Nvidia Locks In 3.5x More Chip Substrate as Rubin Ramps Up

Human Verified 0

Share to: , and

A green computer chip with gold pins on a circuit board inside a semiconductor packaging cleanroom

Nvidia expanded its supply chain commitments on Tuesday, signing multiyear deals to add chip substrate and advanced packaging capacity as demand for its AI platforms keeps climbing. Kinsus will grow Nvidia-dedicated substrate output more than 3.5 times, while packaging partner SPIL gets multibillion-dollar backing.

The Substrate Deals

Under the new Kinsus agreement, advanced ABF substrate capacity dedicated to Nvidia will rise more than 3.5 times from current levels at its Youth campus in Taoyuan, Taiwan. Production should start in the second half of 2027, with the added capacity targeted for the third quarter of 2028.

Samsung Electro-Mechanics amended its existing substrate supply agreement to upgrade operations at its factories in Sejong and Busan, South Korea. The change builds on a multiyear collaboration covering substrate manufacturing for Nvidia’s compute and high-speed networking technologies, which feed the same AI systems now in demand.

Packaging and the Rubin Ramp

On packaging, Nvidia and SPIL signed a multiyear deal backed by multibillion-dollar Nvidia funding to expand advanced packaging operations in Taiwan. Contracted packaging capacity dedicated to Nvidia is set to increase by mid-2029, with new lines for bumping, interposer fabrication, assembly and testing.

Amkor is expanding advanced packaging in Arizona and South Korea under multiyear agreements announced in July and October, totalling nearly $3 billion including prepayments. The release also notes that Vera Rubin NVL72 systems and Vera CPUs began volume shipments in the third quarter, using TSMC’s 5nm, 3nm and CoWoS packaging.

Our Read

This is Nvidia buying certainty, not capacity for its own sake. Substrates and advanced packaging are the quiet choke points of the AI buildout: a chip that cannot be packaged cannot be sold, however fast its design. Locking in dedicated lines years ahead removes that risk from the Rubin ramp.

The timing tells the same story as the earnings. Nvidia guided to $130 billion in fourth-quarter revenue and raised its 2028 growth outlook to 75% on the same day — promises that only hold if factories can deliver. These deals are the supply-side half of that guidance.

Why It Matters

For the AI industry, the deals lower the odds that packaging shortages cap the buildout in 2027 and 2028. Customers from cloud giants to frontier labs have watched lead times stretch; dedicated capacity means Nvidia’s biggest orders move first when supply is tight.

For rivals, the gap is structural, not just technical: Nvidia is securing factory floors years out while competitors still fight for the same substrate and packaging lines. For buyers, the message is that the AI hardware race is won as much in the supply chain as in the lab.

Background

The deals build on an October package with Amkor, IBIDEN, King Yuan Electronics, Samsung Electro-Mechanics, SPIL and Unimicron, which targeted a 2.5-fold substrate capacity increase over three years and a doubling of semiconductor testing capacity by 2027. They arrived on the day Nvidia reported record quarterly revenue of $115.8 billion.

Nvidia’s operations chief, Debora Shoquist, called the supply chain ‘a strategic advantage’ in the announcement. The company has been run since November 12 by interim chief Colette Kress, after Jensen Huang resigned, and the board has ruled out a Huang return.

Source: NVIDIA’s supply chain announcement.

Share to: , and

Faheem R. Khattak

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.