Tesla Inc (NASDAQ:TSLA)'s reported plans to build advanced semiconductor facilities with SpaceX mark a significant step toward addressing what Wedbush analysts see as the “largest bottleneck” in the company’s artificial intelligence (AI) ambitions.
The project involves two chip fabrication plants in Austin, Texas, under a broader “Terafab” initiative aimed at scaling production for Tesla’s vehicles, Optimus humanoid robots, and space-based AI infrastructure.
Wedbush wrote that “current suppliers, including Micron, TSMC, and Samsung, are unable to meet future demand for TSLA’s AI strategy,” highlighting supply constraints as a central driver behind the move.
One facility is expected to support Tesla’s automotive and robotics operations, while the other would focus on chips for orbital AI data centers. The Terafab will produce inference chips for Tesla applications and D3 chips designed for satellites, with development targeting a 2-nanometer process.
“80% of the Terafab’s output is expected to be on computer output for space-based orbital AI satellites and the other 20% focused on ground-based applications,” Wedbush wrote.
The analysts noted that chips designed for space will require different specifications, including the ability to operate under harsher conditions. Musk indicated these chips would function at higher temperatures to “minimize radiator mass,” which Wedbush described as a key constraint in space-based chip performance.
The total cost of the project is estimated at up to $25 billion, with construction planned at Tesla’s Giga Texas site. Wedbush noted that the facility would be “the largest semiconductor fab in human history,” covering chip design, fabrication, memory production, advanced packaging, and testing.
Initial production is projected at 100,000 wafer starts per month, scaling to as much as 1 million per month over time, equivalent to roughly 70% of TSMC’s current global output. Annual output is expected to reach between 100 billion and 200 billion custom AI and memory chips. The analysts noted that this level of spending is not included in Tesla’s current $20 billion capital expenditure guidance. However, Musk has suggested the company may need to build domestic chip capabilities to mitigate geopolitical risks.
“TSLA would have to build its own in-house Terafab, including logic, memory and packaging domestically to protect against any geopolitical risks,” Wedbush wrote.
While the timeline remains uncertain, the analysts said the initiative is expected to accelerate Tesla’s broader AI ambitions and help alleviate supply constraints. The analyst believe that “this will accelerate the company’s ambitious AI path,” adding that it could position Tesla as a larger player in AI over time.
The firm maintained its ‘Outperform’ rating on Tesla shares with a $600 price target, implying significant upside from current levels of $385.