The Attempt to Reform Intel
The chip designer and manufacturer missed multiple major technological opportunities. Its recent pivot to contract manufacturing is heavily backed by the U.S. government.
.jpg){kind=link}
Intel is the most advanced U.S. semiconductor manufacturer and the largest by both revenue and number of employees, with revenues of $54 billion in 2023 and over 130,000 employees worldwide.1 “Fabless” semiconductor companies like Nvidia or AMD design but do not manufacture chips. “Foundries” like Taiwan’s TSMC or China’s SMIC manufacture chips on contract, but do not design them. Rather, like South Korea’s Samsung, Intel is an “integrated device manufacturer,” meaning it both designs and manufactures its own computer chips, including central processing units (CPUs), graphics processing units (GPUs), and more. Despite decades of preeminence in the computer industry strengthened by Intel’s proprietary x86 instruction set architecture (ISA) and long-term partnerships with personal computer companies like Apple and Microsoft, Intel became a dead player by the 2000s. In the face of sharply declining revenues and share prices (INTC) since 2021, the company is now attempting a radical pivot to contract manufacturing in order to become a globally competitive foundry.
An instruction set architecture is a standardized protocol or layer of abstraction for feeding software commands to a CPU, enabling compatibility of software across different CPUs. For decades, Intel’s business model has hinged on its x86 instruction set architecture, which it developed and which gained wide adoption during the personal computer revolution of the 1980s through partnerships with IBM and Microsoft. Intel helped to cement this predominance by heavily committing to backward compatibility of software with its ISA, a highly desirable feature for users but which greatly increased the hardware design complexity of making x86-compatible chips, thus making it harder to compete with Intel.2 Moreover, while other ISA designers like British rival Arm have widely licensed their ISAs to other chip designers, Intel never has, instead implicitly threatening legal action for any infringements on its ISA. The result was Intel having a worldwide quasi-monopoly on designing and manufacturing CPUs for personal computers and servers.
But Intel’s central position has significantly eroded. The mobile computing revolution since the introduction of smartphones in the 2010s is virtually entirely powered by Arm’s ARM instruction set architecture, which lacks x86’s degree of backward compatibility but is far more power-efficient. Intel also missed out on the revolution in parallel computing powered by Nvidia’s GPUs. Meanwhile, as a manufacturer, Intel has been outdone by competitors in East Asia, especially Taiwan’s TSMC, which pioneered the foundry model of chip manufacturing and is now the most technically advanced chip manufacturer in the world. Reforming Intel to become a foundry is now a task that falls to 63-year-old Pat Gelsinger, an Intel veteran who left in 2009 before returning in 2021 to become CEO.
Become a paid subscriber and get a new in-depth investigation of a key live player, institution, or industry in your inbox every Wednesday:
Intel’s Products Are Becoming Outmoded
.png){kind=link}
Intel categorizes its products into three overarching categories, roughly corresponding to chips for personal computers, for servers and data centers, and for network infrastructure. In the second quarter of 2024, these three categories brought in the vast majority of Intel’s revenue, 63% of which came from personal computers, 25% from servers and data centers, and 11% from network infrastructure.3 Over half of Intel’s revenue is from East Asian markets, with China making up 27% of total revenues in 2023.4 This is due to the majority of its products being chips used in personal computers and servers, which are overwhelmingly built and assembled in the region.5 Due to Intel’s reliance on selling chips for personal computers and servers, 40% of the company’s revenues is tied to just three computer manufacturers: U.S. companies Dell and HP, and Chinese manufacturer Lenovo.
The key to Intel’s core line of business is designing and manufacturing x86-based central processing units (CPUs), the central chips in any computer that execute instructions from human-written software. Intel would have a total monopoly on x86 chip design if not for the fact that, when Intel agreed with IBM to supply x86 processors for the first personal computers in 1981, it was stipulated that there should be a second supplier of such processors. The U.S. company AMD was the chosen second supplier and, despite vigorous legal action from Intel running to 1995, has maintained the right to design x86 chips. Unlike Intel, AMD is fabless and outsources manufacturing to foundries. Today, Intel has about 80% of the x86 processor market, while AMD has 20%.6 Notably, since 2019, AMD has somewhat increased its market share of x86 processors against Intel after over a decade of decline, indicating that even in its core business area Intel is facing serious competition.7
But Intel’s bigger problem is that it failed to take advantage of, let alone lead, the multiple fundamental transformations of the computing industry during the 21st century which are now obsoleting Intel’s technology. For example, in the 2000s Intel turned down an offer from Steve Jobs to manufacture processors for the then-brand-new iPhone, because Intel’s then-CEO Paul Otellini believed Apple would not sell enough phones to justify the initial development cost.8 Therefore, the initial iPhone processors were designed and manufactured instead by Samsung, eventually costing Intel untold billions in lost revenue.
Intel failed to later catch up in the smartphone revolution, in part because of its commitment to its x86 instruction set architecture, which was strictly technically inferior to rival Arm’s ARM architecture on mobile devices. Whereas x86 is a “complex instruction set computer” (CISC) where single instructions can execute multiple steps in a row, ARM is a “reduced instruction set computer” (RISC) where instructions are simplified in order to run faster and more power-efficiently, thus preserving operating time on a battery-powered device like a smartphone. ARM 9, the latest version, has 354 base instructions.9 In contrast, x86 has over 1500.10 Hardware constraints in the earlier and less advanced days of semiconductor manufacturing and design often necessitated more divergent CPU core and component designs to achieve different goals, which in turn necessitated more bespoke instructions in the ISA, that were then preserved by Intel and Microsoft’s commitments to backward compatibility, at the cost of greatly increasing the complexity of designing new x86-compatible CPUs.11
This dynamic served to entrench Intel’s leading position until the advent of the smartphone multiplied the number of valuable, battery-powered consumer devices demanding both high performance and efficiency that was only available with ARM-based chips. Unlike Intel, Arm always widely licensed the right to design chips with its ISA and, until 2006, Intel even owned XScale, a designer of ARM-based chips. But Intel doubled down on x86, developing an unsuccessful low-power mobile variant called Atom chips, and sold Xscale to semiconductor firm Marvell Technology.12 Today, Arm has an estimated 99% market share in mobile computing.13
Losing out on the smartphone market was ignored in large part because Intel’s x86 processors saw massive growth in demand from data centers built by cloud computing companies like Amazon Web Services (AWS), but fifteen years later these too are now looking to transition to custom ARM-based chips.14 So are personal computer companies. Between 2020 and 2023, Apple fully transitioned off of x86 for its personal computers, switching to ARM. Even Microsoft is now switching to ARM for its upcoming line of “AI [personal computers].”15
Another significant error was Intel’s failure to build a discrete graphics processing unit (GPU) for high-performance computing early on. From 2008 to 2010, Project Larrabee was Intel’s attempt to make a discrete GPU, but it was shelved due to delays. Though Intel would eventually release dedicated GPUs, named Arc GPUs, in 2020, it lost the vast majority of the market to Nvidia, who pioneered the technology.16 Nvidia, as of 2024, has a 94% market share in GPUs being used in data centers for demanding AI workloads, with AMD having 4% and Intel under 2%.17 Current Intel CEO Pat Gelsinger has argued that the cancellation of Larrabee made redundant many of the acquisitions Intel later made in AI-related companies, as they all relied on Nvidia hardware for training their models.18
Intel also lost its technical lead in manufacturing the most advanced chips. As of mid-2024, the most advanced manufacturing process or “process node” for Intel is 7 nm, while both TSMC and Samsung are producing 3 nm chips.19 The failure was in part attributable to Intel deciding not to buy the Dutch semiconductor equipment manufacturer ASML’s first extreme ultraviolet lithography (EUV) machines in 2014, despite being a large investor in the company at the time.20 The decision was made by then-CEO Brian Krzanich, who believed it would not be economically justifiable.21 In 2018, Intel attempted to leapfrog TSMC and Samsung with a new and more advanced manufacturing process, but ended up failing and further delaying its planned chip releases.22
The two decades of strategic errors and failures on Intel’s part are the result of succession failure. Intel was originally founded in 1968 by luminaries like Robert Noyce—inventor of the silicon microchip—and Gordon Moore—who coined the famous “Moore’s Law”23—with alumni of the first major semiconductor company Fairchild Semiconductor, which was itself founded by former members of the laboratory of transistor pioneer William Shockley.24 Andy Grove, a junior executive at Fairchild, joined Intel on the company’s first day.25 This founding generation of the company led Intel until the early 2000s: Noyce was CEO from 1968 to 1975, followed by Moore from 1975 to 1987, followed by Grove, who remained CEO until 1998 then chairman until 2004. More than anyone, Andy Grove is credited with architecting Intel’s takeover of the global CPU market in the 1980s and 1990s. As soon as Grove retired, Intel began making mistakes. Paul Otellini, who took over as CEO in 2005, was the first Intel CEO not to be an engineer by training.26
In the absence of reform, Intel’s default trajectory would be managed decline as x86 loses adoption compared to ARM and its manufacturing operations continue to be outcompeted by TSMC and Samsung, perhaps eventually licensing x86 more broadly or even becoming a fabless chip designer and giving up on manufacturing entirely. Instead, Intel has chosen to pivot harder into manufacturing and become a contract semiconductor manufacturer for other chip designers, a foundry. This change is being led since 2021 by Pat Gelsinger, an engineer by profession who worked at Intel from 1979 to 2009, where he was mentored by Andy Grove and rose quickly as a chip designer to become the company’s chief technology officer, before leaving the company due to estrangement from Paul Otellini.27
The Plan to Expand Intel’s Manufacturing
Since 2021, Intel’s reform strategy has centered around building out the company’s foundry services, outlined in a strategy referred to as “IDM 2.0.” This involves a target to make Intel the second-largest foundry by revenue by 2030, surpassing Samsung but behind TSMC. This entails reaching $5 billion in foundry revenue for 2027 and $15 billion by 2030.28 The strategy has three parts. First, the company is spending significant sums to increase its internal manufacturing capacity for its own chips. Second, it is prepared to outsource the production of some of its chips to third parties, including TSMC, to speed up the design of its more advanced chips. Intel has, since 2021, already been relying on TSMC to produce its Arc GPUs.29 Third, Intel plans to become a significant foundry for fabless designers by 2030.30 While marketed as a novel strategy, it is a similar strategy to that of Samsung, which in 2017 split off its foundry business from its semiconductor design business.31
Gelsinger has emphasized a cultural shift at the company in his short tenure. He has promoted the adoption of “Grovian” execution. Andy Grove did much to popularize the use of “objectives and key results” (OKRs) to determine executive compensation.32 Gelsinger has also made clear his desire to have a stronger relationship with equipment suppliers, relying more heavily on them for system integration and being willing to pay higher costs for more technical assistance and preferential treatment in shipment orders. The director of Intel Foundry technology, Ann Kelleher, has been largely given free rein to spend on new technology.33 From 2021 to 2023, Intel’s research spending stayed roughly the same but increased its share of the revenue from 19% to 30%.34 Meanwhile, capital expenditure increased by $7 billion, going from 25% of revenue to 45%.35 Gelsinger began his tenure by halting share buybacks, which effectively financially reward stockholders using the company balance sheet.36
Intel currently has five major wafer fabrication facilities and five facilities dedicated to semiconductor assembly, testing, and packaging. Its wafer fabrication facilities are in the U.S., Europe, and Israel, and its assembly and testing facilities are in lower-wage jurisdictions in Costa Rica and East Asia. Intel’s largest current wafer fab operation is in Oregon, where it employs 22,000 people.37 Its second-largest is in Arizona, where it employs 12,000 people. Intel’s Ireland facility is based in Leixlip, Kildare County. The company chose Ireland as its European headquarters in 1989, with two factories today employing 5000.38 Intel also has close to 12,000 employees in Israel, beginning its operations in 1974 and building its first factory there in 1981.39 The Fab 28 facility at Kiryat Gat is capable of manufacturing 7 nm chips, currently Intel’s most advanced process node.40 Both Ireland and Israel have the requisite local workforce to meet Intel’s needs and have historically supported the company through either direct subsidies or generous tax incentives.
In 2021, before Intel’s expansion announcements, it had the sixth-largest capacity for wafer fabrication globally, with the ability to produce nearly 900,000 wafers per month. For comparison, Samsung’s foundry had the capacity to produce 3.1 million wafers per month, followed by TSMC’s 2.7 million. The other significant manufacturers larger than Intel were Micron (1.9 million), SK Hynix (1.9 million), and Japanese memory foundry Kioxia (1.6 million).41 Intel needs to significantly expand its manufacturing capacity to become a globally competitive foundry, and it is.
In addition to expanding its existing wafer fabrication facilities in Arizona and elsewhere, Intel is building in two new locations—Ohio and Magdeburg, Germany—and an additional assembly and testing facility in Poland. The Ohio facility is expected to begin production in 2026, while the European facilities in 2027.42 Wafer fabrication plants have become enormously expensive to build due to the increasing demands of semiconductor manufacturing. A typical fab facility consists of three or four floors. The top floor is filled with fan and filter systems to keep the air in the facility free from particulates. The middle floor, the “clean room,” is where the fabrication process occurs and is filled with thousands of specialized machines. The ground and subterranean floors house pumps and electrical systems to carry electricity, gases, liquids, and waste to and from the clean room.43
The costs of the equipment used in a facility make up 80% of the overall expense.44 Based on estimates from the Institute for Progress, a fab’s cost is roughly 60% specialist manufacturing equipment from companies like Tokyo Electron, another 20% specifically for lithography machines from companies like ASML, and the remaining 20% spread across electrical systems, piping, heating, ventilation, air conditioning, concrete, steel, and site development.45 In 2022, Intel preemptively secured the 2024 stock of ASML’s most advanced extreme ultraviolet lithography machines. This totaled six machines, each costing $370 million.46 Intel’s investment in Ohio is expected to cost $20 billion.47 Because cost is tied to equipment spending, upgrading established plants is about as expensive as setting up new ones. Intel is spending $36 billion to upgrade its wafer fabs in Oregon and committed to spending $25 billion to upgrade its Israeli manufacturing base.48
In June 2024, Intel paused construction of its $25 billion fab project in Israel.49 At the same time, the Ohio plant, originally intended to open in 2025, has been delayed to 2027. The Magdeburg plant has also been delayed from starting construction in 2023 to at least 2025 due to $10 billion worth of subsidies not yet being approved by the EU Competition Authority.50 Given that Intel is not the only semiconductor manufacturer to see ballooning costs or delays in its recent fab construction projects, these problems seem not primarily related to the functionality of these companies in manufacturing semiconductors, but rather stem from construction, regulatory, logistical, and financial problems that are likely to be eventually resolved.51
While previously reporting revenues based only on product types, since 2024 Intel has split off the reporting of all of its manufacturing operations as a separate unit called Intel Foundry.52 In the second quarter of 2024, Intel Foundry reported $4.3 billion in revenue, but Intel also took off $4.3 billion in “intersegment eliminations.”53 This implies that $4.3 billion is the value assigned by Intel to the manufacturing of its own chip designs and that Intel Foundry so far has no customers besides Intel itself, which seems to comport with sparse publicly-available discussion of Intel’s foundry customers except statements like that Intel has $15 billion in “expected lifetime deal value with external customers” by 2030.54 In early August 2024, the company said it expected the “first external customer” to see production begin on its flagship process node, 18A, in the first half of 2025.55
Since 2021, Intel has variously announced that Microsoft, Qualcomm, and AWS would be future customers for its foundry services.56 Intel’s $4 billion in self-assessed and self-supplied quarterly manufacturing revenue pales in comparison to TSMC’s $18.8 billion in foundry revenue in the first quarter of 2024.57 It is higher than second-place Samsung’s estimated $3.4 billion, but this presumably counts only or mostly contract semiconductor manufacturing for external customers—rather than Samsung’s own chips—in which case Intel’s proper figure so far would presumably be $0.58
There is a logic to Intel focusing heavily on foundry capacity. First and foremost, long-term demand growth for foundry capacity is expected. One market research report estimates total foundry revenue for chips will grow from $106 billion in 2022 to $240 billion in 2032.59 Notably, 72% of current wafer fabrication capacity is in Taiwan or China.60 All of Intel’s expansions are in the U.S., Europe, or Israel, where governments perceive a need to increase capacity, meaning Intel receives significant subsidies. It can also market itself as a secure supply chain that is not vulnerable to hypothetical geopolitical shocks like a Chinese invasion of Taiwan. Financial subsidies and political pressure can push through the creation of new facilities and production lines equipped with expensive machines supplied by other functional companies.
The U.S. Government Backs Intel’s Foundry Pivot
Starting with the Trump administration in 2017 and continuing into the Biden administration since 2021, the U.S. government has taken a more protectionist stance on trade with China, especially trade in advanced, valuable technologies like semiconductors. Chinese companies like Huawei have been restricted from participating in U.S.-aligned markets while exports to China of cutting-edge semiconductor technology—like ASML’s extreme ultraviolet lithography machines—have been banned. This change is motivated partially by economic and industrial policy concerns, but more importantly by military and national security concerns. An authoritative report commissioned by the U.S. Congress and Defense Department in 2018 concluded that “if a potential adversary bests the United States in semiconductors over the long term or suddenly cuts off U.S. access to cutting-edge chips entirely, it could gain the upper hand in every domain of warfare.”61
Intel’s foundry pivot thus seems to be in large part a response to the U.S. government’s implicit and explicit offer of financial and political support for expanding semiconductor manufacturing capacity outside of East Asia, rather than an original live player strategy. While Pat Gelsinger became CEO of Intel in February 2021, joining from a decade at the software company VMware, TSMC had announced plans to build new fabs in the U.S. with U.S. government support as early as May 2020.62 It appears that Gelsinger was brought in due to his status as a successful executive and company veteran, rather than expertise in manufacturing.
In 2022, the U.S. Congress passed the CHIPS and Science Act, a sweeping piece of legislation that ordered hundreds of billions of dollars in spending on scientific and technological projects, including $39 billion in direct subsidies for semiconductor manufacturing in the U.S. The CHIPS Act was framed as a major cornerstone of U.S. industrial policy, and was unsurprisingly aggressively pushed by Gelsinger personally.63 While rivals TSMC and Samsung are receiving billions to build fabs on U.S. soil, Intel is the largest direct beneficiary.64 The bill was heavily lobbied for by the Semiconductor Industry Association (SIA), of which Gelsinger is a board member.
The CHIPS Act has given Intel an estimated $44.5 billion of total federal government funding commitments, including $8.5 billion in direct funding, $11 billion in loans, and $25 billion in tax credits.65 Separately, Intel has also received direct U.S. Defense Department funding, receiving $3.5 billion in 2024 to build military chips deemed too critical to manufacture through standard commercial deals.66 There are indications Intel could be the beneficiary of a second CHIPS act in the future, with U.S. Secretary of Commerce Gina Raimondo claiming at an Intel event that it will be necessary for the U.S. government to spend more money to competitively produce the most advanced chips.67
Following the U.S. federal government’s lead, other governments and companies have also stepped in to provide Intel with generous funding. The Ohio state government and New Albany municipal government have committed to $2.1 billion in subsidies.68 Meanwhile, the German government has committed to $10.9 billion in subsidies for the Magdeburg facility. The Canadian financial firm Brookfield, meanwhile, has agreed to provide $15 billion in investment funding for the Arizona facility in exchange for 49% ownership.69 In total, Intel is expected to spend at least $116 billion in capital spending on its fab expansions.70
Despite all this spending and support, there is little reason to expect Intel to surpass or even match TSMC in price or sophistication, although matching or surpassing TSMC in raw capacity may be eventually achievable with enough spending; not even Intel has set the goal of surpassing TSMC, only the goal of becoming the second-largest foundry after TSMC by 2030. The reason is that the U.S., Europe, and Israel all lack the workplace discipline, dense supplier ecosystems, lower salaries, skilled personnel, and traditions of knowledge in manufacturing that TSMC has in Taiwan.
TSMC founder Morris Chang, now 93 years old, made headlines in 2023 for reportedly grilling high-ranking U.S. politician Nancy Pelosi at a private dinner on whether the U.S. really understood the difficulty of building a semiconductor manufacturing base.71 In Chang’s view, the push to “reshore” U.S. semiconductor manufacturing will be “a very expensive exercise in futility” because unit costs will never fall to Taiwanese levels; he cited the example that TSMC’s long-operating fab in Oregon remains 50% more costly per unit compared to Taiwanese fabs, although still profitable.72
Gelsinger has stacked his executive team at Intel with allies from VMware rather than manufacturing experts.73 Intel Foundry’s initial president was Randhir Thakur, a veteran of the semiconductor equipment manufacturing company Applied Materials and Intel’s former head of supply chains. In 2023, just two years into the job, Thakur resigned and moved to become the CEO of Tata Electronics, a subsidiary of the Indian Tata Group.74 He was eventually replaced by Kevin O’Buckley, a semiconductor industry veteran who spent seventeen years at IBM.75
Thakur was just one of several outside hires brought in by Gelsinger to improve Intel’s foundry capacity. A number of critical hires came directly from TSMC, Samsung’s foundry business, and from the Taiwanese foundry UMC.76 For example, Intel’s new vice president of customer enablement, Michael Chang, had been with TSMC for over thirty years.77 But although China’s SMIC and Samsung have in the past poached high-level manufacturing and R&D personnel from TSMC, it is unclear that Intel has poached personnel of comparable manufacturing expertise. If TSMC itself has only achieved manufacturing chips that are 50% more expensive than its Taiwanese-made chips in the U.S., Intel is unlikely to ever surpass this margin.
Gelsinger is attempting a radical reform of Intel. He has not shied away from difficult measures, including announcing the firing of 15% of the company’s workforce to cut costs in August 2024.78 So long as government support continues, it seems likely that Intel will ultimately succeed at expanding its production in the U.S. But it seems unlikely that Gelsinger will succeed at forging a manufacturing culture that goes against the general decline in U.S. manufacturing discipline visible at other companies like Boeing and matches, let alone exceeds, that of a company like TSMC. In recent years, it has taken exceptional live players like Elon Musk to found functional manufacturing organizations in the U.S., like Tesla. Although Intel may well succeed at increasing production, if its unit costs remain high regardless well into the future, Intel will not have been reformed into a functional and dynamic company, but rather a state-backed dead player not dissimilar to Boeing or defense contractors like Lockheed Martin, akin to a state-owned enterprise or “national champion” in countries smaller than the United States.
Become a paid subscriber and get a new in-depth investigation of a key live player, institution, or industry in your inbox every Wednesday:
“Top publicly traded semiconductor companies by number of employees.” Largest Companies by Market Cap, https://companiesmarketcap.com/semiconductors/largest-semiconductor-companies-by-number-of-employees/.
Analyst interview with semiconductor design professional, August 2024.
“Intel Reports Second-Quarter 2024 Financial Results.” Intel, 2024, https://d1io3yog0oux5.cloudfront.net/_41962b00f34b09fd6e3d82d3c6237470/intel/db/887/9077/earnings_release/FINAL+Q2+24+Earnings+Release.pdf
Brodzicki, Tomasz. “The role of East and Southeast Asia in the Global Value Chain in Electronics.” S&P Global, 2 November 2021, https://www.spglobal.com/marketintelligence/en/mi/research-analysis/the-role-of-east-and-southeast-asia-in-the-global-value-chain-.html.
Martin, Dylan. “AMD Gains CPU Share Against Intel In Desktop, Server Segments.” CRN, 13 May 2024, https://www.crn.com/news/components-peripherals/2024/amd-gains-cpu-share-against-intel-in-desktop-server-segments.
“PassMark CPU Benchmarks - AMD vs Intel Market Share.” CPU Benchmarks, https://www.cpubenchmark.net/market_share.html.
Lee, Timothy B. “Intel made a huge mistake 10 years ago. Now 12000 workers are paying the price.” Vox, 20 April 2016, https://www.vox.com/2016/4/20/11463818/intel-iphone-mobile-revolution.
“A64 -- Base Instructions (alphabetic order).” Arm Developer, https://developer.arm.com/documentation/ddi0596/2021-03/Base-Instructions?lang=en
Bailey, Brian. “A Minimal RISC-V.” Semiconductor Engineering, 13 January 2022, https://semiengineering.com/a-minimal-risc-v/.
Analyst interview with semiconductor design professional, August 2024.
Morris, Iain. “Arm is landing blows on x86 amid AI frenzy - Arm is landing blows on x86 amid AI frenzy.” Light Reading, 8 February 2024,
https://www.lightreading.com/semiconductors/arm-is-landing-blows-on-x86-amid-ai-frenzy.
Clark, Don. “Paul S. Otellini, Who Led Intel and Saw It Grow Even More, Dies at 66 (Published 2017).” The New York Times, 3 October 2017, https://www.nytimes.com/2017/10/03/obituaries/paul-s-otellini-who-led-intel-and-saw-it-grow-even-more-dies-at-66.html.
Warren, Tom, and Amelia Holowaty Krales. “Microsoft is confident Windows on Arm could finally beat Apple.” The Verge, 8 April 2024, https://www.theverge.com/2024/4/8/24116587/microsoft-macbook-air-surface-arm-qualcomm-snapdragon-x-elite.
Wright, Arol. “Project Larrabee: How Intel's First Attempt at GPUs Failed.” How-To Geek, 16 July 2023, https://www.howtogeek.com/896521/project-larrabee-how-intels-first-attempt-at-gpus-failed/.
“NVIDIA To Dominate AI Share In 2024 With An Estimated $46 Billion In Data Center Revenue.” Wccftech, 1 February 2024, https://wccftech.com/nvidia-to-dominate-data-center-share-in-2024-46-billion-usd-revenue-expected/.
Szewczyk, Chris. “Intel CEO Pat Gelsinger candidly reveals where Intel dropped the ball in recent years.” PC Gamer, 5 November 2023, https://www.pcgamer.com/intel-ceo-pat-gelsinger-candidly-reveals-where-intel-dropped-the-ball-in-recent-years/.
Valerio, Pablo. “Taiwan Already Has 3nm Chips; Why the U.S. Lags.” EPS News, 9 January 2023, https://epsnews.com/2023/01/09/taiwan-already-has-3-nm-chips-why-the-u-s-lags/.
Kabir, Omer. “The price of failure: Dutch company's chip monopoly born from a bad bet by Intel.” CTech, 15 April 2024, https://www.calcalistech.com/ctechnews/article/y4j6dd07g.
Ibid.
Fitch, Asa. “Intel’s Success Came With Making Its Own Chips. Until Now.” The Wall Street Journal, 6 November, 2020, https://archive.is/WfMa9
Moore's law is the observation that the number of transistors in an integrated circuit (IC) doubles about every two years.
“The Story of Intel.” The History of Computing, 7 March 2023, https://thehistoryofcomputing.net/the-story-of-intel.
Ibid.
Hardy, Quentin. “Intel Chief Executive to Retire in May.” The New York Times, 19 November 2012, https://archive.nytimes.com/bits.blogs.nytimes.com/2012/11/19/intel-chief-executive-to-retire-in-may/
Brock, David C. “Oral History of Pat Gelsinger.” Computer History Museum - Archive Server, 9 January 2019, https://archive.computerhistory.org/resources/access/text/2019/06/102781029-05-01-acc.pdf.
“Intel Targets $15B In Foundry Revenue By 2030, But Analyst Says TSMC Will Be '10X Larger.'” Markets Insider, 3 April 2024, https://markets.businessinsider.com/news/stocks/intel-targets-15b-in-foundry-revenue-by-2030-but-analyst-says-tsmc-will-be-10x-larger-1033218680.
“Intel explains why it is outsourcing chip manufacturing to TSMC.” Gizmochina, 22 September 2021, https://www.gizmochina.com/2021/09/22/intel-explains-why-it-is-outsourcing-chip-manufacturing-to-tsmc/.
“Intel CEO Announces 'IDM 2.0' Strategy for Manufacturing, Innovation,...” Intel, 23 March 2021, https://www.intel.com/content/www/us/en/newsroom/news/idm-manufacturing-innovation-product-leadership.html
Cho, - Hyun. “Samsung's chip division wants the top spot back: What's happening?” ZDNet, 31 March 2024, https://www.zdnet.com/article/samsungs-chip-division-wants-the-top-spot-back-whats-happening/.
“The Essence of OKRs by Andy Grove: a 2-minute overview on OKRs.” Scaled OKRs, 24 February 2023, https://www.scaledokrs.com/post/the-essence-of-okrs-by-andy-grove.
“Ann Kelleher - Intel Corporation.” LinkedIn, https://www.linkedin.com/in/ann-kelleher-6184b89/.
“2023 Annual Report.” Intel, 2024, https://d1io3yog0oux5.cloudfront.net/_2b79cce658ecbeffd57b3484a5150489/intel/db/888/9001/file/426890%281%29_9_Intel+AR_WR_LR.pdf
Ibid.
Richter, Wolf. “Intel to Get $23 Billion in Government Grants & Loans Plus $25 Billion Investment Tax Credits, to Invest $100 Billion in the US, after Wasting $94 Billion on Share Buybacks in 15 Years.” Wolf Street, 20 March 2024, https://wolfstreet.com/2024/03/20/intel-to-get-23-billion-in-government-grants-loans-plus-25-billion-investment-tax-credits-to-invest-100-billion-in-the-us-after-wasting-94-billion-on-share-buybacks-in-15-years/.
“Oregon Campuses.” Explore Intel., https://www.exploreintel.com/campuses/oregon.
“Ireland's role in the global semiconductor industry.” Tyndall National Institute, 25 July 2023, https://www.tyndall.ie/contentFiles/Tyndall_Ireland's_Role_in_the_Global_Semiconductor_Industry.pdf.
“Intel in Israel.” Intel in Israel, https://www.intel.co.il/content/www/il/he/company-overview/intel-in-israel.html.
Scheer, Steven. “Israel grants Intel $3.2 billion for new $25 billion chip plant.” Reuters, 26 December 2023, https://www.reuters.com/technology/intel-get-32-billion-government-grant-new-25-billion-israel-chip-plant-2023-12-26/.
Flaherty, Nick. “Top five chip makers dominate global wafer capacity ...” eeNews Europe, 11 February 2021, https://www.eenewseurope.com/en/top-five-chip-makers-dominate-global-wafer-capacity/.
“Intel Announces Initial Investment of Over €33 Billion for R&D and...” Intel, 15 March 2022, https://www.intel.com/content/www/us/en/newsroom/news/eu-news-2022-release.html#gs.br4ei8.
“What does it take to build a fab?” Intel, https://download.intel.com/newsroom/2022/manufacturing/fab-final-static.pdf.
Potter, Brian. “How to Build a $20 Billion Semiconductor Fab.” Construction Physics, 3 May 2024, link.
Ibid.
Trueman, Charlotte. “Intel acquires ASML's entire 2024 stock of High NA EUV machines.” Data Center Dynamics, 9 May 2024, https://www.datacenterdynamics.com/en/news/intel-acquires-asmls-entire-2024-stock-of-high-na-euv-machines/.
Shankland, Stephen. “Intel's $100B Ohio 'megafab' could become world's largest chip plant.” CNET, 21 January 2022, https://www.cnet.com/tech/computing/intels-100b-ohio-megafab-could-become-worlds-largest-chip-plant/.
Mann, Tobias. “Israel to invest $3.2B in $25B Intel fab project.” The Register, 27 December 2023, https://www.theregister.com/2023/12/27/intel_israel_fab/.
Ibid.
“Intel’s 1nm-class Fabs in Germany Reportedly Delayed Due to Black Soil Concerns and Pending EU Subsidy Approval.” TrendForce, 31 May 2024, https://www.trendforce.com/news/2024/05/31/news-intels-1nm-class-fabs-in-germany-reportedly-delayed-due-to-black-soil-concerns-and-pending-eu-subsidy-approval/.
Mann, Tobias. “Intel's foundry plan is costing far more than expected.” The Register, 16 April 2024, https://www.theregister.com/2024/04/16/intel_foundry_vision/.
“Intel Outlines Financial Framework for Foundry Business, Sets Path to Margin Expansion.” Investor Relations :: Intel Corporation (INTC), 2 April 2024, https://www.intc.com/news-events/press-releases/detail/1687/intel-outlines-financial-framework-for-foundry-busines.
“Intel Reports Second-Quarter 2024 Financial Results.” Intel, 2024, https://d1io3yog0oux5.cloudfront.net/_41962b00f34b09fd6e3d82d3c6237470/intel/db/887/9077/earnings_release/FINAL+Q2+24+Earnings+Release.pdf
“Intel Outlines Financial Framework for Foundry Business, Sets Path to...” Intel, 2 April 2024, https://www.intel.com/content/www/us/en/newsroom/news/intel-outlines-new-financial-reporting-structure.html#gs.dqj8j7.
Moss, Sebastian. “Intel to manufacture Qualcomm chips, package AWS chiplets.” Data Center Dynamics, 27 July 2021, https://www.datacenterdynamics.com/en/news/intel-to-manufacture-qualcomm-chips-package-aws-chiplets/. ; Martin, Dylan. “Intel Launches 'World's First Systems Foundry' With Expanded Road Map To Take On TSMC, Samsung.” CRN, 21 February 2024, https://www.crn.com/news/components-peripherals/2024/intel-launches-world-s-first-systems-foundry-with-expanded-road-map-to-take-on-tsmc-samsung.
Alsop, Thomas. “Top semiconductor foundries revenue by quarter 2024.” Statista, 2 July 2024, https://www.statista.com/statistics/867210/worldwide-semiconductor-foundries-by-revenue/.
Ibid.
Cho, Hyun. “Samsung's chip division wants the top spot back: What's happening?” ZDNet, 31 March 2024, https://www.zdnet.com/article/samsungs-chip-division-wants-the-top-spot-back-whats-happening/.
“China and US Bolster Semiconductor Independence as Taiwan's Foundry Capacity Share Projected to Decline to 41% by 2027, Says TrendForce.” TrendForce, 14 December 2023, https://www.trendforce.com/presscenter/news/20231214-11959.html.
Shivakumar, Sujai, and Charles Wessner. “Semiconductors and National Defense: What Are the Stakes?” CSIS, 8 June 2022, https://www.csis.org/analysis/semiconductors-and-national-defense-what-are-stakes.
“TSMC Announces Intention to Build and Operate an Advanced Semiconductor Fab in the United States.” Taiwan Semiconductor Manufacturing Company, 15 May 2020, https://pr.tsmc.com/english/news/2033.
“President Biden Announces Agreement with Intel.” Rev, 21 March, 2024, https://www.rev.com/blog/transcripts/president-biden-announces-agreement-with-intel
Lu, Marcus. “All of the Grants Given by the U.S. CHIPS Act.” Visual Capitalist, 26 April 2024, https://www.visualcapitalist.com/all-of-the-grants-given-by-the-u-s-chips-act/.
Ibid.
Hawkins, Mackenzie. “Intel Stands to Win $3.5 Billion to Produce Chips for Military.” Bloomberg, 6 March, 2024, https://www.bloomberg.com/news/articles/2024-03-06/intel-stands-to-win-3-5-billion-to-produce-chips-for-military?sref=ZqW0mZJf
Yu, Yifan. “U.S. needs another CHIPS Act to lead world, says Raimondo.” Nikkei Asia, 22 February 2024, https://asia.nikkei.com/Business/Tech/Semiconductors/U.S.-needs-another-CHIPS-Act-to-lead-world-says-Raimondo.
Williams, Mark. “Intel construction report identifies completion delay, spending for New Albany plant.” The Columbus Dispatch, 15 March, 2024, https://eu.dispatch.com/story/business/manufacturing/2024/03/15/intel-columbus-ohio-new-albany-plant-construction-jobs-hiring/72988955007/
“Brookfield Infrastructure Signs Definitive Agreement with Intel.” Brookfield Infrastructure Partners, 23 August 2022, https://bip.brookfield.com/press-releases/bip/brookfield-infrastructure-signs-definitive-agreement-intel.
Burns, Alexander. “Taiwan's Tech King to Nancy Pelosi: U.S. Is in Over Its Head.” Politico, 14 February 2023, https://www.politico.com/news/2023/02/14/taiwan-tech-king-pelosi-powerhouse-microchip-industry-00082646.
“Can semiconductor manufacturing return to the US?.” Brookings Institution, 14 April 2022, https://www.brookings.edu/wp-content/uploads/2022/04/Vying-for-Talent-Morris-Chang-20220414.pdf.
Richter, Wolf. “Intel to Get $23 Billion in Government Grants & Loans Plus $25 Billion Investment Tax Credits, to Invest $100 Billion in the US, after Wasting $94 Billion on Share Buybacks in 15 Years.” Wolf Street, 20 March 2024, https://wolfstreet.com/2024/03/20/intel-to-get-23-billion-in-government-grants-loans-plus-25-billion-investment-tax-credits-to-invest-100-billion-in-the-us-after-wasting-94-billion-on-share-buybacks-in-15-years/.
Martin, Dylan. “Head of Intel Foundry Services resigns just as it gets going.” The Register, 21 November 2022, https://www.theregister.com/2022/11/21/intel_foundry_services_boss_quits/.
“Kevin O'Buckley to Lead Foundry Services at Intel.” Intel, 13 May 2024, https://www.intel.com/content/www/us/en/newsroom/news/intel-foundry-services-leadership-news.html.
Martin, Dylan. “Head of Intel Foundry Services resigns just as it gets going.” The Register, 21 November 2022, https://www.theregister.com/2022/11/21/intel_foundry_services_boss_quits/.
“Michael Chang.” Linkedin, https://www.linkedin.com/in/michael-chang-0627318a/details/experience/.
Hollister, Sean. “Intel is laying off over 15,000 employees and will stop ‘non-essential work’.” The Verge, 1 August 2024, https://www.theverge.com/2024/8/1/24210656/intel-is-laying-off-over-10000-employees-and-will-cut-10-billion-in-costs.