Musk Unveils 'Terafab': SpaceX and Tesla to Build Dual Chip Fabs in Austin

Elon Musk's announcement of two advanced chip factories in Austin marks a watershed moment for the semiconductor industry and the broader tech landscape. By bringing chip manufacturing in-house—or at least into captive, purpose-built facilities—Tesla and SpaceX are attempting to solve the compute bottleneck that has plagued the AI and automotive sectors for years. This move, referred to in internal circles as the 'Terafab' project, represents the ultimate expression of vertical integration, a strategy Musk has used to dominate the electric vehicle and space launch markets. By controlling the silicon from design to fabrication, Musk aims to bypass the global supply chain volatility that has historically hampered production timelines for his various ventures.

For years, Tesla has relied on partners like TSMC and Samsung to manufacture its custom-designed silicon, such as the Full Self-Driving (FSD) computer and the Dojo AI training chip. Similarly, SpaceX requires high-performance, radiation-hardened chips for its Starlink satellites and Starship navigation systems. By establishing dedicated factories in Austin, Musk is not just securing his supply chain; he is creating a feedback loop between chip design and manufacturing that could lead to unprecedented performance gains. This is particularly critical for the development of the Optimus humanoid robot, which requires massive processing power within a limited thermal and energy envelope—a feat that is significantly easier to achieve when the hardware and software are co-optimized at the transistor level.

“

By bringing chip manufacturing in-house—or at least into captive, purpose-built facilities—Tesla and SpaceX are attempting to solve the compute bottleneck that has plagued the AI and automotive sectors for years.

The decision to build two distinct facilities is highly strategic. The first, focused on cars and humanoid robots, addresses the need for high-efficiency 'edge' computing. This facility will likely produce the next generation of FSD hardware and the specialized inference chips required for Optimus to navigate complex human environments in real-time. The second factory, dedicated to AI data centers, aligns with the massive compute requirements of xAI and Tesla’s Dojo supercomputer. This dual approach suggests that Musk views silicon as the fundamental building block for all his future endeavors, from terrestrial transport to interplanetary exploration and the quest for artificial general intelligence.