The four clocks in a magnet lead time
A quoted lead time is the sum of four processes that do not overlap as much as buyers expect. Compressing the wrong one buys nothing.
The question to ask your supplier
Not “can you go faster?” but “which clock is binding on this part, and what would move it?” A supplier who can answer that in one sentence is quoting from a real schedule. A supplier who only offers a percentage expedite fee is quoting from a spreadsheet.
Lead time by sourcing path
The same part number can carry six different lead times depending on how it is sourced. Establish which path you are on before negotiating a date.
| Sourcing path | Typical lead time | Binding clock | When it applies |
|---|---|---|---|
| Stock, shipped as-is | 1–3 days | Transit only | Catalogue size, grade and coating already on the shelf in the U.S. |
| Stock + secondary operation | 1–3 weeks | Production | Stock blank ground to a modified dimension, re-plated, or re-magnetized to a different pattern |
| Custom, existing tooling | 4–8 weeks | Production | Geometry already tooled at the mill; grade or coating change only |
| Custom, new tooling | 8–12 weeks | Tooling | New pressed geometry or a new magnetizing fixture |
| Custom, HRE-bearing grade | 10–16 weeks | Material + border | SH / UH / EH / AH grades, or SmCo — adds alloy allocation and licence review |
| Qualified program (PPAP / FAI) | 16–24 weeks | Qualification | First article, measurement studies, capability run and document package before production release |
Where the schedule usually slips — and it is rarely the factory
Across programs, the largest single consumer of calendar time between enquiry and delivery is drawing and specification churn on the buyer’s side: an RFQ issued without tolerances, without a magnetization direction, or without a stated operating temperature comes back with questions, and each round trip costs days before the clock even starts. A complete package on day one is worth more than any expedite fee. The checklist is in the RFQ guide.
Minimum order quantities and what drives them
MOQ is not an arbitrary commercial floor. It is the point at which a batch process stops losing money, and every step in magnet manufacturing has its own batch economics.
| Driver | Why it sets a floor | Typical effect |
|---|---|---|
| Press die amortization | Tooling cost is fixed regardless of quantity; below a certain volume the per-piece tooling charge exceeds the part cost | Sets the commercial MOQ on new geometries |
| Sintering lot size | Vacuum furnace cycles are ~24 hours and are charged by cycle, not by piece | Small orders share a lot or wait for one |
| Plating tank minimum | Barrel or rack plating has a minimum viable load; setup and bath chemistry cost the same for 200 parts as for 20,000 | Often the binding MOQ on small parts |
| Magnetizing fixture | Custom multipole or true radial fixtures are built per part family | Adds a one-time cost and a volume floor |
| Alloy melt lot | Non-standard compositions are melted in minimum quantities | Significant only for unusual grades |
| Qualification overhead | First article, capability data and documentation cost the same at any volume | Dominant on PPAP and FAI programs |
Practical MOQ ranges
- Catalogue stock: no meaningful MOQ. Single pieces ship for prototyping.
- Custom on existing tooling: typically a few hundred to a thousand pieces, driven mostly by plating.
- Custom with new tooling: commonly 500–5,000 pieces depending on part size and whether tooling is amortized or paid up front.
- Large or complex geometries: lower piece counts, because part value carries the batch.
Getting under the MOQ legitimately
- Pay the tooling separately. Buying the die outright decouples the MOQ from amortization and leaves you free to order in smaller releases.
- Piggyback the plating lot. Accepting a standard coating that runs continuously avoids the tank minimum entirely.
- Design to a stock blank. A geometry that can be ground from an existing pressed size skips the press die and its MOQ.
- Aggregate part numbers. Three similar sizes across three programs often consolidate into one tooled size — see the value engineering guide.
- Order the annual quantity, schedule the releases. A blanket order meets the MOQ on paper while delivering monthly — covered in inventory programs.
The export licence clock
Since 2025 this has become the least understood and most schedule-destructive item in magnet procurement. It is not a tariff question and it is not solved by paying more.
China’s Ministry of Commerce placed a group of medium and heavy rare earth elements — including dysprosium, terbium and samarium — under dual-use export licensing in April 2025. A further expansion announced in October 2025 was suspended in November 2025 for one year, but the underlying April 2025 licensing regime was not suspended and remains in force. Headlines about paused controls routinely get read as “magnets ship freely again.” They do not.
The composition trigger, not the part number
Licensing is driven by what is in the magnet, not by what it is called. In practice the threshold that matters is dysprosium or terbium content at or above roughly 0.1% by weight, individually or combined. That captures essentially every high-coercivity grade — the SH, UH, EH and AH suffixes — and samarium cobalt is captured separately through the samarium listing. Standard N-series NdFeB with no meaningful heavy rare earth addition generally is not.
If your bill of materials calls out N42SH, N38UH, N35EH or any SmCo grade, plan the schedule as a licensed shipment until a composition certificate proves otherwise.
What it does to the schedule
- Statutory review target is around 45 working days under China’s dual-use export control regulations — roughly nine calendar weeks, and that is the target rather than a guarantee.
- The clock starts at application, not at order. A supplier who waits for your PO before beginning the licence file has already lost weeks.
- It runs before production ships, not after. Finished parts sitting in a warehouse awaiting a licence are still not your parts.
- End-use documentation is part of the file. Applications typically require identified end users and end uses; incomplete or vague statements cause resubmission.
- It is revocable and political. In June 2026 China added U.S. rare earth firms including MP Materials and USA Rare Earth to its export control list, a reminder that access can narrow without notice.
The observable result: Bloomberg reported in July 2026 that China’s magnet exports to the United States in the first half of 2026 ran roughly 20% below the 2022–2024 average, despite the trade truce. Volume has not returned to trend, and grade-specific tightness is worse than the aggregate suggests.
How to plan around it
- Design out heavy rare earth content where the thermal duty allows. An N42H part that was specified as N42SH out of caution carries licensing exposure it does not need. Run the load-line check at the real maximum temperature — the method is in the temperature guide.
- Hold safety stock specifically on HRE grades. The buffer that matters is not uniform across your magnet BOM; concentrate it where licensing applies.
- Buy from domestic inventory that has already cleared. Parts already in the United States carry no residual licence risk.
- Ask for the composition certificate. It is the document that determines whether a shipment is controlled — see certificates and compliance documents.
Transit, mode, and the magnetized-cargo problem
Once parts exist and are cleared to leave, mode selection is the last lever — and magnets are not ordinary air cargo.
| Mode | Typical transit | Relative cost | Constraints |
|---|---|---|---|
| Ocean FCL | 4–6 weeks | 1× | Port congestion and blank sailings add variance; the default for production volume |
| Ocean LCL | 5–8 weeks | 1.5–2× | Consolidation and deconsolidation add a week at each end |
| Air freight | 5–12 days | 8–15× | Magnetic field limits apply — see below |
| Air express | 3–6 days | 15–25× | Same field limits; practical only for samples and small quantities |
Magnetized product is regulated cargo by air. A package whose field exceeds the IATA magnetized-material threshold at the prescribed distance must be shielded, declared, and in some cases compass-deviation tested before a carrier will accept it. Packaging and shielding work adds real days to an air expedite and real cost per carton, and a rejected tender at the airport costs more than the days it was meant to save.
The underused option: ship unmagnetized
Sintered magnets can be manufactured, machined, plated and shipped in an unmagnetized state, then magnetized in a fixture at final assembly. Unmagnetized blanks are ordinary cargo — no field limits, no shielding, simpler packaging, lower handling risk, and denser packing. Where your assembly process can accommodate a magnetizing station, this removes an entire category of air-freight friction. It has to be agreed at RFQ stage, because the magnetizing fixture becomes your responsibility rather than the supplier’s.
What actually compresses a schedule
Ranked by days recovered per dollar spent. The first four are usually available; the last two are expensive and situational.
| Lever | Typical time recovered | Cost / trade-off |
|---|---|---|
| Substitute a stocked grade or size | 4–10 weeks | Requires a design review; often free. The highest-yield lever by a wide margin. |
| Relax a non-functional tolerance | 1–3 weeks | Free. Removes a grinding pass and its queue — see tolerances and acceptance. |
| Split shipment: air the bridge, ocean the balance | 3–5 weeks on first parts | Air premium on 5–15% of the quantity only. Usually the cheapest way to hit a build date. |
| Ship unmagnetized | 3–10 days | Needs a magnetizing station at assembly; removes air-cargo field constraints. |
| Accept a partial first article | 2–4 weeks | Dimensional and magnetic data released ahead of the full document package, with the balance following. |
| Pay for tooling priority | 1–3 weeks | Real but bounded — a die takes as long as it takes; only queue position moves. |
| Expedite fee on production | 0–1 week | The least effective lever. Sintering and plating cycle times are physical, not commercial. |
What does not compress a schedule
- Paying more for a licensed shipment. Regulatory review is not a queue you can buy into.
- Splitting the order across two suppliers who use the same mill. Common, and it duplicates cost without duplicating capacity. Verify the mill, not the vendor — the second-source logic is in inventory programs.
- Escalating on a part waiting for tooling. Nothing downstream can start.
- Upgrading the grade to “whatever is available.” A higher grade is not automatically a drop-in; coercivity, not energy product, usually governs. See the grades chart.
Building lead time into the purchase order
The durable fix is not faster expediting; it is a reorder policy that assumes the lead time you actually have.
Reorder point
The quantity on hand at which a replenishment order must be released:
worked example
A part consuming 2,000 pieces per week with a standard deviation of 400 pieces per week, on a ten-week lead time, targeting 98% service:
- Demand over lead time: 2,000 × 10 = 20,000 pieces
- Standard deviation over lead time: 400 × √10 ≈ 1,265 pieces
- Safety stock: 2.05 × 1,265 ≈ 2,600 pieces
- Reorder point: 20,000 + 2,600 = 22,600 pieces — roughly eleven weeks of cover
Lead time variability costs more than lead time length
A reliable sixteen-week lead time is easier and cheaper to plan around than a ten-week lead time that sometimes takes eighteen. If supply is erratic, the safety stock term must be sized against variability in both demand and lead time — which is precisely why the 2025–2026 licensing environment drove buffer requirements up even where nominal quoted lead times held steady.
Quick planning reference
| Part profile | Suggested cover | Rationale |
|---|---|---|
| Catalogue stock size, N-series grade | 2–4 weeks | Replenishable from domestic inventory |
| Custom geometry, N-series grade | 8–12 weeks | Full production cycle plus transit |
| Custom geometry, H / SH grade | 12–16 weeks | Adds alloy allocation |
| Any UH / EH / AH or SmCo part | 16–26 weeks | Adds export licence review and material scarcity |
| Single-sourced, line-stopping part | 26+ weeks | Buffer sized against a supply interruption, not a delivery slip |
If holding that much inventory on your own balance sheet is unattractive — and for most manufacturers it is — the structures that move the carrying cost off your books while keeping the cover in place are covered in blanket orders, consignment and VMI.