Intel’s state-of-the-art Fab 52 in Arizona. (Credit: Intel)
Intel has undergone a tremendous amount of change over the past year-plus. It has been in the news for reasons it would probably rather not be, including major layoffs, an unexpected change in CEO, and the U.S. government’s surprise acquisition of a stake in the company. More broadly, over the past few years it has lost market share in processors for both the datacenter and PCs, and so far it has missed the datacenter GPU craze altogether. Meanwhile, its old foe AMD’s growth in servers and PCs has been at its expense, while NVIDIA has set the tone for AI (and market valuation) like Intel used to set the tone for CPUs.
Yet amidst all of this upheaval, Intel still plays a bellwether role in the chip industry, in part for its sheer size, but more importantly because it is the only company that can provide both onshore leading-edge silicon manufacturing and onshore R&D to create new IP for future generations of chip production. While TSMC has leading-edge production nodes in its Arizona fabs, it keeps the crown jewels of its IP in Taiwan. Which hints at the real problem facing the U.S.-based chip industry.
Please understand, I am not casting any aspersions on TSMC, which has been a reliable supplier serving many different U.S.-based semiconductor design companies for decades. But the inescapable realities of the current chip landscape — especially American companies’ dependency on Taiwan-based production — should keep U.S. policy makers up at night. And Intel is the only company that can supply the complete antidote.
The Political and Geopolitical Issues at Stake
Let’s pull back to get a better view of the issues. At the level of U.S. local and national politics, saying “Let’s make stuff in America” always plays well, and even more so under the current administration. A U.S.-first view also fits President Trump’s approach to deal-making and tariffs. Please keep in mind that I’m not picking any sides here, but simply describing the lay of the land.It is worth noting that TSMC and Samsung (the other leading-edge chip foundries besides Intel operating in the United States) can also take advantage of a U.S.-first policy approach by building more fabs and creating more jobs here in the United States. But Samsung Foundry is just now competing hard on leading-edge nodes, and TSMC must send the cutting-edge wafers it manufactures in the U.S. back to Taiwan for advanced packaging — at least until Amkor, TSMC’s neighbor in Arizona, gets its new packaging facility operational. That’s a potentially dangerous open loop.
Going past U.S. politics to deeper geopolitical considerations, there is substantial risk created — in both economic and national-security terms — when more than 90% of the U.S.’s most advanced chips are made in Taiwan. That concentration of production makes Taiwan a chokepoint that threatens vital U.S. interests.
In this context, it is obviously easy to imagine chaotic scenarios involving major moves by China, which has always regarded Taiwan as part of its own sovereign territory. But even setting aside that dimension of the problem, what happens if Taiwan suffers a major earthquake or other natural disaster that knocks TSMC facilities offline, even for a few months? Beyond that, what happens if the Taiwan government decides to play hardball with the U.S. for any reason, let’s say in response to tariff threats?
Here’s my point: the ripple effects of any significant disruption in Taiwan-based production would be significant for many U.S.-based chip makers, starting with NVIDIA, AMD, Broadcom, Marvell, and Qualcomm. Let’s just pause for a moment to consider the impact if those five companies had to delay or cancel shipments of chips to their many customers that manufacture servers, PCs, automobiles, smartphones, networking equipment, military equipment … It’s a very long list, and the ripple effects would get even worse if any disruption extended long enough to affect transitions to the next production nodes for new generations of chips.
In short, there is a positive view that emphasizes wanting the U.S.-based chip industry to be competitive on the world stage. That is a good thing for the U.S. economy and job creation, and it’s well worth promoting. Meanwhile, the negative view is much more sobering. There is a new sort of cold war going on between the United States and China, and it is playing out largely in the commercial sphere. We cannot risk the disruption to the U.S. economy that would arise from any significant interruption of leading-edge chip production in Taiwan. Therefore, we need to promote U.S.-based leading-edge chip production capacity to mitigate that risk.
We also can’t chance any significant disruption in the progression of foundry R&D as each generation of leading-edge nodes gives way to the next. These cycles of production and innovation must be secure from either accidents or intentional actions beyond our control occurring in other countries. The potential downsides — for both the economy and national security — are too big to be complacent about.
How Intel Fits into this Picture
My colleagues Matt Kimbell and Anshel Sag have recently written about different aspects of Intel’s place in the chip industry and the broader tech economy. Kimbell’s paper focuses more on Intel’s history of innovation on the R&D side and how it is now reestablishing its leadership. Meanwhile, Sag’s article digs into how Intel Foundry’s leading-edge production nodes could positively impact other chip makers.Those production nodes have been drawing a lot of attention — as they should. The launch of the new Intel 18A node is the final piece in the “five nodes in four years” (5N4Y) plan to revamp production quality that may go down in history as one of former CEO Pat Gelsinger’s two greatest legacies at Intel. (The other one: leading the team that built the x86 architecture back in the 1980s.) Intel 18A is up and running now for Intel products, and I’m bullish about it. From everything I’ve seen — and all the Intel customers I’ve talked to — I’m very confident that Intel will use 18A to successfully produce tens of millions of leading-edge chips in the Panther Lake family and beyond in the near term. I believe 18A can stack up against any production node in its class in the world, so it really is getting Intel back to industry leadership.
It is also a crucial stair-step in the progression toward the subsequent nodes on the roadmap. Next up is Intel 18A-P. I’m confident about the progression to it from 18A; Intel will spend the next year optimizing 18A-P for performance and yield, and I expect it to be in production before 2026 is out. The final step in the current sequence: Intel 14A. That’s further toward the horizon for volume production, but it has an increased likelihood of success because of the great work already done on 18A and 18A-P. Intel customers I’ve talked with who have seen this one say that 14A is the real deal. And as Sag hinted in his article, 14A should be highly competitive not only in the datacenter and PC markets, but in mobile chips as well, which will be an important shift for the company. I am eager for Intel to get its 14A .5 PDK out and start hearing feedback about it. Make no mistake, though: even without the PDK out in the wild, I am already hearing very positive things, given the 18A progress, as each new node builds upon the others.
There are precautions that should be taken along the way. This is because, as much as I like and respect CEO Lip-Bu Tan and the rest of the Intel leadership, there are industry-scale risks that inevitably arise if we have just one U.S.-based leading-edge foundry. Particular attention must be paid to the issue of which companies get wafer allocation from Intel Foundry. The prospective customers I talk with — and I have spoken recently to nearly every CEO of potential Intel Foundry anchor tenants — consistently express their need to have real confidence that wafer allocations will be scrupulously fair. This will prevent even the appearance of any kind of unfair self-dealing if Intel’s design business has a big client that it wants to satisfy.
To its credit, Intel has announced its intention to establish Intel Foundry as a wholly owned subsidiary of Intel with a distinct leadership team and an independent operating infrastructure governed by an independent board. This will allow Intel Foundry to operate autonomously and accept external investments, with the aim of accelerating its progress as a trusted partner delivering wafer fabrication and advanced packaging services for both Intel and external customers. This kind of thing can certainly take time for a company of Intel’s scale and complexity. But I can’t declare that Intel has solved this problem until I — and, more importantly, prospective Intel Foundry customers — have a chance to know more details around leadership and governance, including wafer allocation. Will the new structure be enough to assuage the concerns of potential Intel Foundry clients? Maybe. But we can’t know that until the new structure is actually in place.
Extending Intel’s Unique Role in the U.S. Chip Industry
In an earlier phase of my career, I spent many years as an executive in the semiconductor industry thinking about how to make my company as competitive as it could be. I spent nearly a decade before that in the systems business running and marketing product lines where CPU choice was a big deal. These days, as an analyst working with many chip companies, their customers, and their customers’ customers, I spend a lot of time thinking about the importance of the semiconductor industry as a whole, and what the industry should look like in the future.The main takeaway here is that for all the reasons laid out above, the health of our onshore chip production and foundry IP development is too important to the U.S. economy and national security to leave to chance. The geopolitical threats we face dictate that we support and protect onshore leading-edge chip manufacturing. And the existing realities of the U.S. chip industry — in terms of engineering prowess, manufacturing footprint, and commercial relationships — dictate that it’s vital for Intel specifically to reestablish and maintain its leadership position in cutting-edge chip manufacturing.
ANALYST INSIGHT: Why Intel’s Onshore Chip Production Is So Important for U.S. Industrial Policy
Intel has undergone a tremendous amount of change over the past year-plus. It has been in the news for reasons it would probably rather not be, including major layoffs, an unexpected change in CEO, and the U.S. government’s surprise acquisition of a stake in the company. More broadly, over the...
moorinsightsstrategy.com