When Intel established its Intel Foundry Providers division in early 2021, it was clear that it wanted the contract chipmaking unit to be on par with Samsung and Taiwan Semiconductor Manufacturing Co. when it comes to scale as fabs and manufacturing nodes are getting costlier. The purpose was bold from the beginning, and it seems like the corporate intends to be aggressive, with plans to change into the second-largest foundry by 2030.
“Our ambition is to be the No. 2 foundry on this planet by the tip of the last decade, and [we] count on to generate main foundry margins,” stated Randhir Thakur, the president of Intel Foundry Providers, in an interview with Nikkei Asia.
Attending to the No. 2 spot on the worldwide foundry market implies that Intel should beat Samsung Foundry (the present No. 2, based on TrendForce), which generated over $20 billion in income in 2021 and is on monitor to exceed these leads to 2022. As of Q1 2022, Samsung Foundry managed some 16.3% of the worldwide foundry income, being dramatically behind market chief TSMC (53.6%) however considerably forward of its closest friends UMC (6.9%) and GlobalFoundries (5.9%)
In contrast, Intel’s IFS enterprise unit has generated $576 million in income up to now this yr. As soon as the acquisition of Tower Semiconductor closes in early 2023, Intel will add some $1.5 billion in income per yr to its IFS division. It will instantly make IFS the world’s No. 7 or No. 8 foundry, however it’s going to nonetheless be significantly smaller than Samsung Foundry when it comes to income.
To change into the world’s second-largest contract maker of chips, Intel should undertake a multi-faceted technique that includes the next:
- Growth of modern course of applied sciences which might be aggressive to these of Samsung Foundry and TSMC when it comes to energy, efficiency, and space (PPA); yields, and time-to-market.
- Construct modern capability for IFS shoppers. Primarily, the corporate should personal extra superior capability than Samsung Foundry by the late 2020s.
- Preserve operations and aggressive positions of Tower Semiconductor by innovating on the mature applied sciences entrance.
- Land orders primarily from shoppers that use TSMC and Samsung Foundry, maybe steal somebody from GlobalFoundries and SMIC too.
Want Huge Actions
Up to now, Intel has revealed a moderately aggressive course of know-how roadmap that includes excessive quantity manufacturing of chips on its 18A (18 angstroms or 0.18nm-class know-how) in 2025 and using Excessive-NA excessive ultraviolet lithography instruments for 18A if potential. Intel’s manufacturing node plan is significantly extra aggressive than that of Samsung Foundry and TSMC, which each plan to begin making 2nm-class (20 angstroms-class) chips in 2025 (which implies that these will likely be out there in very late 2025 on the earliest, or moderately in 2026).
On the semiconductor capability entrance, Intel’s plans are not any much less aggressive. The corporate is constructing its 20A-capable Fab 52 and Fab 62 at its camp close to Chandler, Arizona; establishing the primary two 18A/20A-capable modules of its website close to Columbus, Ohio; constructing its $3.5 billion facility for superior packaging operations; ending up a brand new Intel 4-capable module at its location close to Leixlip, Eire; and establishing an all-new fab close to Magdeburg, Germany. Total, Intel plans to speculate (or moderately co-invest with governments and semiconductor co-investment companions like Brookfield) round $100 billion in new semiconductor fabrication services within the coming years.
However Samsung is not any much less aggressive with its CapEx spending. Whereas Intel lately minimize its capital expenditures from $27 billion in 2022 to $25 billion, Samsung will make investments over $33 billion in new semiconductor manufacturing capability this yr and can retain its spending at across the similar stage subsequent yr, the corporate introduced lately. After all, it’s unclear how a lot of those sums are going to be invested in reminiscence (3D NAND and DRAM) manufacturing services and the way a lot will likely be used to increase logic capability of Samsung Foundry, however the South Korean firm is clearly very aggressive with its semiconductor enterprise, so it is going to be robust for Intel to match SF’s superior capacities.
Stealing shoppers from TSMC and Samsung Foundry will likely be even tougher as massive prospects like Nvidia or Qualcomm have provide agreements with their foundry companions that span for years. Moreover, it stays to be seen whether or not Intel’s fabs situated in Europe and America will be capable to supply the identical pricing as TSMC’s and Samsung Foundry’s fabs situated in Taiwan and South Korea.
It’s notably noteworthy that Intel is increase manufacturing capability within the U.S. and Europe and, up to now, has not introduced any plans to construct fabs in Taiwan or South Korea (it’s going to proceed to function a Tower Semiconductor fab in Japan). Whereas working fabs in Europe and America is costlier than in Taiwan or South Korea, constructing new fabs within the U.S. and Europe is smart each from buyer relationship and geopolitical factors of view.
On the one hand, U.S. prospects will likely be keen to make use of fabs in America as a consequence of logistics, threat administration, and different components. Alternatively, there are usually not so many chip designers in Europe that want modern manufacturing applied sciences. On the finish of the day, Intel’s product portfolio is increasing, so it’s going to want new fabs for itself anyway. As well as, based on Intel, a lot of its potential foundry shoppers see benefits in fabs situated in America or Europe.
“As we’ve engaged with foundry prospects since launching IFS, it has change into abundantly clear that many of those firms see the necessity for a extra resilient and geographically balanced semiconductor provide chain,” Thakur advised Nikkei Asia.
But, since each TSMC and Samsung Foundry are increase new modern fabs in Arizona and Texas, the geographical benefits of Intel’s new fabs in America could also be overestimated.
Since politicians in Europe and the U.S. need to construct home semiconductor provide chains in order to not depend on Taiwan so considerably, they’re desperate to co-invest in new services with Intel.
Make or Break
For Intel, the foundry enterprise is a strategy to improve its manufacturing volumes shortly and subsequently match its CapEx monetary capabilities with these of TSMC and Samsung. Getting a brand new income stream is important for Intel however just isn’t as necessary as rising manufacturing volumes.
Due to this fact, if the corporate succeeds in touchdown orders from many consumers that want superior manufacturing applied sciences, this will likely be successful by itself as it’s going to permit it to maintain investing within the growth of modern nodes and more and more costly fabs. If it doesn’t, it could in all probability stop to be a number one built-in design producer (IDM) and CPU provider over time. If the corporate manages to change into the world’s second largest foundry within the technique of organising its foundry enterprise pursuing its essential goal, this will likely be an excellent larger success. However volumes is the first goal.