Intel: Spinning Off and Over IDM
By: Justin Voronoff & Roy Katznelson
The Ivey Business Review is a student publication conceived, designed and managed by Honors Business Administration students at the Ivey Business School.
Sil-Icon of Silicon Valley
Intel was incorporated in 1968 by Robert Noyce and Gordon Moore, and with it, began the rapid innovation that characterizes “Silicon” Valley. Intel’s business model revolved around designing and manufacturing microchips to be sold to computer companies such as Dell, who packaged these chips into their own computers. The integration of these two steps, design and manufacturing, within one company made Intel an Integrated Device Manufacturer (IDM).
Selling chips as an IDM was the predominant business strategy until the 1980s, when competitors began to focus on just one of these two supply chain steps. Those specializing in chip manufacturing became known as foundries or “fabs”, and those that focused exclusively on design became known as fabless.
As chips became more complex over time, continuous capital investment in manufacturing processes became a steep barrier to entry. Today, the upfront capital expenditures associated with opening a new fab is estimated to be a minimum of $10 billion in upfront capital. In addition, the industry became characterized by long lead times of up to four months. Therefore, a highly agile supply chain is required in the industry. Manufacturing processes need to constantly keep up with the breakneck speed of chip innovation, changing manufacturing needs, and high demand uncertainty.
Chipping Away At The Competition
The world’s first independent foundry, Taiwan Semiconductor Manufacturing Company (TSMC), has gone on to become the dominant player in the global semiconductor manufacturing market. TSMC’s strategy of investing heavily in research and development to lead the industry in engineering innovation has contributed to its rapidly increasing market share, particularly among high-end chip manufacturing. This strategy has earned TSMC a commanding 54 percent of the semiconductor manufacturing market and by some estimations, a 2 to 3 year lead in technology innovation.
In the early 2000s, Advanced Micro Devices (AMD) recognized that dividing its business into two units — one design-focused, one manufacturing-focused, would be strategically beneficial. AMD’s foundry business was spun off as a separate entity named GlobalFoundries (GF).
AMD, the former design unit, shifted its focus to innovating chip design while outsourcing their manufacturing to GF and TSMC. This strategy proved fruitful as AMD could innovate freely, delivering a powerful 10nm-equivalent chip to the market ahead of Intel. They have now shifted their operations to making chips for next-generation technologies like 5G. These initiatives have unlocked a $20 to $30 billion market for AMD while keeping fabless product development costs low.
Following the spin off, GF maintained operational efficiency and expanded its chip manufacturing business to six factories based in the United States, Germany, and Singapore. These developments have allowed GF to become the third largest global semiconductor manufacturer with about a 7 percent share of the $86 billion semiconductor manufacturing market today.
COVID-19 Pandemic: A Symptomatic Supply Chain
A sudden increase in semiconductor demand, catalyzed by the COVID-19 pandemic, has created a worldwide shortage of chips that has revealed hidden risks in the global supply chain. The automotive industry has been adversely impacted by this shortage, with companies such as Toyota and Ford forced to scale back vehicle features or temporarily cease production.
Governments have started to intervene to increase global supply. For example, South Korea is pledging to invest $452 billion into domestic chip manufacturers. The U.S. is also taking action by investing $50 billion directly into the domestic supply of semiconductors.
With various projections for global semiconductor sales set to double from approximately $450 billion in 2019 to $1 trillion by 2030, major foundries are trying to capitalize on this opportunity to gain market share. TSMC, Samsung, and GF have pledged $100 billion, $151 billion, and $4 billion respectively to increase global capacity over the coming years.
Manufacturing Consolidation: A Comparative Disadvantage
High labour inputs for semiconductor manufacturing pushed stakeholders to outsource to developing economies, leading to the consolidation of manufacturing in Asia — where 60 percent of chip manufacturing occurs today. Conversely, 55 percent of fabless chip operations remain concentrated in the United States, due to the need for highly-skilled engineering labour for microchip design. The pandemic uncovered hidden costs of concentrating high-capital expenditure operations in one geographic region, as supply chains failed to adapt to changing demand forecasts. Semiconductor consumers such as automobile companies quickly became exposed to the limited suite of manufacturing options for highly technical chips outside of Asia, particularly with TSMC, leading to product delays and increased costs. Intel’s strategy of continuing to operate as an IDM and selling in-house designed chips makes them unfit to capitalize on the excess demand coming from spaces such as the automotive industry.
Navigating Chippy Waters
Intel’s struggle has been evident in recent years — in 2020, they lost Apple as a client to TSMC due to stagnant innovation after being their primary chip supplier for 14 years. To overcome its struggles and capitalize on the current market conditions, Intel overhauled its management team. Newly-hired CEO Patrick Gelsinger’s “IDM 2.0” strategy pushed Intel to design and manufacture its own chips while introducing a new business unit with foundry services for third-party fabless customers.
This ambitious strategy will see Intel increase its use of external foundries, such as TSMC, to manufacture some of its chips. To quickly enter the foundry space, Intel explored a deal to acquire GF, for around $30 billion in July 2021. However, the deal never gained traction due to GF’s concerns that the combination would upset key customers such as its former parent AMD, Intel’s largest competitor.
In 2002, IBM proposed an IDM strategy similar to Intel’s IDM 2.0, offering foundry services to clients such as Sony and AMD while also manufacturing chips for its own high-end servers group. However, customers began switching suppliers because they worried that IBM would keep cutting-edge technologies for its own servers, only sharing them once the technology became mainstream. Ultimately, the strategy failed in 2014 as IBM gave GF $1.5 billion to take control of its foundry unit in exchange for manufacturing services of its chips for the next 10 years.
Like IBM, Intel’s IDM 2.0 strategy may target and serve Intel’s direct competitors. This could lead to the same poor customer relationships or limit the foundry division’s ability to target a significant share of the market. Intel’s strategy also involves the increased use of external foundries which could create internal competition with Intel’s own foundry services.
An Intel-ligent Solution
Intel should take advantage of the current market conditions by repositioning its business model. To overcome increasing global competition and avoid IBM’s foundry mistakes, Intel should spin off its manufacturing division into an independent company.
Spinning off Intel’s foundry division would strengthen Intel’s manufacturing arm by allowing the new unit to focus on key success factors unique to the manufacturing space. These factors include prioritizing winning customer contracts to expand their weakening market share, and directing capital allocation towards manufacturing capabilities and increasing capacity.
By focusing on winning new customers and outsourcing chip design, Intel’s foundries can access fabless customer innovations and build their manufacturing capabilities around these needs. This would position the foundry unit to meet trending product innovations from a manufacturing standpoint, while increasing the likelihood of winning incremental customers who are creating bleeding edge technology. This contrasts with the current IDM model, which is solely reliant on Intel’s internal design unit for manufacturing innovation, and instead follows TSMC’s successes of crowdsourcing innovations from its wide range of customers as an independent foundry.
With this strategy, Intel’s foundries will have better access to M&A opportunities with less strategic and financial constraints. Conflicts of interest such as those arising from the GF merger would be eliminated, allowing Intel’s foundries to strategically align with acquisition targets in the market during dealmaking. There are over 190 large foundries outside of Asia that can be targeted for acquisition across a range of chip sizes, which would increase the geographic diversity of the foundry.
From a financial standpoint, raising capital to make these acquisitions can be streamlined as it will be explicitly sourced for foundry M&A needs. This will align shareholders in the foundry unit around the manufacturing function, which solves Intel’s current issue of budgeting across two units with different financial profiles when raising capital. This issue was evidenced by Intel’s recent one-day 10-percent stock price drop on October 22, 2021 as investors became concerned that Intel’s capital expenditure into the foundry business could significantly decrease overall company profitability in the coming years.
Focusing investments into solely developing leading-edge chip designs will allow Intel to better innovate in the future. The spin off will also give the design teams more flexibility in choosing manufacturing processes for their chips. Intel can use the foundry spin off for production of their current products and others in the low-complexity chip market, and increase its outsourcing to TSMC for innovative products supplying higher-end markets.
Since the foundry spin off meets the U.S.’s goal of increasing domestic chip supply, it can apply for subsidies to speed up foundry construction. Intel can also lobby for larger separate subsidies to fund design group research. The market appetite for a potential Intel foundry spin off would be high. Skeptics may look no further than GF’s IPO on October 28 at $47, as it raised about $2.6 billion and saw a 37-percent increase in its share price following its first week of trading. (Exhibit 5)
Disconnecting Chips
Intel has various options for how to structure the spin off transaction including spinning off either unit (fab or fabless) and involving either the public markets or private financial sponsors in the deal. Intel should shift some of its internal leadership to exclusively manage the foundry unit before offering a majority share of the new company to the public. The fabless unit should maintain a minority share in the new entity for at least the first stage of the foundry’s transition to an independent entity, as it builds a customer rapport among fabless customers and establishes its unique capital expenditure needs. This would allow Intel’s fabless division to partake in the new entity’s upside without directly impeding on the unit’s strategic focus on foundry operations.
Choosing to go public over finding a financial sponsor would be favourable with public semiconductor companies driving the NASDAQ to record highs in late 2021. A strong public share price would enable the foundry entity to begin seeking out acquisition targets, using public shares as potential acquisition currency to strengthen its market position.
Intel, once the leading semiconductor company across design and manufacturing, has lagged competition in both their design and manufacturing functions. Competitor case studies and strategic analyses have made it clear that long term success in the semiconductor supply industry lies outside of the IDM model. The undersupply of chips worldwide and government policy interventions offer a compelling opportunity for Intel to improve both their division’s long term market positions through a spin off transaction.