THE COMPONENT SIGNAL

Issue #006 · Monday, February 17, 2026 · Memory + Metals + Design

• •

  Samsung Pyeongtaek Line 4 commissioning delayed to Q3. Equipment installation is running 6–8 weeks behind schedule due to ASML EUV tool delivery constraints. Line 4 was supposed to bring 30K wafer starts/month of advanced DRAM and HBM4 capacity online by Q2. The delay tightens an already sold-out HBM4 supply picture. SK hynix's Cheongju M17 fab remains on track — giving hynix a temporary capacity advantage in the HBM race. Source: ETnews, DigiTimes

• •

  LG Energy Solution expanding Michigan battery plant, sourcing Korean components. The Ultium Cells expansion in Holland, Michigan will consume more Korean-origin power electronics — including EMI shielding for BMS modules. Tier 1s supplying into this plant should expect Korean-origin component audits as the tariff situation evolves. Source: Korea JoongAng Daily

• •

  Hyundai Mobis opens new ADAS R&D center in Seoul. 500 engineers focused on L3+ autonomous driving. Every ADAS module — radar, camera, V2X — needs board-level EMI shielding qualified to CISPR 25 Class 5. This is net new shielding demand that didn't exist in Hyundai's previous-gen platforms. Source: Hyundai Mobis press release

Metals

Indonesia nickel update: The Indonesian parliament's Commission VII passed the nickel ore export tax bill out of committee on February 12. Final floor vote expected by March. If enacted, the 10–15% export tax on processed nickel products would structurally raise input costs for any shield alloy containing nickel — including nickel-silver (typically 18% Ni), mu-metal (80% Ni), and nickel-plated steel.

If you currently spec nickel-bearing shield materials, flag this to procurement now. We've already had two customers ask us to re-quote nickel-silver shields in tin-plated CRS as a hedge. The price impact won't hit immediately, but it will hit before year-end.

Everyone knows you need ground vias around a shield can footprint. The question is: how many, and how close together?

The rule: via spacing ≤ λ/20 at your highest frequency of concern. At 3 GHz, that's 5 mm spacing. At 6 GHz, that's 2.5 mm.

But the mistake most designers make isn't spacing — it's via-to-pad registration. If your ground vias don't connect to the shield can's solder pads with a continuous copper pour on the pad layer, you've created a gap in the Faraday cage. The vias go to ground, but the shield can floats above them on a solder joint that may or may not make continuous contact.

Fix: Extend the ground pour on the shield pad layer to overlap every stitching via by at least 0.3 mm. Verify in your Gerber review — this is invisible in schematic and easy to miss in layout.

Last week we reviewed a customer's layout where the stitching vias were placed perfectly at 3 mm spacing — but the ground pour had a thermal relief pattern on the shield pad layer. Every via had a 0.2 mm gap in the copper connection to the shield. At 2.4 GHz, those gaps leaked enough energy to fail Wi-Fi immunity. Filled the thermal reliefs, problem solved. Zero BOM change.

If you spec nickel-silver or mu-metal shields, ask your procurement team to get an alternate quote in tin-plated CRS now — before the Indonesia vote creates a price floor you can't negotiate around.