Summary of Key Points
Elon Musk's SpaceX and Tesla have jointly launched the Terafab superchip factory in Texas, USA, with an initial investment of $16.8 billion. The goal is to create a computing foundation that will support Tesla's autonomous driving capabilities, humanoid robots, and SpaceX's AI infrastructure. To overcome the technical barriers in advanced semiconductor manufacturing, Tesla has recruited Gary Jiang, a 18-year veteran from Intel (with a background in materials science from Tsinghua University), to collaborate on using Intel's next-generation 14A manufacturing process. This move is part of Musk's strategy to reduce reliance on foundries like Samsung and TSMC by gaining control over chip production through vertical integration. However, building an advanced wafer factory is far more challenging than expected, as even giants like Intel and TSMC have experienced delays in their new facility constructions. There is still a long way to go before Terafab reaches mass production.
Detailed Analysis
1. Terafab: The “Heart of Musk’s AI Empire”
Terafab is not just another chip factory; it represents a critical step towards strengthening Musk's AI initiatives. In recent years, Tesla's autonomous driving (FSD chips), Optimus robots (edge computing), SpaceX's satellite AI systems (Starmind), and xAI model training have all required substantial computational power. Although Tesla attempted to develop its own AI training chips through the Dojo project, it failed in 2025 and had to rely on Samsung and TSMC for manufacturing.
Musk does not want to be at the mercy of others when it comes to chip production capacity and costs. Therefore, Terafab aims to achieve full control over the entire chip manufacturing process—from design to fabrication, packaging, and testing. In the future, it will serve as the “heart” that powers all of Musk’s AI initiatives.
2. Gary Jiang: The Right Expert for the Job
Why was Gary Jiang chosen? Building an advanced wafer factory is not primarily about money; the real challenge is transforming technology into a stable, operational production line. Jiang's background perfectly addresses this issue:
- Solid Education: Bachelor’s and Master’s in Materials Science from Tsinghua University, Ph.D. in Advanced Materials from The Ohio State University;
- Relevant Experience: 8 years at Rudolph Corporation in semiconductor equipment, followed by 18 years at Intel, where he participated in the construction of several wafer factories (including Fab68) and the implementation of the 18A manufacturing process, focusing on technology transfer and equipment deployment for mass production.
In short, Jiang possesses the expertise to transform a chip design into a production line that can produce thousands of chips daily. He acts as a bridge between SpaceX and Intel, ensuring the successful integration of their technologies.
3. Collaboration with Intel: Leveraging Expertise
Why choose Intel? Advanced manufacturing processes are not easily attainable on one’s own. Tesla enlisted Intel’s help:
- Using Intel’s 14A Process: This is the next-generation process after 18A, representing even greater technological advancement (the smaller the number, the more advanced the process);
- Intel’s Manufacturing Support: Intel will provide technical assistance and equipment calibration.
For Intel, this represents a new market opportunity to supply Tesla’s AI needs; for Tesla, it allows them to access cutting-edge manufacturing technology without starting from scratch.
4. The Challenges of Building an Advanced Wafer Factory
Although $16.8 billion is a significant investment, the challenges of building an advanced wafer factory go beyond funding:
- Intel’s Ohio Plant: Initially planned for 2025 production, it was delayed to 2030-2031;
- TSMC’s Arizona Plant: Launched in 2020 with a $12 billion investment, postponed to 2025 due to worker and equipment issues;
- Samsung’s Texas Plant: Started in 2021 with a $17 billion investment, delayed to 2026.
These giants have all encountered delays because building advanced wafer factories requires a stable supply chain of equipment (e.g., lithography machines), experienced engineering teams, and mature manufacturing processes (such as improving chip yield). These factors cannot be quickly resolved by money alone; they require time and effort.
5. The Future of Terafab: Can Musk Repeat SpaceX’s Success?
Musk has previously succeeded with vertical integration—creating rocket engines and reusing rockets to break NASA’s monopoly, as well as developing batteries and software that transformed the electric vehicle industry. However, semiconductor manufacturing is different:
- Higher Barriers: Chip production involves hundreds of precise steps;
- Strong Industry Interdependencies: It requires collaboration with equipment manufacturers (e.g., ASML), material suppliers (e.g., photoresists), and engineering teams.
With funding, a location, partners, and Gary Jiang, Terafab still faces numerous challenges, such as equipment procurement, team building, and yield improvement. Whether Musk can replicate SpaceX’s success and create another “TSMC” remains to be seen.
Conclusion
Terafab marks another bold attempt by Musk in the semiconductor manufacturing sector. Success would grant him control over the core computing power of the AI era, while failure could result in significant financial losses. Regardless of the outcome, this move is a crucial step in the battle for future AI dominance among tech giants.