In a sign of how the AI hardware race is moving beyond just designing chips, a South Korean business newspaper reports that Tesla and Nvidia are planning to pour 3.7 trillion won (roughly $2.7 billion) into Samsung Electro-Mechanics to expand production of advanced chip packaging materials. The investment would target the company's Sejong facility, which makes FC-BGA substrates—a key component used to connect high-end AI chips to the rest of a system.
According to Maeil Business Newspaper, Tesla would contribute 1.1 trillion won and Nvidia 2.6 trillion won, with the new production lines tied to a supply contract running from 2029 through 2033. Samsung Electro-Mechanics told MT Newswires it cannot share client details, and neither Tesla nor Nvidia immediately commented. But the structure of the reported deal—customers funding capacity years in advance—is raising eyebrows across the industry.
What are FC-BGA substrates and why do they matter?
FC-BGA stands for flip-chip ball grid array. It's a type of substrate that sits underneath a chip and acts as a bridge between the silicon and the circuit board, carrying thousands of tiny electrical connections. For high-performance chips used in AI data centers, these substrates are essential. They're also notoriously difficult to manufacture: the process requires extreme precision, and the specialized equipment takes time to install and calibrate.
That difficulty is why the reported investment is significant. Rather than simply placing a purchase order, Tesla and Nvidia would be effectively pre-funding new production capacity years before they need it. This kind of arrangement is rare and suggests that the companies are worried about future shortages of a component that's hard to substitute.
The move fits a broader trend in the semiconductor industry. As AI models grow more complex, the packaging that holds chips together—and the materials used in that packaging—are becoming as strategic as the chips themselves. Companies like Samsung Electro-Mechanics, which specialize in these substrates, are finding themselves in a position of unusual leverage.
What it means for investors
For markets, a 3.7 trillion won commitment is a signal that FC-BGA capacity is being reserved far ahead of actual demand. If the reported timeline holds, Samsung Electro-Mechanics would gain something manufacturers value highly: long-dated volume commitments. These can make factory utilization steadier and reduce the risk that new capacity sits idle, which is a common worry when companies build out expensive production lines.
For Tesla and Nvidia, putting money up front is a way to lock in a hard-to-substitute input during a period when competition for AI-related components could be intense. The 2029-2033 window is far enough out that it suggests the companies are planning for sustained growth in AI computing, not just a short-term spike.
The bigger takeaway for everyday investors is where pricing power sits in the AI supply chain. When customers feel they need to pre-fund a supplier's expansion, it's usually a sign that the bottleneck is the component supplier, not the end-market brand. That could mean companies like Samsung Electro-Mechanics are in a stronger negotiating position than their customers, which may translate into better margins over time.
This isn't the first time AI-related funding concerns have rattled markets. Recent weeks have seen AI funding worries rattle tech stocks, and investors are watching how companies plan to pay for the massive infrastructure AI requires. The reported Samsung deal suggests that at least some major players are willing to commit capital well ahead of need, which could ease fears about future supply constraints.
Still, it's worth noting that the report is unconfirmed. Samsung Electro-Mechanics declined to comment on specific clients, and Tesla and Nvidia have not issued statements. Investors should treat the news as a strong signal rather than a done deal.
For those tracking the AI trade, the story reinforces a key theme: the value in AI is increasingly shifting to the companies that make the physical components—substrates, packaging, and advanced materials—rather than just the chip designers. As AI demand hits funding walls, the companies that control scarce inputs may be the ones that benefit most.
For now, the reported investment is a reminder that the AI boom isn't just about software and models. It's also about the unglamorous, precision manufacturing that makes those models possible. And when the world's biggest tech companies are willing to write big checks years in advance, it's worth paying attention.
