Minimum Order Quantity
A minimum order quantity, or MOQ, is the smallest wafer volume a foundry will accept for a given process and order type, set at the point where the fixed costs of a mask set and process qualification stop dominating the per-wafer economics.
The floor exists because mask and qualification costs do not scale down. A full mask set for a leading-edge node can cost several million dollars and a qualification run consumes real tool time regardless of whether the final order is for 25 wafers or 25,000, so a foundry quoting below its MOQ is effectively absorbing a fixed cost that a larger order would have spread thin enough to ignore.
Multi-project wafer runs exist specifically to get small customers under the MOQ line. A shuttle run splits a single mask set and a single qualified process lot across many unrelated small designs, so each customer pays only a fraction of the fixed cost instead of the full mask bill, which is why MPW programs such as those run through Europractice and MOSIS have long been the standard path for university research and early-stage startups that could never meet a standard foundry MOQ on their own.
| Customer size | Typical path below MOQ | Trade-off |
|---|---|---|
| Startup, first tapeout | multi-project wafer (MPW) shuttle run | shared mask cost, fixed shuttle schedule |
| Small fabless, established | volume surcharge on standard run | pays a premium per wafer |
| Mid-size, growing volume | negotiated reduced MOQ with price floor | foundry bets on future growth |
| Large, take-or-pay customer | MOQ is irrelevant, floor is contractual | sits well above any standard MOQ |
MOQ and take-or-pay sit at opposite ends of the same spectrum. A take-or-pay capacity agreement guarantees a volume floor far above any standard MOQ in exchange for locked-in capacity, while a customer below MOQ is on the opposite end, paying extra or sharing a mask precisely because its volume is too small to matter to the foundry on its own terms.
A rising MOQ is one of the quietest signs of a tightening market. During the 2020-2022 shortage, several foundries raised effective minimum order volumes on mature nodes as capacity tightened, which pushed smaller customers either toward the MPW shuttle model or out of certain fabs entirely, reinforcing the same allocation-queue dynamics that determine who gets served first when supply is constrained.
Read MOQ through a *fixed-cost-per-slot* lens rather than a *minimum sale* lens: the foundry is not refusing small orders out of preference, it is refusing to let one customer's mask and qualification costs eat a production slot that a larger order could fill more profitably, which is exactly the economics a multi-project wafer run is built to route around for everyone too small to clear that line alone.