THE COMPONENT SIGNAL

Issue #005 · Monday, February 10, 2026 · Tariffs + Metals

The 25% tariff on Korean electronics imports — announced January 31, effective March 1 — isn't hypothetical anymore. It's in procurement spreadsheets. If your BOM includes Korean-origin passives, connectors, or shielding components, your landed cost is about to shift.

What's actually affected: finished electronic components and assemblies with Korean country of origin. Raw materials (nickel, tin, copper) routed through Korea for processing may also trigger the tariff depending on HTS classification and substantial transformation rules.

What we're seeing: Three of our US customers asked for Vietnam-origin certificates within 48 hours of the announcement. The parts are identical. The routing changes the duty rate from 25% to 0%. The teams that moved fast are already re-quoted. The ones still waiting for "official guidance" will pay the tariff on their next PO.

What to do: Ask your Korean-origin suppliers where the product is actually manufactured vs. where it ships from. If there's a Vietnamese or Southeast Asian production option, request it now — before March 1 creates a queue.

Metals

What's driving copper: The AI infrastructure buildout is consuming copper at a rate nobody modeled correctly. Each hyperscale datacenter uses 30,000–60,000 tonnes of copper for power distribution, cabling, and busbars. Goldman Sachs revised their 2026 copper demand forecast upward by 4% in January, citing "AI-driven electrification." If your shielding alloy contains copper — and most tin-plated CRS does — watch your material cost through Q2.

Every shield can has a resonant frequency determined by its internal dimensions. For a rectangular can, the dominant TE₁₀ mode resonance is approximately:

f = c / (2 × L) where L is the longest internal dimension.

A 30 mm shield can resonates around 5 GHz. A 50 mm can resonates around 3 GHz. If your switching regulator's harmonics land on that resonance, the shield amplifies instead of attenuates.

Rule of thumb: Keep your shield can's longest internal dimension below λ/4 at your highest frequency of concern. If you can't shrink the can, add an internal absorber strip or partition wall to break the resonant mode.

We had a customer last month with a 45 mm shield can over a 2 MHz switching regulator. Clean at fundamentals. But the 5th harmonic at 10 MHz was bouncing around inside the can — the 45 mm dimension put the TE₁₀ resonance right at the harmonic. Dropped the performance by 8 dB compared to an open board. Swapped to two smaller cans with a partition wall. Problem gone. Same BOM cost.

If you're sourcing Korean-origin components, request Vietnam or Southeast Asian routing options from your suppliers this week — before March 1 creates a queue. The teams that wait will pay 25% more on the same parts.