As Intel faces what it calls the most significant crisis in its 56-year history, Samuel J. Palmisano, former CEO of IBM, has some hard truths for the Silicon Valley giant. With Intel exploring options like divesting assets, selling shares, and even.....
As Intel faces what it calls the most significant crisis in its 56-year history, Samuel J. Palmisano, former CEO of IBM, has some hard truths for the Silicon Valley giant. With Intel exploring options like divesting assets, selling shares, and even breaking up parts of the company, Palmisano suggests Intel might be better served by forging deep alliances with leading semiconductor foundries such as Samsung or Taiwan's TSMC.
Intel has already hinted at expanding its partnerships with both TSMC and Samsung, having outsourced some of its chip production to TSMC. Insiders suggest that Intel's product division, which is under pressure to keep pace with industry trends, is more inclined to work with external foundries than continue relying on its own manufacturing plants.
Palmisano, speaking in an interview with Bloomberg TV, argued that Intel needs to rethink its IDM 2.0 strategy, which integrates both chip design and manufacturing. His perspective marks one of the first prominent public critiques of Intel's current direction from a tech industry veteran.
While industry analysts have long debated the viability of Intel's vertically integrated model, Palmisano's view is rooted in experience—he led IBM through its own difficult restructuring, including the controversial decision to sell its personal computer business to China's Lenovo in 2005.
Though Intel's crisis is well-publicized, some may question whether IBM—fresh off massive layoffs in its R&D department in China—is in a position to offer advice on navigating the chip industry's upheaval. Yet, while IBM may not be top of mind when it comes to semiconductors, its history in the sector is substantial. IBM has long operated semiconductor fabs and developed cutting-edge processes, but always in a B2B model, unlike Intel's consumer-facing B2C approach with its CPU products.
Few in the tech world forget IBM's impact on semiconductors. Take Lisa Su, CEO of AMD, a rising star in the AI-driven chip race. Su spent 12 years at IBM, where she held key roles in the company's Semiconductor Research Center and later became a critical figure in the company's emerging product strategies. It was her IBM background that laid the groundwork for her turnaround at AMD.
While media outlets like Reuters and Bloomberg continue to report on Intel's restructuring, Bloomberg's interview with Palmisano was likely driven by his history of making bold moves. In addition to selling IBM's PC division, Palmisano also presided over the company's decision in 2015 to offload its wafer manufacturing business to GlobalFoundries. The deal wasn't a typical sale—IBM not only provided 16,000 patents but also handed over US$1.5 billion in cash to sweeten the deal for GlobalFoundries.
For Palmisano, tough decisions like selling core assets or divesting in legacy businesses are just part of how companies evolve with changing industry trends. IBM, a century-old company, has had to pivot many times to stay relevant, and Palmisano doesn't see why Intel should shy away from doing the same, whether it involves splitting off fabs or selling legacy divisions like PCs.
Palmisano's suggestion that Intel deepen its partnerships with foundries like Samsung or TSMC draws on IBM's own experience. When IBM sold its fabs to GlobalFoundries, the initial agreement was for GF to manufacture server chips for IBM.
However, after GlobalFoundries indefinitely shelved its 7nm process development, IBM found itself without a key manufacturing partner. This led IBM to strike a deal with Samsung, which now produces its advanced chips. From IBM's perspective, the most important factor wasn't controlling the fabs but ensuring the performance and quality of its server chips remained top-tier.
For Intel, the question remains: will its future chips be marked by "Samsung Inside" or "TSMC Inside"? According to Palmisano, it's a possibility Intel needs to seriously consider if it wants to navigate the current semiconductor landscape effectively.
Intel has long been a staple of the semiconductor industry, but its challenges are mounting. Starting in 2024, the company will split its manufacturing and product divisions into separate profit-and-loss entities. While this doesn't necessarily mean an immediate breakup, Intel is under increasing pressure to streamline operations and become more competitive. It's a move that echoes some of IBM's restructuring under Palmisano—though Intel's financial situation means it may not yet have the luxury to operate its divisions independently.
Palmisano's blunt advice for Intel might sound like tough love, but given his track record of steering IBM through its own crises, it's advice Intel would do well to consider. After all, IBM's willingness to make hard choices—divesting businesses, outsourcing production, and redefining its strategy—has helped it endure as a tech giant for over a century. For Intel, the time may have come to decide what it's willing to hold onto, and what it can afford to let go.