After Samsung Foundry, Intel has managed to win an order for Tesla's supercomputers, as the company will now be responsible for packaging and testing.
Intel's EMIB Technology Acts as a Viable Alternative to TSMC's SoW, With Higher Modularity
It seems like Elon Musk's Tesla is actually looking to pivot away from TSMC by diversifying its sources for different products. Recently, we reported on how the automaker managed to secure a huge contract with Samsung Foundry for the next-gen AI6 chips, which will be built on the 2nm process. Now, based on a report by ZDNet Korea, it is claimed that Tesla has opted for Intel's packaging services for its Dojo supercomputer. This means that Intel Foundry might also bag a massive customer, allowing them to get a breathing space.
Interestingly, Samsung and Intel both have independent chip and packaging service offerings, so Tesla opting for both companies means that they like 'nitpicking' their partners. It is claimed that the next-gen Dojo 3 supercomputer might separate the "chip mass production and module packaging" while discussing supply chain partners, which is why both Samsung and Intel are expected to benefit, despite them being rivals in the segment.
One reason Tesla cannot opt for TSMC right now is that the custom D1 chips used in the Dojo supercomputer are large, with multiple chips arranged on a single wafer. Tesla previously utilized TSMC's SoW packaging to benefit from the redistribution layer, but given how populated TSMC's production lines are with CoWoS and its derivatives, it isn't viable for the company to satisfy limited orders from Tesla. However, since Intel is desperate for a breakthrough, they have showed readiness offering packaging services for Dojo chips.
Team Blue will likely rely on its EMIB (Embedded Multi-Die Interconnect Bridge) technology for Tesla's orders, which connects multiple dies using tiny bridges, which is an alternative to SoW. Moreover, it offers higher modularity, which means that Tesla could integrate additional functionalities or fine-tune existing ones as well, hence the inclination towards including Intel is pretty high for now.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
