Samsung Electro-Mechanics and Nvidia partner on ABF substrates for AI chips

NVIDIA’s supply chain plans sit inside the wider buildout of AI hardware, where advanced chips depend on a chain of specialized manufacturing steps rather than only on chip design. Substrates, packaging, testing, and optical-engine capacity are all part of the infrastructure needed to turn demand for AI systems into shipped hardware. The reference to 2028 shows that this is a multi-year production plan, not a one-off procurement move.

For chip builders and their suppliers, these agreements point to more structured access to the parts of the manufacturing process that often determine how quickly AI systems can be assembled and delivered. Samsung Electro-Mechanics’ role in ABF substrates places a major component supplier directly into NVIDIA’s AI-chip supply chain. That can reshape who participates in the hardware stack and how much coordination is needed across vendors.

The story also shows how AI competition is now extending beyond model software into industrial capacity, materials, and packaging. Companies that can secure substrate and testing capacity may be better positioned to meet large-scale orders, while firms without similar access face tighter bottlenecks. The trade-off is a more resilient supply chain on one side and a more concentrated one on the other, with fewer critical suppliers carrying more of the load.

Still open is how quickly these capacity deals will translate into output and whether they are enough to meet demand through 2028. The two announcements identify the partners and the technical areas involved, but not the volume, timing, or contract structure behind them. More detail on those points would show how deep NVIDIA’s expansion really reaches and how much Samsung Electro-Mechanics will be tied into it.