While it remains somewhat premature to conclusively assess the success of Intel’s 18A process technology (1.8nm-class), clarity regarding the success of Pat Gelsinger’s ambitious plan to launch five new technology nodes in four years is expected within the next year. However, Intel’s current inability to compete in the AI processor market is already apparent, providing Morris Chang, the founder of TSMC, with ample grounds to argue that Intel might have been better served by concentrating on AI processor development rather than pushing the boundaries of process technology. Chang has openly criticized the lack of both a clear strategy and a CEO at Intel, noting the challenges these gaps will present.
“I am not sure why Pat stepped down,” Chang mentioned at a press event for his biography launch, as per Reuters. “It’s unclear whether his strategy was flawed or poorly implemented.” […] “He appeared to prioritize transforming into a foundry over focusing on AI, which, in retrospect, may have been a misstep.”
Chang’s critique holds weight considering Nvidia’s annual revenue from AI processors amounts to tens of billions of dollars, far surpassing what TSMC makes from manufacturing those chips. A strong AI processor line could potentially have been more lucrative for Intel than its service as a chip foundry, underscoring the missed opportunity in not prioritizing AI technology earlier.
Upon assuming the role of CEO in 2021, Gelsinger introduced his IDM 2.0 strategy, which included the development of five cutting-edge process technologies over four years and aimed to position Intel as the world’s second-largest contract chip manufacturer by 2030. Despite this focus, the market has yet to witness significant new product launches under Gelsinger’s leadership.
Just last month, Intel reported that it would fall short of its $500 million revenue target for its Gaudi 3 AI accelerators due to software complications. Considering the typical three-plus years required to design, finalize, and mass-produce a chip, Gaudi 3’s development predates Gelsinger’s tenure, absolving him of responsibility for the hardware. Yet, the apparent lack of emphasis on AI software development is quite clear.
The recent publication of Chang’s biography offers additional insight into his viewpoints. The memoir covers his career from 1964 to 2018 and highlights TSMC’s significant collaborations with giants like Apple and Qualcomm. It also revisits Intel’s 1980s decision to pass on an investment opportunity in TSMC, a decision that, in hindsight, represents another missed strategic opportunity.
Today, Intel’s future appears fraught with uncertainty. With no clear successor to Gelsinger and no definitive new strategic direction announced, Intel seems poised to continue its focus on developing competitive products and maintaining its manufacturing capabilities, even as it faces persistent challenges. Although Gelsinger has departed, the Board of Directors has not yet confirmed whether a new AI-focused strategy has been established.
“They are currently without a new strategy or a new CEO,” Chang reiterated. “Securing both will be a formidable challenge.”
