Intel's CEO Pat Gelsinger has announced his retirement, effective December 1, amid the company's struggles with design and manufacturing separation and ongoing financial losses. In the interim, CFO David Zinsner and EVP of the Client Computing Group Michelle Johnston Holthaus will jointly serve as acting co-CEOs while the board seeks a permanent successor.
The Taiwanese supply chain draws parallels between Intel's current leadership transition and the 2018 scenario when CFO Bob Swan temporarily took over as CEO following Brian Krzanich's resignation. After nearly four years as CEO, Pat Gelsinger's retirement announcement has sparked speculation about underlying challenges, including potential setbacks with the 18A process or disputes over separating design and manufacturing. The uncertainty now centers on identifying a suitable successor to lead Intel.
A tenure of nearly four years marked by challenges
Pat Gelsinger, who joined Intel in 1979 and became its youngest vice president at 32, served as CTO in 2001 before leaving in 2009 after a restructuring. He returned in February 2021 as Intel's eighth CEO during a crisis period, a stark contrast to his earlier tenure when the company flourished in a thriving global PC and semiconductor market.
During his nearly four-year tenure, Pat Gelsinger openly addressed Intel's longstanding missteps that contributed to its challenges. These included misjudging market trends, missing collaboration opportunities with Apple for the iPhone, and failing to penetrate the smartphone chip market. Following his departure in 2009, Intel halted progress in AI and GPU technologies and neglected to develop a cohesive foundry strategy, further compounding its difficulties.
Upon his return, Pat Gelsinger sought to revitalize Intel with the IDM 2.0 strategy and an ambitious roadmap to deliver five nodes in four years. However, his remarks about geopolitical tensions regarding the Taiwan Strait and lobbying against US subsidies for TSMC drew significant criticism, overshadowing some of his initiatives.
Nvidia's rise poses a significant threat to Intel
As Pat Gelsinger's tenure neared its end, Intel faced mounting challenges, including US-China tensions, volatile global conditions, sluggish PC demand recovery, AMD's re-emergence, missed partnerships with Apple, and rising competition from Arm. Additionally, Nvidia's dominance in AI and TSMC's foundry leadership intensified Intel's financial woes. Despite these struggles, Intel's status as a key US corporation has led many to anticipate government intervention to ensure its stability.
The timing is critical for Intel as TSMC has solidified its leadership in foundry services, outpacing Intel in manufacturing technology and impacting Intel's product design capabilities. While Intel once dominated the PC and server markets through cutting-edge, high-cost manufacturing, TSMC's advancements in sub-7nm processes over the past decade have enabled it to surpass Intel, securing nearly all major chip contracts, especially in the AI sector.
Conversely, Intel struggles with delays in manufacturing processes and challenges in re-establishing itself in the foundry market. To stay competitive, it has grown reliant on TSMC for advanced production under 7nm nodes, with its latest Lunar Lake and Arrow Lake CPU Tile cores manufactured using TSMC's N3B process.
Meanwhile, Nvidia has risen to prominence with its AI GPUs, securing a dominant position in the AI sector and outpacing Intel despite Intel's two-decade history in AI research. The US government's push for domestic manufacturing and revitalizing semiconductor production could also contribute to Intel's challenges, potentially accelerating its decline.
Five crises Intel faces
Intel's crises are far from being resolved with Gelsinger's departure. At least five pressing issues still loom.
Firstly, with Gelsinger's departure, the question arises: Who in Silicon Valley is willing and capable of stepping into this challenging role? Following Brian Krzanich's resignation in 2018, it took Intel until early 2021 to find a suitable successor in Gelsinger. At present, there appears to be no clear figure in Silicon Valley who can steer Intel out of its current difficulties, though former Intel President Renee James might have been a potential candidate.
Recently, rumors have circulated within the semiconductor industry that Intel has inquired about the willingness of former TSMC co-CEO Mark Liu, who retired in June 2024, to succeed Gelsinger. However, non-compete clauses could pose significant hurdles, especially as Liu recently announced collaborations with his alma mater to establish a think tank.
Secondly, organizational instability could impede Intel's ability to roll out new products, processes, and technological initiatives, potentially damaging its longstanding partnerships within the Taiwanese supply chain. This disruption may weaken Intel's strategic position and ability to compete effectively in the semiconductor market.
Thirdly, the 18A process may fall short of meeting Gelsinger's promises of full-scale production, which could lead to Intel continuing its reliance on TSMC for next-generation PC platforms. This might even extend to outsourcing server platform processors for the first time, further impacting Intel's independence and competitiveness.
Fourthly, after securing US government subsidies, how can Intel revive its foundry operations to satisfy the new Trump Administration?
Fifthly, the contentious spin-off plan either risks stagnation or fragmentation, with a tumultuous future ahead for the company.
TSMC founder's discussion on relationship with Intel in autobiography
On another note, TSMC founder Morris Chang elaborated on the longstanding relationship between TSMC and Intel in the second volume of his newly published autobiography, highlighting the reasons behind Intel's decline.
Chang noted that Intel has been TSMC's longest-standing major customer, sharing a historical connection with him. Initially, when Chang sought funding for TSMC, he approached Intel but was turned down. However, within a year of TSMC's establishment, Intel's president visited the company, leading to close interactions and orders beginning to flow from Intel to TSMC starting in 1990.
Chang noted that in the 1990s, Intel took pride in its fabs and production technologies. However, Intel turned to TSMC for foundry services because TSMC's "nodes" advanced every 18 to 24 months, which encouraged Intel to allocate all its resources—factories, equipment, and talent—towards developing cutting-edge products while letting TSMC handle the production of older technologies.
From Intel's perspective, partnering with TSMC offered significant advantages. Intel was particularly impressed by TSMC's ability to use the same technology to manufacture different products for multiple customers, showcasing a level of flexibility that Intel found remarkable.
This outcome is merely a result of TSMC's business model. Intel's approach focused on designing and manufacturing high-tech products, emphasizing maximum efficiency at every stage. A key example of Intel's strict discipline is its requirement for each factory to operate identically to others (a strategy known as "copy exactly"). In contrast, TSMC's business model offers specialized foundry services to multiple clients, allowing for greater flexibility and the ability to produce various products using the same technology.
After 2010, Intel showed interest in entering the foundry business, but its established practices as an IDM made achieving success difficult. In early 2021, Intel appointed a new CEO intending to expand into foundry services yet it also increased its orders from TSMC. Reflecting on this duality of Intel—being both a client and a competitor to TSMC—Chang wrote in his book, "I wish them good luck!"
