Paradoxical change at Intel: hire another partner to produce chips, while the chip production segment receives outsourcing for other partners

If Intel was once known for its ability to design and manufacture those chips on its own, now the tremendous growth in chip manufacturing capacity of other competitors is forcing the company to transform. A series of recent reports show that, while Intel’s chip design department is outsourcing chip processing, Intel’s foundry division is moving to produce ARM chips for external partners.

In early April, a report from the news site ctee.com.tw said that, according to sources close to TSMC and Intel, the US technology giant is preparing to use TSMC’s chip foundries. for outsourcing Battlemage and Celestial generation GPUs. More specifically, Intel’s Battlemage series of GPUs will be manufactured on TSMC’s 4nm process in the second half of 2024, while the Celestial series of GPUs will be manufactured on the 3nm process in the second half of 2026.

Although Intel has invested heavily in the development of 4nm process and wafers using more advanced manufacturing processes, the company’s discrete GPU product line continues to maintain cooperation with chip manufacturers with TSMC is developing. is the sole partner.

The reliability of this report is quite high when TSMC also processed Intel’s first GPU line, code-named Alchemist (Xe). These GPUs were launched in mid-2022 but only started to hit the shelves from the second half of 2022. These chips all use TSMC’s 6nm process and therefore, Intel’s next generation of GPUs continue to rely on TSMC. is understandable.

While Intel’s GPUs are being processed by TSMC, Intel’s foundry division itself is working with chip designer ARM to be able to produce processors according to this architecture for external partners. other than Intel.

The deal will help Intel’s foundry division not only manufacture x86 processors, but also manufacture processors for mobile devices from other partners. The division’s target is customers like Apple, Nvidia, which are dependent on TSMC and Samsung to produce chips for them.

The deal will help partners manufacture their ARM chips on Intel’s upcoming 18A (1.8nm) technology production line, which is expected to begin mass production by the end of 2024. Therefore, it is not clear whether this line will be used to produce chips for Intel itself.

Intel’s 18A technology line is said to be able to compete with TSMC’s upcoming 2nm process, which is rumored to launch in 2025. Besides, Intel is also planning to build foundry factories. new in both the US and Europe.

Current Intel and ARM both promise that the 18A process will deliver “a breakthrough in transistor technology for improved power and performance”. Both companies are planning to cooperate to optimize ARM chip design for Intel’s new chip manufacturing technology. Both companies said that the partnership will initially focus on mobile processor SoC designs, but “could also open up other designs for autonomous cars, IoT, hubs.” data, aerospace and government applications”.

This new deal also opens the door to make it easier for Intel to attract customers for its foundry business, which is lagging behind two big rivals, TSMC and Samsung.

Refer to PCMag, Wccftech