The short answer
Allocation changes four things about an assembly project, and price is only the fourth. It changes which parts you can actually get in a given month, who owns the inventory while it waits, how long a quotation stays valid, and finally what the part costs. Teams that treat allocation as a pricing problem usually meet the other three the hard way, after the purchase order is already placed.
A working definition helps. A supplier is allocating when it accepts your order but limits the quantity, the delivery date, or your right to substitute a part. A project is affected by allocation when its build schedule is set by component arrivals rather than by the launch plan.
This article is about the cost sheet. It works through where allocation money actually goes on a printed circuit board assembly, which of those lines are negotiable, and which ones you will pay whether or not anyone writes them down. It is not a price forecast. Component markets in 2026 have moved faster than any forecast cycle, so every figure below is either a dated observation or a clearly labelled illustration.
What allocation actually means on a purchase order
Most buyers meet allocation as a clause rather than as an announcement. The language is usually mild: quantity confirmation subject to material availability, price valid for seven days, no cancellation once material is committed. Each of those phrases transfers a specific risk, and each one costs money somewhere in the project.
The mechanisms repeat across suppliers, and they are worth naming precisely because they are usually mixed together in one paragraph.
None of these is unreasonable on its own. A distributor that has to buy memory on a rising market cannot offer a sixty-day firm price without charging for the option. The problem is that buyers often accept all five mechanisms without pricing them, and then compare two quotations as though they were the same product.
- Hard allocation. The supplier can supply a fixed quantity per customer per month, and the excess becomes a backorder with no committed date.
- NCNR terms. Non-cancellable, non-returnable material. Once the order is placed, the inventory risk is yours even if your own schedule slips.
- Validity compression. The quote holds for a shorter window than your approval cycle, so it expires before your internal sign-off finishes.
- Substitution rights. The supplier may fit an alternative part, or may refuse to without a written approval that takes days to obtain.
- Priority tiers. Existing customers with long-run programmes get served first, and new programmes are quoted with longer lead times that are honest but unwelcome.
Two shortages, two different clocks
The 2026 component market is usually described as one shortage. It behaves like two, and they need different responses.
Memory is the fast clock. Capacity has been pulled toward high-bandwidth memory and high-density server modules, and the price of commodity DRAM and NAND has been reset by contract negotiations that happen monthly or quarterly. When that market moves, it moves by tens of percent in a single round. A buyer who confirms a price on Monday may find it is no longer offered on Friday, not because the supplier is unreliable but because the supplier's own input price changed.
Passives are the slow clock. A general-purpose 0402 ceramic capacitor is not scarce. What is scarce is the qualified part: the specific dielectric, voltage rating, case size and temperature characteristic that sits on an approved vendor list for an automotive or industrial programme. Qualification takes months, capacity for high-reliability grades is added slowly, and reported lead times for those grades have stretched well beyond a normal planning horizon.
The difference matters because the responses are opposite. Memory problems are solved with money and timing: buy earlier, accept a shorter validity, or redesign around a different density. Passive problems are solved with engineering: identify which parts are single sourced, which alternates are already qualified, and which ones can be re-qualified in weeks rather than quarters.
A useful rule for 2026: if the shortage is on the price list, you can usually buy your way out of it. If the shortage is on the approved vendor list, you cannot.
What actually moved in 2026
The table below is a summary of what suppliers and trade reporting described during 2026. It is deliberately coarse. Ranges vary by package, grade, density and volume, and the numbers are not a substitute for a quotation dated the day you read this.
The practical lesson is that a BOM has stopped being a single risk profile. Two lines on the same BOM can be in completely different markets, and treating the whole BOM as one lead time is how projects end up with everything except one part.
| Category | What was reported in 2026 | What it does to an assembly quote |
|---|---|---|
| DRAM and NAND | Contract prices reset sharply upward in early 2026, with single-round increases far outside normal ranges | Memory lines in a BOM can move more in one quarter than the whole assembly margin |
| High-reliability MLCC | Lead times extended into many months for qualified automotive and industrial grades | Forces early ordering or a documented alternate, and pushes first-article dates |
| Large-case and polymer capacitors | Price rounds in the tens of percent, with allocation on selected case sizes | Small share of BOM value, large share of schedule risk |
| Analogue and mixed-signal ICs | Mixed picture: commodity parts normal, specialised parts still extended | Requires per-line checking rather than a category assumption |
| Laminate and copper foil | Separate cost cycle, covered in our CCL timeline | Raises the bare-board line, not the component lines |
Why an assembly quote expires faster than a fabrication quote
A bare-board quotation is anchored to a material index. Laminate and copper foil prices are published, discussed and revised on a monthly rhythm, and most fabricators will hold a price for a defined window because they can see their own input cost for that window. The uncertainty is real but it is bounded.
An assembly quotation is anchored to a component market where individual lines can move without notice. That is why a 2026 assembly quote that claims sixty days of validity is either very conservative on price or quietly conditional. Neither is dishonest, but the buyer needs to know which one they are holding.
There is a second, less obvious reason. Assembly pricing includes a working-capital decision. If the supplier buys your components on day one and you pay on day sixty, the supplier is financing your inventory. In a rising market that financing has a real cost, and it appears in the unit price as a percentage that nobody itemises. Shortening the validity is one way suppliers manage that exposure without raising the headline number.
A worked example: 1,000 boards, three moving lines
Numbers make the mechanism visible. The BOM below is illustrative, not a quotation, and it is rounded for clarity. It represents a mid-complexity industrial control board with roughly 180 placements.
The scenario applies three changes: laminate up 20 percent, memory up 60 percent, passives up 25 percent. Everything else is held constant, including labour and margin.
The result is a 14.7 percent increase in unit price from three lines that together represent 36 percent of the original cost. Memory alone contributes 10.8 points of the 14.7. If memory doubles instead of rising 60 percent, the same board lands near 118.00 with no other change.
Two conclusions follow. First, the argument about labour cost is almost irrelevant in a year like this; the volatility is in materials. Second, a buyer who negotiates hard on assembly labour while accepting a seven-day validity on memory has optimised the wrong line.
| Line | Before (USD per board) | Share | After (USD per board) |
|---|---|---|---|
| Bare PCB, 6 layers | 12.00 | 12% | 14.40 |
| Memory | 18.00 | 18% | 28.80 |
| Passives | 6.00 | 6% | 7.50 |
| Other ICs | 22.00 | 22% | 22.00 |
| Connectors and mechanical | 8.00 | 8% | 8.00 |
| Assembly labour | 14.00 | 14% | 14.00 |
| Test | 5.00 | 5% | 5.00 |
| Packaging and freight | 3.00 | 3% | 3.00 |
| Overhead and margin | 12.00 | 12% | 12.00 |
| Total | 100.00 | 100% | 114.70 |
The four cost lines that never appear on the quote
The unit price is the visible part. Four other costs are real, are usually larger than buyers expect, and are almost never in the comparison spreadsheet.
Carry cost of pulled-forward inventory
When allocation forces you to order eight weeks earlier than the schedule requires, someone finances that inventory. At an annual cost of capital around eight percent, eight extra weeks on a material value of 46.00 per board is roughly 0.57 per board, or about 570 on a 1,000-piece build. That is modest. The number becomes serious when the pull-forward is a full quarter and the build is 20,000 pieces.
Expedite and broker premium
When a line is short, the part is still available, at a price. Broker premiums of 15 to 40 percent over contract price are common in tight markets, and the risk is not only cost. Parts from outside the authorised channel carry a traceability question that matters for any programme with an audit trail.
Requalification cost of a substitute
Approving an alternate capacitor is not a data-sheet comparison. It is engineering time, a deviation or change note, first-article inspection, and often a functional test run. For a safety-related or medical product it can be a formal change process measured in weeks. None of that appears on a bill of materials, and all of it consumes schedule.
Scrap and obsolescence on a cut allocation
If your allocation is cut after you have committed to a build plan, you may be left with kits that cannot be completed, or finished boards without the one part that makes them sellable. The cost is not the part; it is the partially built inventory and the missed delivery window.
Adding these four lines typically moves the real cost of an allocation event from the single digits into the high single or low double digits as a percentage of project value. The exact figure is project-specific, but the direction is consistent.
Turnkey, consignment or hybrid
The three sourcing structures move risk rather than remove it. The table compares them on the dimensions that actually differ.
Full consignment looks attractive in a shortage because it removes the supplier's margin on components. It also moves every delay, every quality question and every customs entry onto your team. For a 500-piece build with one short memory line, hybrid is usually the rational middle: consign the line that is scarce and price-sensitive, let the supplier handle the long tail of passives and hardware.
| Dimension | Turnkey | Full consignment | Hybrid (consigned critical lines) |
|---|---|---|---|
| Who buys components | Supplier | You | Split, defined line by line |
| Who owns inventory risk | Supplier, priced into unit cost | You | You, for the consigned lines only |
| Price validity | Short in a rising market | Not applicable to consigned lines | Longer on the assembly, short on consigned lines |
| Visibility of component cost | Low, bundled | Full | Full on the lines you consign |
| Administrative load | Low | High: purchase orders, shipping, customs, kitting | Medium |
| Best fit | Small and medium builds, simple BOMs | Programmes with existing supply agreements | Builds where one or two lines dominate cost or risk |
Seven questions to ask before you accept an allocation clause
The answers are more useful than the price. A supplier who can answer all seven has a real material plan. A supplier who answers only the first has a price list.
- Which lines in this quotation are firm, and which are subject to reconfirmation?
- What is the exact validity date, and what happens to price if the order is placed one day later?
- If a line is short, who sources the substitute, and does that require my written approval?
- Is the substitution list agreed in advance, or negotiated at the moment of shortage?
- Which parts are single sourced, and what is the qualified alternate for each?
- If my schedule slips, what is the cancellation window before material becomes NCNR?
- Can you show the lead time per line rather than one lead time for the build?
A triage method for the BOM
Most BOMs have two hundred lines and four that matter. Finding those four does not require software, only five columns and an honest hour.
Sort every line by four attributes: annual spend, number of approved sources, qualification time for an alternate, and how far the current lead time exceeds the build plan. Then rank by the product of the first three and the gap in the fourth. The lines that rise to the top are your risk register, and they are usually not the most expensive lines. A three-dollar connector on a six-month qualification cycle can stop a build as effectively as a thirty-dollar memory device.
Two patterns show up almost every time. First, the highest-spend line is often well managed, because everyone watches it. Second, the lines that actually stop builds are cheap, single sourced and boring: a specific terminal block, a particular relay, a connector housing with a tooling lead time. Those are the lines worth consigning or stocking, and they are rarely the lines a buyer negotiates.
If you only do one thing after reading this, list the ten cheapest single-sourced lines on your BOM. That list, not the memory line, is usually where the schedule risk lives.
Lead time is not the number to plan against
A build is not finished when the average part arrives. It is finished when the last part arrives, and that is a different number. Buyers routinely plan against a weighted average lead time, which is arithmetically reassuring and operationally meaningless.
Consider a 300-line BOM where 295 lines arrive in four weeks and five lines take sixteen. The average is about 4.2 weeks. The build starts in sixteen. The gap between those two numbers is where expediting budgets disappear, and it exists because the plan was built from an average rather than from a maximum.
Planning against the maximum has a cost of its own: you hold inventory longer and tie up cash. The usual compromise is to split the BOM into a kitting decision and a build decision. The long lines determine when you can start; the short lines determine when you can finish. Once those two dates are separate, the schedule conversation becomes specific, and specific conversations are the ones that produce a plan instead of an apology.
There is a second-order effect worth noting. When five lines control the start date, the negotiation leverage changes. Those five are worth paying a premium for, and the other 295 are worth squeezing. Buyers who apply one negotiation strategy across the whole BOM usually overpay on the lines that do not matter and lose time on the ones that do.
What a good quotation package looks like in an allocation year
A quotation in 2026 should be readable as a risk document, not only as a price. Four things make the difference, and all four are cheap for a supplier to provide.
A dated validity statement, with the date of the price check rather than the date of the PDF. A per-line lead time, or at least a per-category one, so the buyer can see which lines set the start date. A written note on which lines are firm and which are subject to reconfirmation. And a substitution position: either an agreed alternate list, or an explicit statement that no substitution is permitted without written approval.
None of this makes the price lower. It makes the price comparable. Two quotations that differ by four percent on unit price but differ by six weeks on start date and by a validity window of seven versus forty-five days are not really four percent apart, and the spreadsheet will never show you that unless the supplier writes it down.
What we can and cannot do
Being clear about this saves both sides a meeting.
We can review your BOM line by line, flag the parts where a single source or a long qualification cycle creates schedule risk, and quote against approved alternates that you document in writing. We can tell you which lines we are confident about for a given delivery window and which ones we would want to confirm again before you commit. Where a part is short, we will say so at quotation rather than after the order.
We cannot guarantee allocation. No manufacturer can, because allocation is set upstream by wafer capacity, by qualified-material availability and by distributor policy. We also will not fit an unapproved substitute to protect a ship date. A board that arrives on time with an unqualified part has not solved the customer's problem; it has moved it to the customer's customer.
The working arrangement that holds up is boring and effective: agree the alternates before the build, put them in writing, and let the quotation state which lines are firm.
The next 30 days
- Pull the BOM and mark every line with a single approved source. Those are the risk lines, regardless of price.
- For each risk line, write down one qualified alternate and one that would need re-qualification. If you cannot name either, that is the first conversation.
- Ask your supplier for lead time per line, not per build, and compare it with your assembly date rather than your ship date.
- Decide which two or three lines are worth consigning. In most 2026 builds it is memory plus one specialised IC, not the whole BOM.
- Fix the validity expectation in writing before the quotation, so nobody is surprised when it expires.
- Move the review earlier. A BOM review three weeks before build start is a formality; six weeks before is a decision.
How to keep this from being a surprise next time
The pattern that repeats across projects is not a shortage. It is discovering the shortage at the wrong moment. Teams that handle 2026 well tend to do three things differently, and none of them is clever.
They keep a live alternates list attached to the BOM rather than to a folder. They treat quote validity as a project milestone, with a named owner, instead of a line in a PDF. And they separate the decision about what to build from the decision about when to buy, so that a component delay changes the purchasing plan rather than the product.
None of that requires a forecast to be right. It only requires knowing which lines can hurt you before the market tells you. The CCL side of the same problem, where laminate rather than components drives the number, is covered in the 2026 CCL price timeline, and the way those material increases reach a bare-board quotation is explained in why PCB quotes keep expiring. The rest of the cost picture, including duty and landed cost, is indexed in the sourcing and cost section.
Questions, answered
Before your next build
Is allocation the same as a long lead time?
No. A long lead time is a date. Allocation is a limit on quantity, or a condition attached to the order. A part can have a normal lead time and still be allocated, which is why asking only about lead time misses the risk.
Should we switch to full consignment while parts are short?
Only if you already have the supply agreements and the logistics to run it. Consignment moves the shortage to your desk. For most small and medium builds, consigning two or three critical lines gives most of the benefit with a fraction of the administration.
How long should a PCBA quote be valid in this market?
Shorter than most buyers expect. Ask for the validity date in writing and treat it as a real deadline. If your approval cycle is longer than the validity, agree a reconfirmation step in advance rather than discovering the gap at order placement.
Does a higher memory price always mean a higher board price?
It means a higher material cost. Whether it reaches the unit price depends on margin, on whether the part is consigned, and on when the material was bought. A supplier holding older inventory may quote below the current market for a short period, which is an opportunity rather than a guarantee.
What is the single most useful thing to send with an RFQ?
The BOM with approved alternates marked, plus the date you need boards, not the date you need parts. That one document turns a price request into a material plan.