Tesla (NASDAQ:
TSLA) and Nvidia (NASDAQ:
NVDA) are investing in new production capacity for AI chip substrates at Samsung Electro-Mechanics in South Korea, according to a report by Korea's Maeil Business Newspaper, as cited by Investment Monitor on October 22, 2026.
What was reported
According to the report:
- Tesla is investing about 1.1 trillion won, or about $830 million, as part of a 4.27 trillion won investment programme by Samsung Electro-Mechanics in FC-BGA substrates in South Korea
- the Tesla-backed line in Sejong will make substrates for AI chips used in Tesla's vehicles and its Optimus humanoid robot
- Tesla will have priority access to output under a five-year agreement covering 2029 to 2033
- Nvidia is investing about 2.6 trillion won to expand capacity at Samsung Electro-Mechanics' sites in Sejong and Busan, for substrates used in data centre networking and AI computing chips
- construction is expected to be completed by the end of 2027, with test production in the first half of 2028 and commercial production from the second half of 2028
Nvidia disclosed the partnership at its GTC conference in Germany on October 21 but did not publicly reveal the investment amount. Tesla and Samsung Electro-Mechanics have not publicly confirmed the reported figures.
What substrates are
A chip substrate is the base on which a processor is mounted. It connects the tiny circuits of the chip to the larger circuit board of a device and helps manage power and heat. FC-BGA, or flip-chip ball grid array, substrates are used for high-performance chips, including processors for AI. As AI chips become larger and more powerful, they need more advanced substrates, and supply of these components has become an important part of the AI supply chain.
Why chip buyers invest in suppliers
Large chip buyers sometimes invest directly in suppliers to secure capacity for key components. In return, they may receive priority access to production for several years. This approach can help companies plan production of their own products, but it requires large upfront commitments.
Context for Nvidia
On October 21, Nvidia also announced capacity expansion plans with several supply chain partners, including Samsung Electro-Mechanics, Amkor, IBIDEN, KYEC, SPIL and Unimicron. Nvidia designs the graphics processors used to train and run AI models, and it relies on partners to manufacture and package its chips.
Context for Tesla
Tesla designs its own AI chips for its driver assistance and self-driving software and for its Optimus robot. According to the report, Samsung Electronics' foundry business is expected to manufacture Tesla's next-generation AI chip. Securing substrates is one step in building a supply chain for these chips.
Competition among suppliers
Samsung Electro-Mechanics competes with other substrate makers, including Ibiden in Japan and Unimicron in Taiwan. Demand for advanced substrates has grown with the expansion of AI data centres, and several suppliers have announced new investments.
Timeline of the project
According to the report, the project will move through several stages. Construction of the new facilities is planned to finish by the end of 2027. Equipment will then be installed and test production will take place in the first half of 2028, followed by commercial production in the second half of 2028. Deliveries to Tesla are planned to begin in 2029.
What to watch
Key items include any official confirmation of the investment amounts, progress on construction, and comments from Tesla and Nvidia on their supply chains in upcoming results.
Risks and uncertainties
The figures come from media reports and have not been fully confirmed by the companies. Construction and production plans may change, and new capacity may take longer to come online than expected. Technology shares can be volatile.
What this means for investors
For Investing.one users following TSLA or NVDA, developments in their AI chip supply chains may be relevant. The information above is based on media reports.
Sources
This is not investment advice. Your capital is at risk.