran. Su, forty-four years old, an electrical engineer by training and temperament, sat down with her communications chief Drew Prairie on day one to write a letter to employees³. The three-point plan she put on paper — great products, deep customer trust, simplified operations — looked modest against the scale of what was broken. It would take five years to know whether she had read the situation right.

Su arrived in the United States from Tainan, Taiwan at three years old, growing up in Queens, New York. Her father, a mathematician who had come to New York for graduate school, quizzed her on multiplication tables at the dining room table when she was seven². At ten she began dismantling her brother's remote-control cars¹ — not to break them but, as she would later recall, out of 'a great curiosity about how things worked.' She attended the Bronx High School of Science, graduated in 1986⁴, and enrolled at MIT that fall. During her undergraduate years, a stint as a research assistant introduced her to semiconductor physics and fixed her direction. She earned her bachelor's, master's, and doctoral degrees in electrical engineering at MIT, finishing her PhD by her mid-twenties. The discipline would define not just her career but her way of seeing problems: decompose, test, rebuild.

In 2000, IBM assigned Su to a year-long post as technical assistant to CEO Lou Gerstner¹ — a position that put her in the room for decisions that most engineers never see. Afterward, she founded IBM's Emerging Products division, describing herself at the time as 'basically director of myself — there was no one else in the group.' She hired ten people to work on low-power semiconductors and biochips. The division then joined a collaboration with Sony and Toshiba to build a next-generation game processor; the project's goal, set by Sony's Ken Kutaragi, was to improve game machine performance by a factor of 1,000¹. Su's team proposed a nine-processor design that became the Cell microprocessor, the chip that powered the PlayStation 3⁵. The experience taught her what a small team with a focused mandate could accomplish when the organization got out of its own way — a lesson she would apply at a much larger scale fourteen years later.

In the winter of 2014, a few weeks into her tenure as AMD's CEO, Su convened a series of town hall meetings at company facilities across the country, listening before she dictated⁶. The picture that emerged was of a company that had fragmented its bets across too many product lines and lost the engineering confidence that once let it challenge Intel directly. She stripped the strategy down: gaming, data centers, high-performance computing. Then came the deeper wager. Meeting with AMD's chief architects, Su decided to redesign the company's CPU architecture from scratch, a project that became known as Zen, and simultaneously commit to chiplet design — splitting processors into multiple smaller dies rather than a single monolithic chip. 'Is this the time we're going to bet the company on going to chiplets?'⁷ she recalled asking her team. The answer was yes. Ryzen, the first Zen-based processor line for consumers, launched in 2017⁸. AMD's server chip market share at that moment was less than 1 percent. By the second quarter of 2022, it had crossed 22 percent⁹. The Austin campus AMD had sold to survive was, by then, merely a detail in a much larger story.

By 2022, AMD had surpassed Intel in both market value and annual revenue for the first time in the company's history². The market capitalization that stood at roughly $3 billion when Su took over had grown to more than $200 billion⁵. Time named her CEO of the Year for 2024 — the first woman to receive the designation twice¹. MIT named its new nanotechnology research center the Lisa T. Su Building¹⁰. In the semiconductor industry she built her career in, Su remains most associated with a specific decision made in a conference room in Austin: the choice to design a new architecture and break AMD's chips into smaller, more flexible pieces when the company had neither the money nor the margin for error to be wrong. AMD's share of the desktop CPU market grew from 18 percent in 2017 to 36 percent by the end of 2023, while Intel's fell from 82 to 61 percent¹¹. The building in Cambridge now bears her name, and the chiplet design she bet on has since become the industry's dominant approach to scaling processor performance.