Nvidia is putting money down to secure the back end of its supply chain. The company announced a $1.1 billion advance to King Yuan Electronics (KYEC) to reserve capacity, with the funds earmarked for new semiconductor testing equipment in the United States. The move shifts attention to a question investors have circled for months: whether manufacturing and testing can keep pace with demand.
The agreements stretch across several links in the chain rather than a single production step. Beyond chip testing, Nvidia is also expanding substrate capacity and advanced packaging, an approach that targets multiple potential bottlenecks at once instead of betting on one fix.
Doubling tests, 2.5x substrates, 300% packaging
Total testing capacity is slated to double from calendar year 2026 to 2027, according to Nvidia. The advance to KYEC is what reserves that capacity — a distinction worth keeping in mind, since it does not signal completed facilities or extra output available right now. The timeline matters: these deals build out future capability, not an immediate ability to fill every order.
Substrate arrangements with suppliers call for capacity to reach 2.5 times current levels within three years. Advanced chip packaging is covered too, with Amkor targeting capacity growth of more than 300% in that segment by the end of 2028 — another date on the expansion calendar, and a target rather than reported additional production.
A $35 billion order book Nvidia won’t promise to fill
Why the extra capacity matters was underscored by a Reuters report from the previous day. ByteDance, Alibaba and Tencent together ordered more than three million RTX PRO 5500 graphics cards, with a combined list value above $35 billion. Nvidia warned customers and partners at the same time that it cannot guarantee fulfillment. Demand outruns planned production, so the list figure reflects neither secured deliveries nor revenue already booked.
For investors, the split is clean. Those orders demonstrate demand; the supply agreements announced today address the supply side. How much of the planned expansion actually turns into shipments will be settled in execution.
Should investors sell immediately? Or is it worth buying Nvidia?
Berlin backdrop: simulation software and RTX Spark laptops
The capacity news lands as Nvidia’s GTC Berlin runs from October 21 to 22, with a keynote from the chief executive. The conference offers a moment to weigh recent product announcements — expanded use of Nvidia technology in chip development and AI compute arriving on more notebooks — though technical progress and shipping products are not the same as quantified revenue.
On October 14, Nvidia introduced cuPotential, a library within CUDA-X aimed at AI agents simulating atomic and molecular interactions. Early tests sped up already GPU-optimized workflows by a further two to three times, according to the company. The gain comes from specialized software rather than new hardware, a distinction that matters when reading the numbers.
Separately, Nvidia pointed to Lam Research, where the use of NanoVDB accelerated simulations thirtyfold — a different application that should not be conflated with the cuPotential results. Nvidia also reported uptake of its CUDA-X libraries across the semiconductor industry for AI in chip design and manufacturing. These are technical benchmarks, not revenue or profit figures.
Late last week came word that RTX Spark notebooks are available, with ASUS, Dell, HP, Lenovo, Microsoft and MSI among the named providers. The systems offer up to 128 GB of unified memory and up to one petaflop of AI compute in FP4 format. Where cuPotential targets specialized simulation, RTX Spark brings comparable compute to portable Windows machines; the roster of manufacturers shows the breadth of the offering, not its sales.
The yardstick for Berlin stays the same: keep technical capability, product availability and commercial results in separate columns. Recent announcements give solid evidence for the first two. A number attached to business success is not among them.
Ad
Nvidia Stock: Buy or Sell?! New Nvidia Analysis from October 21 delivers the answer:
The latest Nvidia figures speak for themselves: Urgent action needed for Nvidia investors. Is it worth buying or should you sell? Find out what to do now in the current free analysis from October 21.
Nvidia: Buy or sell? Read more here...
