Musk Eyes 'Gigantic Chip Fab' for Tesla AI, Opens Door to Potential Intel Partnership

date
15:37 07/11/2025
avatar
GMT Eight
CEO Elon Musk announced Tesla may build a "gigantic chip fab" to secure enough AI chips for its autonomous and robotics ambitions. He suggested a possible partnership with Intel, whose shares rose 4% on the news. The proposed fab would produce at least 100,000 wafer starts per month. Musk aims for the in-house chips to be significantly more efficient than Nvidia's, costing only more than 10% to produce.

At Tesla’s annual shareholder meeting on Thursday, CEO Elon Musk suggested the company may need to build a massive semiconductor fabrication plant — a facility he dubbed a “terafab” — to produce the enormous number of artificial intelligence (AI) chips required for Tesla’s self-driving and robotics ambitions. He also hinted that Intel could become a potential partner in this effort.

Musk explained that Tesla is already developing its fifth-generation AI processor, known as the AI5, which will power the next stage of the company’s autonomous systems. Despite optimistic supply projections from existing partners, Musk said the chip volume would still fall short of Tesla’s long-term requirements.

“I think we may have to do a Tesla terafab… I can’t see any other way to get to the volume of chips we’re looking for,” he told shareholders.

Although Musk did not outline a construction timeline or location, he said the envisioned facility could handle at least 100,000 wafer starts per month, a key industry measure of chip production capacity.

His comments appeared to influence investor sentiment: Intel’s stock rose roughly 4% in after-hours trading. Intel — which operates its own chip foundries and recently secured a significant U.S. government investment — has been seeking strategic collaborations to advance its AI manufacturing capabilities. The company did not immediately comment on Musk’s remarks.

“Maybe we’ll do something with Intel,” Musk added. “We haven’t signed any deal, but it’s probably worth having discussions.”

Currently, Tesla relies on TSMC in Taiwan and Samsung in South Korea to manufacture chips for its Full Self-Driving (FSD) systems. Musk confirmed that limited AI5 chip production is scheduled for 2026, followed by broader output in 2027. The next-generation AI6 chip is expected around mid-2028, offering nearly double the processing performance using the same manufacturing framework.

According to Musk, the custom Tesla-designed chips will be optimized for the company’s software, providing significant gains in energy efficiency and cost. He estimated that the upcoming chip could operate on roughly one-third the power consumption of Nvidia’s flagship Blackwell processor while costing just over 10% of its production price.

The ambitious manufacturing vision coincided with Tesla shareholders voting to approve Musk’s controversial pay package, which could be worth up to $1 trillion over the next decade. The vote further reinforces Musk’s push to transform Tesla from an electric vehicle manufacturer into a global leader in AI and robotics.