When Intel establishes its Intel Foundry Services division in early 2021, the higher cost of fabs and production nodes will require contract chip manufacturing units comparable in size to Samsung and Taiwan Semiconductor Manufacturing Co. It was clear. The goal has been ambitious from the start, with plans to make him the second-largest foundry by 2030, and the company seems intent on getting aggressive.
big ambitions
“Our ambition is to be the number two foundry in the world by the end of the decade. [we] Randhir Thakur, president of Intel Foundry Services, said in the following interview: Nikkei Asia.
This means that Intel needs to beat Samsung Foundry to become #2 in the global foundry market (currently #2, trend force), expected to generate more than $20 billion in revenue in 2021 and surpass these results in 2022. As of Q1 2022, Samsung Foundry controls approximately 16.3% of global foundry revenue, well behind market leader TSMC (53.6%), but closest competitor UMC. (6.9%) and GlobalFoundries (5.9%).
In contrast, Intel’s IFS business unit has generated $576 million in revenue so far this year. When his Tower Semiconductor acquisition closes in early 2023, Intel will add about $1.5 billion in annual revenue to his IFS division. This would make IFS quickly become the world’s seventh or eighth largest foundry, but still significantly smaller than Samsung Foundry in terms of revenue.
To become the world’s second largest contract maker of chips, Intel must adopt a multi-pronged strategy that includes:
- Development of leading-edge process technology that rivals Samsung Foundry and TSMC in terms of power, performance and area (PPA). Yield, time to market.
- Build state-of-the-art capacity for your IFS clients. Basically, the company should own more advanced production capacity than Samsung Foundry by his late 2020s.
- Keep Tower Semiconductor operational and competitive by innovating at the forefront of mature technology.
- Land orders from clients that primarily use TSMC and Samsung Foundry will probably steal someone from GlobalFoundries and SMIC as well.
big action needed
So far, Intel has a pretty aggressive process to mass-produce chips at 18A (18 Angstrom or 0.18nm class technology) in 2025, adopting high NA extreme UV lithography tools for 18A if possible. Revealed technology roadmap. Intel’s production node plans are significantly more aggressive than Samsung Foundry and TSMC, which plan to begin manufacturing 2nm-class (20 Angstrom-class) chips in 2025 (i.e. they will be available in 2025 at the earliest). means it will be available in the second half of the year (or in 2026).
In terms of semiconductor capacity, Intel’s plans are less aggressive. The company builds his 20A capable Fab 52 and Fab 62 at a camp near Chandler, Arizona. Built the first two of his 18A/20A capable modules at a site near Columbus, Ohio. Building his $3.5 billion facility for advanced packaging operations. Finishing new Intel 4 capable modules at his location near Leixlip in Ireland. We are building an all-new fab near Magdeburg, Germany. Overall, Intel plans to invest (or co-invest with the government) Semiconductor co-investment partners like Brookfield) Approximately $100 billion in new semiconductor manufacturing facilities will be built in the next few years.
But Samsung is similarly aggressive in its CapEx spending. Intel recently cut capital spending to $25 billion from $27 billion in 2022, while Samsung said it will invest more than $33 billion in new semiconductor capacity this year and maintain spending at about the same level next year. The company recently announced Of course, it’s unclear how much of these dollars will be invested in memory (3D NAND and DRAM) production facilities and how much will be used to expand Samsung Foundry’s logic capacity, but the South Korean company is clearly very Aggressive. As such, it will be difficult for Intel to match his sci-fi prowess.
Stealing customers away from TSMC and Samsung Foundry is made even harder because large customers like Nvidia and Qualcomm have long-term supply agreements with their foundry partners. Additionally, it remains to be seen if Intel’s fabs in Europe and the US can offer the same prices as his TSMC and Samsung Foundry fabs in Taiwan and South Korea.
geographical advantage
It’s especially notable that Intel is ramping up production capacity in the US and Europe, and so far has not announced plans to build fabs in Taiwan or South Korea (Japan will continue to build Tower Semiconductor fabs). plans to operate). Operating fabs in Europe and the US is more expensive than in Taiwan or South Korea, but building new fabs in the US and Europe makes sense both from a customer relationship and geopolitical point of view. .
On the one hand, US customers are eager to use US fabs due to logistics, risk management and other factors. On the other hand, there are not many chip designers in Europe who need cutting-edge production technology. After all, Intel’s product portfolio is expanding, so a new fab will be needed anyway. Additionally, according to Intel, many potential foundry clients see an advantage in fabs located in the United States or Europe.
“Having worked with foundry customers since the launch of IFS, I know that many of these companies recognize the need for a more resilient and geographically balanced semiconductor supply chain. It’s clear enough,” Thakur said. Nikkei Asia.
However, with both TSMC and Samsung Foundry building new state-of-the-art fabs in Arizona and Texas, Intel’s new fab geographic advantage in the US may be overstated.
Western politicians want to build a domestic semiconductor supply chain that is less dependent on Taiwan, so they are keen to co-invest in new equipment with Intel.
make or break
For Intel, the foundry business is a way to ramp up production quickly, thus aligning its CapEx financial capabilities with those of TSMC and Samsung. Acquiring new revenue streams is essential for Intel, but not as important as increasing production.
Therefore, if the company succeeds in winning orders from many clients that require advanced production technology, it will be a success in itself, as it will continue to invest in developing cutting-edge nodes and increasingly expensive fabs. . Otherwise, the company could lose its position as an integrated design manufacturer (IDM) and a major supplier of CPUs. If the company manages to become the second largest foundry in the world in the process of launching the foundry business that pursues its main target, this will be an even greater success. But volume is the primary target.
