Elon Musk hasn't dropped his plans to build Tesla's very own chip network at all, and instead he 'doubled down' on his 'TeraFab' commitment in the latest earnings call.
Elon Musk Claims Supply Constraints & Geopolitical Tensions Ultimately "Force" The TeraFab Project
Tesla's CEO has indicated the company's intention to enter the foundry business during the company's last shareholder meeting, where Musk revealed that, to ensure Tesla's custom silicon gets 'widespread' adoption, the company would need sufficient production capacity. Musk has shared a vision of having a fab network that could output a whopping 100 billion to 200 billion chips per year, and now, in his latest comments, the to-be trillionaire wants TeraFab to handle "logics, memory, and packaging" all in one go, to ensure that Tesla doesn't rely on any external partner.
One of the things I'm trying to figure out is; how do we make enough chips? To eliminate constraints that are likely to arise in the next 3 to 4 years, we must build a gigantic chip fab, Tesla Terafab. It must be a very large-scale domestic production facility that includes logic, memory, and packaging.
I can't see any other way to get to the volume of chips that we're looking for.
- Elon Musk
The idea of a fabrication venture from Tesla doesn't sound odd at all, given how massive a bottleneck the semiconductor supply chain is becoming for fabless manufacturers. Musk has indicated an extensive roadmap for Tesla's custom chips, scaling up from the existing AI5 to AI9, and the idea is to achieve cost-effective compute, something Tesla's CEO finds necessary for this venture to turn out as needed. And, for those who think Musk doesn't have sufficient resources onboard for the chip business, well, here's what he has to say:
That’s crazy fabs are really hard! Yes I know they are really hard, I didn’t think they are easy, but we do a lot of hard things. It would be crazy to not try a Terafab.
While we cannot comment on the demand Tesla sees for custom chips, Musk has previously said that supply constraints imposed by existing partners like TSMC, Micron, and Samsung, along with geopolitical uncertainties in the chip industry, make the TeraFab project much more appealing. However, as argued previously in a post, creating a chip network involves several key elements that make this venture much more difficult, such as talent, equipment sourcing, capital, and more. If Tesla wants to have a 'standalone' chip business, it will take a lot.
The other prospect often talked about in the industry is that Tesla teams up with an existing foundry partner, providing them with the necessary resources, mainly capital, to build up chip facilities. We already know that TSMC has opened the option for customers to invest in future production lines to get early access, so Musk could cut down on the overhead involved in maintaining 'exclusive' supply lines.
Considering that Dojo 3 is back and Tesla's CEO has repeatedly talked about owning a fab, it seems the idea is under serious consideration by Musk, which means that, apart from physical AI and autonomous vehicles, Tesla might be diving into the semiconductor business as well.
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.
