TSMC Founder Claims Intel Lacks Clear Strategy and Effective Leadership

While it’s premature to definitively assess the success of Intel’s 18A process technology—a 1.8nm-class—it will be clear in about a year whether Pat Gelsinger’s ambitious plan of launching five nodes in four years has paid off. However, Intel’s apparent shortfall in delivering a competitive AI processor is already conspicuous today. This has led Morris Chang, the founder of TSMC, to suggest that Intel might have been better served by concentrating on AI chip development rather than pushing the boundaries of chip manufacturing technology. Chang also mentioned that Intel is currently lacking in both strategy and leadership, with filling these gaps posing a significant challenge.

At a press event launching his biography, Chang expressed his uncertainty about Gelsinger’s departure from the company, as reported by Reuters. “I don’t know why Pat resigned,” Chang said. “It’s unclear whether his strategy was flawed or poorly implemented.” He further noted, “It appeared he was more interested in Intel becoming a foundry, though in hindsight, focusing on AI would have been better.”

Chang’s critique seems justified when considering the financial success of Nvidia, which makes billions annually from its AI processors—a sum that far exceeds what TSMC earns from manufacturing those chips. Had Intel developed competitive AI processors, they potentially could have earned more than what they make from their foundry services. Therefore, prioritizing AI should have been a strategic move made years earlier.

Upon assuming leadership in 2021, Gelsinger’s primary strategy, dubbed IDM 2.0, aimed to develop five advanced process technologies over four years and to position Intel as the world’s second-largest contract chipmaker by 2030. Although Intel had a clear product strategy, the market has yet to see significant products emerge under Gelsinger’s tenure.

Recently, Intel announced a shortfall in reaching its $500 million sales goal for the Gaudi 3 AI accelerators this year, attributing the miss to software setbacks. Typically, the development, deployment, and mass production of a chip spans more than three years. Since Gaudi 3’s initial development started before Gelsinger’s time as CEO, it’s unfair to place the entire blame on him for this hardware. However, a lack of sufficient focus on AI software is apparent.

The recent release of Chang’s biography offers deeper insights into his thoughts. Covering his career from 1964 to 2018, the memoir highlights TSMC’s critical partnerships with giants like Apple and Qualcomm and recalls Intel’s missed opportunity in the 1980s to invest in TSMC—a decision that, in retrospect, was a strategic error.

Currently, Intel’s future appears uncertain. With no clear successor to Gelsinger, the company seems to be maintaining its course—focusing on competitive product development and manufacturing. Yet challenges remain, and the Board of Directors has not confirmed whether a new AI-focused strategy has been formulated.

“They currently lack both a new strategy and a new CEO,” Chang reiterated. “Finding both will be quite challenging.”