NVIDIA Expands Commitments With Supply Chain Partners to Support AI Infrastructure Buildout

Glass Epoch DeskOriginal digestSource: NVIDIA Newsroom

AI-assisted digest, checked against the original source before publishing. How we work

NVIDIA is locking in substrate, packaging, wafer and system manufacturing capacity with six partners as it scales current and next-generation AI platforms.
NVIDIA Expands Commitments With Supply Chain Partners to Support AI Infrastructure Buildout
Photo: Brecht Corbeel · free stock license

NVIDIA announced expanded commitments spanning six supply chain partners — Amkor Technology, Foxconn, Kinsus Interconnect Technology, Samsung Electro-Mechanics, Siliconware Precision Industries (SPIL) and TSMC — to align manufacturing capacity with demand for its accelerated computing and AI infrastructure platforms. The agreements cover substrates, advanced packaging, wafer production and contract manufacturing, and are positioned by the company as long-term supply agreements meant to support sustained growth rather than a single product cycle.

On substrates, the most concrete figure concerns Kinsus, which plans under a new agreement to grow advanced ABF package substrate capacity dedicated to NVIDIA by more than 3.5x over current levels, built out at its Youth campus in Taoyuan, Taiwan. Samsung Electro-Mechanics is taking a different route: its agreement adds substrate capacity and improves production capabilities at existing factories rather than standing up new sites, a pattern that suggests NVIDIA is pushing volume through plants already qualified for its parts.

For advanced packaging, NVIDIA says its work with Amkor and SPIL expands capacity for both current and next-generation accelerated computing platforms, without disclosing output targets. Debora Shoquist, NVIDIA's executive vice president of operations, framed the supplier push as a competitive edge: "Our supply chain is a strategic advantage," she said, adding that the partnerships build on the company's manufacturing foundation across components, advanced packaging and system production.

Downstream, NVIDIA reports that Vera Rubin NVL72 systems and Vera CPUs began volume shipments in the third quarter, drawing on TSMC's 5nm and 3nm process nodes alongside TSMC's CoWoS advanced packaging. Foxconn is deepening its collaboration with NVIDIA, tapping a global production network to support next-generation GPU and networking platforms, which puts chassis, rack assembly and system integration in scope alongside the silicon-side inputs.

The announcement leaves several open questions for anyone tracking the buildout. NVIDIA did not disclose contract values, agreement durations, or capacity targets for Samsung Electro-Mechanics, Amkor, SPIL, Foxconn or TSMC — so the Kinsus 3.5x figure is the only benchmark against which progress can be measured from the outside. Also unresolved is timing: how much of the expanded substrate and packaging capacity comes online after the third-quarter Vera Rubin volume ramp, and whether new sites are involved or, as with Samsung Electro-Mechanics, the work stays inside existing factories.

Key facts

  • • Six partners named: Amkor, Foxconn, Kinsus, Samsung Electro-Mechanics, SPIL and TSMC.
  • • Kinsus to grow NVIDIA-dedicated ABF substrate capacity by more than 3.5x.
  • • Kinsus expansion located at its Youth campus in Taoyuan, Taiwan.
  • • Vera Rubin NVL72 and Vera CPUs began volume shipments in Q3.
  • • Vera platforms use TSMC 5nm, 3nm and CoWoS packaging technologies.

Why it matters

Builders should read this as capacity being pre-committed well ahead of demand signals, with substrates and advanced packaging — historically the tightest links in the AI hardware chain — getting contractual guarantees rather than spot orders. Expect future platform launches, including Vera Rubin rollouts, to be gated more by these suppliers' ramp schedules than by GPU design readiness, and watch

Spotted an error? Send a correction.

Get the day's AI news in one email

The Daily Epoch · 60-second read · unsubscribe anytime

Related