We Went To Intel’s Arizona Chip Fab To See If It Can Regain Its Edge

Key Concepts

• Intel 18A: Intel’s advanced process node, aiming for parity with TSMC’s most advanced chips. Currently primarily used for internal Intel chip production.
• Moore’s Law: The observation that the number of transistors on a microchip doubles approximately every two years, though its pace has slowed.
• EUV Lithography (Extreme Ultraviolet Lithography): A crucial technology for manufacturing advanced chips, involving expensive machines from ASML. Intel’s delayed adoption of EUV contributed to past setbacks.
• Foundry: A semiconductor manufacturing business that produces chips for other companies (e.g., TSMC). Intel is attempting to become a major foundry player.
• RibbonFet & PowerVia: Intel’s innovative transistor architecture (RibbonFet – gate-all-around) and power delivery system (PowerVia – backside power) designed to improve performance and efficiency.
• IDM (Integrated Device Manufacturer): A company that designs, manufactures, and sells its own chips (like Intel), potentially creating conflicts of interest as a foundry.
• Chips Act: US legislation providing subsidies for domestic semiconductor manufacturing.

Intel’s Comeback Attempt: A Deep Dive into Fab 52 and Beyond

I. The Current Situation & Historical Context

Intel is currently attempting a significant comeback in the semiconductor industry, marked by the high-volume production of its 18A node at Fab 52 in Chandler, Arizona – a 1,000,000+ sq ft cleanroom facility. This effort comes after years of delays, missteps, and a 10% equity stake taken by the U.S. government. Founded in 1968 by Robert Noyce and Gordon Moore, Intel dominated the semiconductor landscape from the late 1970s through the early 2000s, reliably advancing process nodes in line with Moore’s Law. Jim Johnson, a 30+ year Intel veteran, recalls the 1990s as a period of “wonder and excitement” where Intel was the world’s largest and most profitable semiconductor company.

II. The Stumbles & Leadership Changes

The company’s recent struggles stem from a loss of “discipline of cycle time” during the 10nm and 7nm node transitions. A decision not to invest in costly EUV lithography machines in the 2010s resulted in significant delays (3+ years for 10nm, 2+ years for 7nm). This led to the loss of Apple as a major customer in 2020, who transitioned to TSMC for its M-series chips. Intel also missed opportunities in the mobile (rejecting Apple’s original iPhone processor deal) and AI revolutions (a failed GPU project under Pat Gelsinger in the mid-2000s).

In March 2024, Lip-Bu Tan replaced Pat Gelsinger as CEO, initiating significant restructuring. This included a 15% workforce reduction, a four-year delay of the Ohio chip factory, and paused projects in Germany and Poland, driven by over-investment without sufficient demand.

III. Intel 18A: Technology & Challenges

Intel 18A is positioned as competitive with TSMC’s 2nm node in certain benchmarks, like transistor density. Key technological advancements include:

• RibbonFet: A “gate-all-around” transistor architecture offering improved power control by fully surrounding the transistor channel. This provides a >15% performance per watt improvement over Intel 3.
• PowerVia: A backside power delivery system, moving power wiring from the front to the back of the chip, enhancing efficiency.

Despite these advancements, Intel 18A currently lacks significant external customer commitment. Intel itself is the primary customer, utilizing the node for its Core Ultra 3 processors launching in January, destined for PCs from Dell, HP, Lenovo, Asus, Acer, and Samsung by late 2026. Yield issues remain a concern, though Intel reports month-over-month improvements. Nvidia also experienced initial yield challenges with its Blackwell chips at TSMC, highlighting the inherent difficulties of advanced node production.

IV. Foundry Strategy & Customer Trust

Intel’s foundry strategy, spearheaded by Naga Chandrasekaran, aims to attract customers like Nvidia, Apple, and AMD. The core challenge is regaining customer trust, particularly given Intel’s history of delays and its status as an IDM (Integrated Device Manufacturer) – a potential competitor. Chandrasekaran emphasizes a shift towards “predictable execution” and a “customer-first mindset.”

Early deals with Microsoft and Amazon are considered positive signs, but their volumes are small compared to major players like Nvidia. AMD is reportedly considering Intel for manufacturing, and there’s speculation Apple might return for some Mac chip production by 2027.

V. Manufacturing Capabilities & Sustainability

Fab 52 is capable of running over 10,000 wafer starts per week of 18A and houses at least 15 EUV lithography machines. TSMC’s new Arizona fab is currently producing 4nm chips, while its most advanced 2nm is still limited to Taiwan.

Intel is committed to sustainable manufacturing, with nearly 100% renewable energy powering its Arizona facilities. The facility used over 3 billion gallons of water in 2024 but recycled 2.4 billion gallons back into the local supply.

VI. Government Support & Future Outlook

The U.S. government’s 10% equity stake in Intel reflects its commitment to reshoring semiconductor manufacturing, spurred by the $52 billion Chips Act. SoftBank also invested $2 billion, and Nvidia $5 billion (though without a commitment to use Intel’s foundry).

Intel is currently focused on cost-cutting and delaying projects while awaiting customer commitments for its next node, 14A, which is slated for development in Oregon with volume production targeted for 2028. The company believes 14A has garnered greater customer confidence. Construction is underway on Fab 62 in Arizona, potentially for 14A production.

VII. Key Quotes

• Jim Johnson (Intel Client Computing Head): “We lost the discipline of cycle time…We started talking ourselves into, hey, we can have longer cycle times and try and lift more or do more.”
• Naga Chandrasekaran (Intel Foundry Head): “I have to become part of their team and convince them that they can trust Intel Foundry to execute.”
• Lip-Bu Tan (Intel CEO): “We are making yield improvements, defect density improvements month-over-month and hitting our goals, so I believe we have turned the corner.”

Conclusion:

Intel faces a critical juncture. While Intel 18A demonstrates technological competitiveness, securing external customers and establishing a reliable foundry business remain significant hurdles. The combination of technological advancements, government support, and a renewed focus on execution under Lip-Bu Tan offers a path to recovery, but success hinges on regaining customer trust and delivering on its promises. The stakes are high, not only for Intel but for the U.S.’s strategic position in the global semiconductor landscape.