In its most recent earnings, Intel’s CEO, Lip-Bu Tan stated that future investments in next generation process technology at its fabs would only take place if they found significant new external customers to supplement their own internal needs. While the next generation process technology, designated 14A, is not expected to achieve volume production until 2028/29, initial engineering investments have already started. Given Intel’s difficulties, does this mean we should expect 14A to be eliminated? Are we destined to see future leading edge products from Intel be built on a fabless semiconductor model used by most other chip companies?
Is the threat to 14A real?
There are several reasons to believe that Intel will indeed ultimately deploy 14A in its fabs. First, they are already in the process of proving the basic technology for 14A based on their implementation of Intel’s current leading edge process, 18A. By learning from the challenges in 18A, 14A should be a relatively smoother transition. And there are several companies already working with Intel on evaluating its 14A process for their chip needs, so its highly likely that Intel could get at least one major customer in the near term, and likely others as well. That’s in addition to the US government committed to using Intel for its advanced chips for defense and space needs.
Next, it’s likely that the statement by Intel’s CEO was meant as much about leverage as about future investments. It's highly unlikely that given the pursuit of US based chip making emphasized by the US government and major funding already committed, that it would not support and incentivize Intel to indeed built out its 14A manufacturing capabilities. Beyond the technical issues, there are also political realities in play. Can the US afford to lose a leading edge foundry company after it set this as a major priority? And is the threat to shut down a way for Intel to leverage more US incentives? We expect to see the political ramifications play out over the next 1-2 years, but its likely the government will get involved if things look problematic.
Third, while there are examples of foundries running profitably on older processes, e.g., Global Foundries, is that a model that Intel could follow successfully? Much of what Intel produces needs leading edge processing technology to be competitive, especially in its PC and server chips. It currently does outsource some of its higher performance chips to TSMC, and has done so for some time, but this is generally on a supplemental basis to its own manufacturing. Can TSMC really handle all of the additional chips that Intel would need if Intel no longer had leading edge processing capability in house? Probably, but at what cost to Intel profitability? In theory Intel making its own chips should increase its margins, but that is not what has happened in the past with its less than efficient foundries. The issue here is, can Intel turn around its foundries to achieve market equivalent efficiency? It's seems to be progressing on the right path but it likely will take another 1-2 years to achieve.
Finally, it is still possible that Intel will get out of chip making altogether and sell its fab business, as has been speculated. But given the dynamics of the chip industry and Intel’s needs, it seems less likely now than it was 1-2 years ago. It is terribly expensive to build out leading edge nodes, costing $40B or more, so with the cost-cutting currently going on at Intel, this could be a possibility. But there are very few companies that could afford to buy outright all of Intel’s foundry business, and it’s unlikely that Intel would simply decide to shut it down.
Bottom Line
The next 1-2 years at Intel Foundry will be challenging. They need to create a much more efficient manufacturing capability and especially when it comes to the most modern process nodes. But there is every reason to believe that at the end of the day, Intel can turn around its foundry business to be a competitive internal and external chip maker and advance to its 14A process node. Given some of the uncertainties, we will need to watch closely what happens in the next 1-2 years before we can draw any real conclusions on long term viability of Intel Foundry Services. But the picture looks much better than it did a short while ago.
Jack Gold is founder and principal analyst at J.Gold Associates, LLC. With more than 45 years of experience in the computer and electronics industries, and as an industry analyst for more than 25 years, he covers the many aspects of business and consumer computing and emerging technologies. Follow him on Twitter @jckgld or LinkedIn at https://www.linkedin.com/in/jckgld.