Tesla Partners with Samsung for AI5 Chip Production as Musk Claims Korean Fab Outperforms TSMC's Arizona Facility

Tesla Expands Chip Manufacturing Partnership

Tesla CEO Elon Musk announced during the company's third-quarter earnings call that the electric vehicle manufacturer will dual-source its next-generation AI5 processor from both Taiwan Semiconductor Manufacturing Company (TSMC) and Samsung Foundry 1

1

. This represents a significant strategic shift from Tesla's previous plan to have TSMC exclusively manufacture the AI5 chips 2

2

.

The decision to implement a dual-sourcing strategy stems from Tesla's ambitious plans for the AI5 chip, which will power a diverse range of applications including autonomous vehicles, robots, and data centers. "I am confident [the AI5] will be winner, next level [of AI performance], so it makes sense to have both Samsung and TSMC focused on AI5," Musk stated during the investor call 1

1

.

Samsung's Texas Facility Gains Recognition

In an unusual comment from a fabless chip designer, Musk praised Samsung Foundry's fabrication facility in Taylor, Texas, claiming it possesses more advanced equipment than TSMC's Fab 21 phase 1 in Arizona. "Technically, the Samsung fab [in Texas] has slightly more advanced equipment than the TSMC fab [in Arizona]," Musk explained 1

1

.

The technical basis for this claim lies in the manufacturing capabilities of each facility. Samsung's Taylor fab is designed to produce chips using 3nm-class process technologies, including the company's SF2 node, which was previously designated as SF3P but was reclassified due to its performance advantages 1

1

. In contrast, TSMC's Fab 21 phase 1 was designed to manufacture chips on 5nm and 4nm-class production nodes, making it less advanced by design.

AI5 Chip Specifications and Performance Expectations

The AI5 processor represents a substantial advancement over Tesla's current AI4 chip, with Musk claiming performance improvements of up to ten times the previous generation 3

3

. The chip is designed to consume up to 800 watts at peak performance, significantly higher than the AI4's 140-watt consumption, while fitting within a "half reticle" size of approximately 430mm² 1

1

.

Tesla's design team has made significant architectural changes to the AI5, removing the legacy GPU block and focusing die space on low-precision tensor compute hardware to maximize performance and cost efficiency 1

1

. This specialized approach allows the chip to excel in AI-specific workloads across Tesla's expanding product ecosystem.

Strategic Implications for Samsung Foundry

This partnership represents a major victory for Samsung Foundry as it competes against TSMC for market share in the advanced semiconductor manufacturing space. The company had previously secured a $16.5 billion deal to manufacture Tesla's future AI6 chips, and the addition of AI5 production further solidifies the relationship 2

2

.

Samsung's financial performance has benefited significantly from AI-driven demand, with the company expected to report its largest quarterly profit in more than three years. The surge in demand for AI hardware has boosted sales of Samsung's High-Bandwidth Memory (HBM) products while constraining supply and driving up prices for conventional memory chips 2

2

.