Segmenting the spend
A magnet category splits cleanly on two axes — supply risk and business impact — and each quadrant gets a different treatment:
| segment | typical parts | strategy |
|---|---|---|
| Strategic (high impact, high risk) | Custom rotors, high-temp classes with heavy rare earths, sensor magnets with tight field specs | Deep qualification, contracts with change-notification and continuity clauses, dual-source or stocking insurance per the second-sourcing worksheet |
| Bottleneck (low impact, high risk) | Legacy customs, low-volume odd geometries, single-fixture magnetization patterns | Convert to stock sizes where possible; otherwise buffer with inventory — qualification spend here is usually wasted |
| Leverage (high impact, low risk) | High-volume standard geometries in common grades | Competitive tension on landed cost, annual should-cost, volume consolidation |
| Routine (low impact, low risk) | Catalog discs, blocks, rings, pot magnets, fixture and jig magnets | Catalog/stock supply, minimal transactions — U.S. stock, blanket releases, done |
The classic category error is treating everything as Leverage — running three-bid events on Strategic parts (where switching is slow and risky) and on Routine parts (where the event costs more than it saves).
Should-cost & the rare earth floor
A sintered NdFeB part has a calculable material floor: magnet weight × alloy cost, where the alloy price is driven by Nd/Pr oxide — plus a multiple for Dy/Tb content in high-temperature classes. Layer on processing (pressing, sintering, grinding by surface count, coating by system), tooling amortization, margin, and landed-cost elements, and you have a model good enough to sort quotes into three bins:
- Near the model: negotiate the normal things — volume, terms, logistics.
- Well above the model: ask what you're specifying that's expensive. The answer is usually a tolerance, a temperature class, or a custom geometry — the specification levers ranked in the cost reduction guide outsave negotiation by an order of magnitude.
- Well below the floor: walk away. Substituted grade, misrepresented material, or margin recovery planned for lot two — the red-flag math from the qualification guide.
Index exposure & contract structures
Because the raw material is a traded commodity, every contract structure is a decision about who holds the index risk:
Whichever structure: put the pass-through symmetry on the scorecard. Fast-up, slow-down pricing against the index is the most common quiet margin leak in the category.
Tariffs in the total cost
Duty exposure on Chinese-origin magnets is now a first-order cost element, not a rounding error — Section 301 action on permanent magnets stepped up in January 2026, stacking with other measures, while China's export-licensing regime on rare earth materials and magnets adds lead-time risk on top. The category-level implications:
- Compare quotes on landed cost with the duty math shown. HTS classification, origin, and the current tariff stack belong in the comparison sheet — the worked arithmetic is in the tariffs & landed cost guide.
- Assign the mid-contract change. Duty modifications arrive faster than contract cycles; the agreement should state who absorbs a change and how it's evidenced.
- Mind the licensing clock, not just the rate. Export licenses add calendar risk independent of price — a part that clears at an acceptable duty but waits on a license still stops your line.
- Re-run the footprint math when policy moves. Each tariff or control round shifts the break-even between China-direct, third-country, U.S.-stocked, and ex-China supply — which is the next section.
The sourcing-footprint decision
| footprint | strengths | watch |
|---|---|---|
| China direct | Deepest capability and capacity, lowest ex-works cost at volume | Full tariff and export-control exposure; you own import, licensing, and logistics risk |
| Third-country / SKD routes | Can moderate duty exposure for assemblies | Origin rules are substance-over-form; classification aggressiveness is a compliance risk, not a savings |
| Ex-China manufacturing | Policy-risk reduction; increasingly required by downstream customers | Premium pricing, limited grade/geometry range, capacity still scaling — qualify early, expect longer ramps |
| U.S. stockist / value-add supplier | Landed pricing, inventory buffer inside the border, engineering interface, absorbs import mechanics | Qualify their upstream chain one level up — named sites, change notification — per the qualification guide |
For most mid-volume categories the resilient answer is a portfolio: direct or ex-China relationships for Strategic parts, U.S.-stocked supply for Leverage and Routine, and aggressive custom-to-stock conversion to shrink the Bottleneck quadrant.
Where dual sourcing pays
Dual sourcing magnets is more expensive than it looks — duplicate tooling, duplicate qualification, and the physics problem that the drawing doesn't fully define the part (grade windows, house-built magnetizing fixtures, alloy recipe differences). The second-sourcing guide scores the decision; the category summary:
- Dual-qualify: line-stopping parts with long requalification times and stable multi-year volume — the tooling duplication amortizes against the outage cost.
- Buffer instead: parts where a stocking agreement or consignment inventory covers the realistic disruption window at a fraction of dual-qualification cost.
- Design out: Bottleneck parts whose risk is really a specification problem — a stock-size redesign or a grade-window-tolerant design removes the exposure permanently.
- Cross-qualify properly or not at all: a “backup” supplier who has never run production tooling is a phone number, not a second source.
Levers & the annual review
Ranked by realized savings across magnet categories, the levers run specification-first — the full ten, with effort ratings, are in the cost reduction guide:
- Specification: temperature-class re-justification, tolerance relaxation off non-functional surfaces, custom-to-stock conversion, coating right-sizing.
- Structural: volume consolidation across plants and products, blanket structures with index clauses, landed-cost optimization of the footprint.
- Commercial: should-cost-anchored negotiation, pass-through symmetry enforcement, tooling ownership and amortization terms.
Run the annual review in that order — engineering-facing levers first while the commercial calendar is quiet — and end each cycle by re-scoring the segmentation grid: parts migrate between quadrants as volumes, policy, and designs change.