[News] Intel Reportedly to Outsource More Arrow Lake Orders to be Manufactured by TSMC’s 3nm

---

---

In order to counter the competition from chip rivals such as AMD and NVIDIA, Intel reportedly plans to scale up its outsourcing efforts by handing over to TSMC more 3nm orders for its upcoming Lunar Lake and Arrow Lake chipsets in 2025, according to industrial sources cited by Commercial Times.

TSMC will continue to secure a large volume of outsourced business from IDMs, maintaining a strong cooperative relationship with Intel, acording to the Commercial Times report.

According to the report, Intel aims high for Arrow Lake as the chipset will feature two TPUs (Tensor Processing Units), allowing it to maintain high performance and high clock speeds while reducing power consumption by at least 100 watts. As the result, the product is regarded by Intel as a critical advantage for maintaining its lead in the AI PC market.

Intel’s Arrow Lake, its 15th generation CPU, features significant changes in both architecture and manufacturing process, along with a new name—Core Ultra 200S, according to its press release.

According to the Commercial Times, the processor is not only built using TSMC’s 3nm process, with a substantial reduction in computational core area and energy consumption, but also moving away from the traditional SoC design by adopting Intel’s exclusive 3D Foveros technology.

Foveros, Intel’s 3D advanced packaging technology, is a first-of-its-kind solution that enables the building of processors with compute tiles stacked vertically, rather than side-by-side, according to its press release. The focus of this new design is on energy efficiency, reducing packaging power consumption, and enhancing multi-core performance, the report notes.

According to the supply chain sources cited by the Commercial Times report, the Intel 7-Series chipsets in the 13th and 14th generations, in spite of adopting an 8P+16E (8 Performance-core and 16 Efficient-core) core configuration, the compute tile area still accounted for as much as 70% of the total chip area. However, by switching to TSMC’s 3nm process, the same core configuration now takes up only a third of the total area, while allowing the space for an additional NPU unit.

Though Intel has not given up its foundry unit, the struggling giant does seem to gradually loose competitiveness in advanced nodes, and it has outsourced several products to TSMC. The company’s latest flagship AI processor, Gaudi 3, is fabricated with TSMC’s 5nm.

Recently, in order to reduce costs and better prepare for its in-house 18A process, Intel has decided to abandon the introduction of the 20A node, and leverages TSMC’s process for the Arrow Lake chipset.

Read more

• [News] Intel Impresses the Market with Upbeat Q4 Guidance; 18A to Begin High-Volume Production in 2H25
• [News] Intel Shifts to TSMC for Chip Production: A Looming Threat for AMD?

(Photo credit: Intel)