In the wake of Intel's profound financial difficulties, a quartet of past Intel board members has suggested a drastic measure in a Fortune magazine piece. They propose that the U.S. government leverage its nearly $20 billion in grants, loans, and promised federal funding to compel Intel to bifurcate into two distinct entities, fully separating its manufacturing and design operations. They argue this is the sole path to Intel's survival.
On the contrary, Craig Barrett, Intel's former CEO, has expressed his disagreement in his own Fortune article, suggesting that such a split would be detrimental to Intel and could undermine the United States' objective of maintaining semiconductor leadership.
Former Intel Board Members Urge Immediate Intel Split
While Intel is implementing its IDM 2.0 strategy, aiming to excel as both a chip designer and manufacturer, it has reported losses of $5.87 billion and a significant decrease in market value over the past 2.5 years. The company is now implementing cost-cutting measures, including laying off 15,000 employees globally, pausing construction of its wafer fab in Germany and packaging plant in Ireland, and selling two-thirds of its worldwide real estate assets. Despite substantial investments in its 18A process, the long-term effectiveness of these efforts remains to be seen.
Yet, David Yoffie, Reed Hundt, Charlene Barshefsky, and Jim Plummer—four past Intel board members—insist that Intel's survival hinges on splitting into two separate entities.
Rationale for Splitting Intel
They contend that Intel's fabrication facilities, engineers, and intellectual property are critical to the national security and technological advancement of the United States. As the only U.S. company capable of mass-producing advanced logic chips, Intel's decline could render America overly reliant on TSMC, largely based in Taiwan, China, thereby presenting considerable geopolitical and supply chain risks.
Former TSMC executives have underscored the necessity for America to maintain a robust semiconductor capability to stay competitive in AI and defense technology. However, Intel has lost its competitiveness against TSMC in leading-edge process technologies.
Moreover, companies like Broadcom, NVIDIA, and Qualcomm, while seeking alternatives to TSMC, hesitate to use Intel Foundry services due to their rivalry with Intel. Similarly, Samsung Foundry faces similar hurdles as major chip designers like Apple and NVIDIA avoid it due to competitive concerns.
Although Intel's product division remains competitive, its Foundry services have been losing money. Current Intel management still needs to prove its ability to operate a successful foundry business amid criticism over missed deadlines and targets. These former directors caution that continuing along the current trajectory could lead to Intel's bankruptcy, a scenario disastrous for both the national economy and security.
Proposed Split Model
The former directors urge Intel to separate its foundry from its design business, akin to AMD's creation of GlobalFoundries in 2009. This would entail Intel's product design division entering long-term supply agreements with newly independent foundries. They argue that such an arrangement would stabilize the foundry's operations until it secures third-party orders.
The U.S. government, they assert, should play a pivotal role. The Chip Act has allocated $39 billion to boost U.S. semiconductor production, with potential support for Intel totaling $8.5 billion in grants and $11.5 billion in loans.
However, these former directors express concern that Intel's mismanagement could echo the failure of Solyndra, a solar company that went bankrupt despite substantial government funding. To prevent this, they urge the government to ensure Intel fully divests its manufacturing and design divisions.
This move, they argue, would not only enhance Intel Foundry's competitiveness but would also offer companies from the U.S., South Korea, Japan, and Europe access to a crucial secondary source of advanced chips, bolstering the global supply chain's resilience.
The former directors emphasize the need for urgency. Intel's rapid decline in the semiconductor market could worsen if corrective measures are delayed, potentially leaving the U.S. lagging further behind as TSMC advances. Both Intel and the government need to act swiftly to secure the future of American semiconductor manufacturing.
Challenges of the Split and Potential Issues
The directors present a clear vision for Intel's manufacturing spinoff, drawing parallels with AMD's 2008-2009 restructuring of its foundry operations. However, this comparison reveals challenges.
GlobalFoundries, primarily manufacturing for AMD and other clients like Broadcom, NXP, and Qualcomm, struggled with profitability before its IPO, which coincided with a semiconductor demand surge. It halted advanced process node developments below 10nm to curtail losses.
AMD's split, some argue, provided limited help to GlobalFoundries, which continued to underperform. Between 2009 and 2021, Mubadala, the majority shareholder, incurred over $22.4 billion in losses on GlobalFoundries.
Given this precedent, it is difficult to envision investors willing to take over Intel's expansive manufacturing business. Intel operates more fabs and projects than AMD or GlobalFoundries, and if it were to become truly independent, staggering losses could ensue, especially if its design division shifts chip production to TSMC.
Even assuming exclusive supply agreements prevent this, not maintaining a technological edge could render Intel's product division noncompetitive with TSMC-made products from AMD, Nvidia, and Qualcomm. This situation could lead to broader unprofitability, with losses potentially attributed to Intel, which would likely hold a significant ownership stake in any independent foundry company.
Former Intel CEO's Counterargument
Craig Barrett, former Intel CEO, disputes the necessity of the split. He argues that dismantling Intel would weaken the U.S.'s semiconductor ambitions. He states that robust investment is crucial for competitive chipmakers, with only three globally—Intel, Samsung, and TSMC—having sufficient revenue to sustain advanced process R&D for future products.
Barrett warns that while Intel's design division might endure independently, the foundry's survival remains uncertain. Comparing it to AMD's and GlobalFoundries' 2008 split, he notes that despite AMD's success with Ryzen CPUs (manufactured by TSMC), GlobalFoundries has lagged behind TSMC due to a lack of distinct leadership technologies.
Limited resources forced GlobalFoundries to abandon advanced process R&D, reducing its competitiveness. Prospective customers tend to favor foundries with more advanced technologies, limiting production and revenue—which further constrains R&D investment at GlobalFoundries.
Barrett posits that a similar fate could befall Intel if its manufacturing arm is spun off. This potential reliance on TSMC or Samsung could endanger America's chip production autonomy, even as they're encouraged to establish U.S.-based factories. Additionally, these companies, headquartered in nations under external threat, could become focal points for conflict.
Barrett advocates for a focus on technological leadership rather than division, emphasizing performance to reestablish Intel's global dominance in chip manufacturing. He commends Pat Gelsinger, current Intel CEO, for steering Intel correctly, notably with the announced Clearwater Forest Xeon chip built on Intel's 18A node.
He further emphasizes that the U.S. must maintain its lead in the semiconductor sector. Despite investing more this past year than in the last 28 combined, he suggests further emphasis, particularly on academic research. While the National Semiconductor Technology Center (NSTC) establishment marks a positive step, Barrett points out its five-year budget is dwarfed by Intel's annual R&D expenses.
Finally, Barrett reminisces on Intel's resilience after the Internet bubble burst, leading to a robust comeback, suggesting a similar approach now would be invaluable for Intel's future.
