Update, 18 November 2026, 00:39 UTC: Samsung Electro-Mechanics has now published its side. In a Korean-language newsroom release on 18 November (our translation), Samsung Electro-Mechanics says it will keep expanding supply of advanced package substrates for Nvidia’s next-generation GPUs and strengthen operations at its Sejong and Busan sites. It says Nvidia agreed in October to provide large-scale funding for dedicated substrate factory expansion at those two sites, and that Nvidia is now reviewing additional funding for more production infrastructure and equipment there, on the basis of which Samsung Electro-Mechanics will have full-scale mass production in place from 2029. The cooperation now also covers MLCCs (multilayer ceramic capacitors), which the release ties to stable power supply in high-density AI computing platforms. The release gives no amount for either the October funding or the additional funding under review. By our reading of the two timelines, Samsung Electro-Mechanics’ 2029 full-scale production comes after Kinsus’ additional capacity, which Nvidia’s release targets for the third quarter of 2028. ZDNet Korea reported the release at 00:29 UTC.
Kinsus plans to expand its advanced ABF package substrate capacity dedicated to Nvidia by more than 3.5 times over current levels under a new multiyear agreement, according to an Nvidia press release issued on 17 November alongside its third-quarter results. The release says production is expected to begin in the second half of 2027, with the additional capacity targeted for the third quarter of 2028.
Our number: the only dollar figure the release gives for these partner deals is Amkor’s “nearly $3 billion”, about 0.8% of the $390 billion of supply commitments in Nvidia’s third-quarter CFO commentary, by our arithmetic. The release does not say how much of the $390 billion these agreements represent.
Business Pill · Lead Time
An evergreen Business Engineer explainer on lead times. It teaches the general idea only and does not discuss Nvidia, its suppliers or this release.
The key insight: Nvidia named six manufacturing partners and a Kinsus plan to lift Nvidia-dedicated ABF substrate capacity more than 3.5 times by the third quarter of 2028; the release prices only Amkor’s nearly $3 billion, about 0.8% of Nvidia’s $390 billion of supply commitments, by our arithmetic.
What the Release Names
The release lists six partners: Amkor Technology, Foxconn, Kinsus Interconnect Technology, Samsung Electro-Mechanics, Siliconware Precision Industries (SPIL) and TSMC. It says Nvidia is expanding capacity “across substrates, advanced packaging, wafer and contract manufacturing”.
The Kinsus capacity would come from its Youth campus in Taoyuan, Taiwan. Nvidia and Samsung Electro-Mechanics “have amended their existing substrate supply agreement to improve operational capabilities at existing factories”, building on a collaboration covering substrate manufacturing in Sejong and Busan, South Korea.
On packaging, the release says Nvidia supports Amkor’s expansion in Arizona and South Korea through agreements announced in July and October 2026 “totaling nearly $3 billion”, including prepayments. It says SPIL is expanding advanced packaging in Taiwan “under a multiyear agreement supported by multibillion-dollar funding from NVIDIA”, increasing contracted capacity dedicated to Nvidia by mid-2029.

How It Compares With October
The release says October’s commitments with Amkor, IBIDEN, KYEC, Samsung Electro-Mechanics, SPIL and Unimicron included “an aggregate 2.5x increase in substrate capacity over three years”, a doubling of semiconductor testing capacity from 2026 to 2027 and a 15x increase in SPIL’s optical-engine capacity by mid-2027.
The Kinsus step of more than 3.5 times applies to one supplier’s capacity dedicated to Nvidia, while October’s 2.5 times was an aggregate across suppliers, so the two figures measure different things. Our packaging bottleneck explainer and the earlier packaging squeeze piece cover why substrates and packaging gate output; IBIDEN’s own plan is part of October’s round.
What Nvidia Said on the Call
On the third-quarter call, interim chief executive Colette Kress said: “With demand continuing to outpace supply, we are working closely with our partners to accelerate capacity expansion across critical components, including wafers, substrates, and advanced packaging”. The quotations from the call are from our transcription of the webcast audio.
The CFO commentary says supply commitments rose to $390 billion from $279 billion, “primarily related to the multi-year procurement of memory supply”. Our commitments piece compares that with Nvidia’s cash, and our results piece and call piece have the rest of the quarter.
The release also says the third quarter marked the first volume shipments of Vera Rubin NVL72 and Vera CPUs, built on TSMC’s 5nm, 3nm and CoWoS advanced packaging technologies.
The Structural Read
As we read the release, the new agreements cover the steps after the wafer: substrates at Kinsus and Samsung Electro-Mechanics and packaging at SPIL, on top of Amkor’s earlier agreements.
As we read the timelines, the capacity arrives in stages: Kinsus production from the second half of 2027 and its additional capacity in the third quarter of 2028, and SPIL’s contracted capacity by mid-2029.
Colette Kress, NVIDIA interim CEO, Q3 FY2027 earnings call
“With demand continuing to outpace supply, we are working closely with our partners to accelerate capacity expansion across critical components, including wafers, substrates, and advanced packaging”
Three Implications
FOR SUBSTRATE MAKERS The release names Kinsus and Samsung Electro-Mechanics for this round, after IBIDEN, Samsung Electro-Mechanics and Unimicron in October.
FOR PACKAGING HOUSES The release says Amkor’s agreements include prepayments and SPIL’s expansion is supported by multibillion-dollar funding from NVIDIA.
FOR NVIDIA SHAREHOLDERS The CFO commentary ties most of the rise in supply commitments to memory; this release covers other inputs and gives one dollar figure. This is not investment advice.
The Business Engineer Lens
This story maps onto the Business Engineer framework The Nine Layers of AI.
The framework says: “This entire ecosystem has been shaped by constraints, bottlenecks, chokepoints, and scarcity at every layer.”
As we read it, Nvidia’s release shows it paying to widen two of those constraints, ABF substrates and advanced packaging, with capacity that arrives from 2027 to 2029. The framework predates this release and does not discuss it.
What Is Not Established
The release gives no dollar value for the Kinsus, Samsung Electro-Mechanics, SPIL, TSMC or Foxconn arrangements, and does not say whether any of them includes prepayments other than Amkor’s. We did not find filings from Kinsus or Samsung Electro-Mechanics on the agreements at the time of writing. The 0.8% is our arithmetic on the one figure the release states. We did not contact Nvidia or its suppliers.
The Bottom Line
Nvidia named six manufacturing partners for its next expansion, led by a Kinsus plan to lift Nvidia-dedicated ABF substrate capacity more than 3.5 times by the third quarter of 2028. The only dollar figure in the release, Amkor’s nearly $3 billion, is about 0.8% of Nvidia’s $390 billion of supply commitments, by our arithmetic. This is not investment advice.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The partner agreements, capacity multiples, timelines and quotations from partners are from NVIDIA’s press release of 17 November 2026; the $390 billion and $279 billion supply commitments are from its third-quarter fiscal 2027 CFO commentary filed with the SEC on Form 8-K; the Colette Kress quotation is from our transcription of the earnings call webcast. The 0.8% is our arithmetic. This is not investment advice. We did not contact Nvidia or its suppliers.
Sources: Samsung Electro-Mechanics newsroom, 18 Nov 2026 · ZDNet Korea, 18 Nov 2026 · NVIDIA press release, NVIDIA Expands Commitments With Supply Chain Partners to Support AI Infrastructure Buildout, 17 Nov 2026 · NVIDIA CFO Commentary on Q3 FY2027 results, Ex.99.2 to Form 8-K, 17 Nov 2026









