Create a free Supply & Demand Chain Executive account to continue reading

Why Just-in-Time Fails for Critical Semiconductor Components

The inventory target is to carry enough usable material to finish the right builds and redirect what those builds no longer consume.

Digital warehouse with holographic circuit board overlay and data visualization in blue and orange neon lighting
YM Creative Studio AdobeStock_754862606

Just-in-time inventory strategies fail for critical components because a single missing part can halt production of expensive systems. Manufacturers need deliberate buffer strategies, multi-source qualification, and supply agreements to protect production schedules while managing capital efficiently.

  • A single inexpensive component like an MLCC capacitor can stop production of systems worth thousands of dollars, making unit cost irrelevant to true risk.
  • AI infrastructure demand is shifting supplier capacity away from traditional manufacturers, making historical forecasts and lead times unreliable.
  • Count complete builds using "square sets"—the number of full production runs supported by available inventory—rather than aggregate inventory valuation alone.
  • Critical components require multi-source qualification, direct supplier relationships, and supply agreements that protect against allocation shifts and end-of-life decisions.
  • Inspection, authentication, and exit planning for surplus inventory must be built into procurement strategy before purchasing, not after inventory arrives.

A manufacturer can have shelves full of expensive memory and processors but still have nothing to ship. For example, a multi-layer ceramic capacitor (MLCC) may cost less than $1 and appear on the bill of materials (BOM). Without it, the build stops. The capital tied up in everything else also stops working. That penny part deserves more attention than its purchase price suggests.

Semiconductor inventory decisions have often revolved around how little material a company can carry while keeping production supplied. Artificial intelligence (AI) infrastructure demand complicates that calculation.

Capacity is moving toward customers and applications that command suppliers’ attention, while manufacturers outside those markets find their commitments slipping. Even an automotive business with flat demand can lose access to components also used in AI data centers. Its own forecast may be accurate, but the supply assumptions behind it may no longer hold.

The unit cost tells only part of the story

So, what happens if this component disappears for 12 weeks? If production stops or a customer shipment slips, the answer gives procurement and finance something more meaningful to discuss than cost alone. A $5 semiconductor holding up a $5,000 system has different economics from an equally priced part with several qualified substitutes.

The BOM review then gets specific. How many qualified sources exist? How long would engineering take to approve another? Does the quoted lead time reflect actual availability, including allocation risk?

A single-source component with a six-month qualification cycle remains exposed even when the supplier quotes delivery in 12 weeks. Commitments can shift from 12 weeks to 52 weeks, forcing buyers to revise production schedules around components that have yet to arrive.

Memory attracts attention because it consumes so much capital. Older microcontrollers, NOR Flash, analog integrated circuits (ICs), and specialty passives are easier to overlook. Those modestly priced parts are difficult to replace due to capacity migration and end-of-life decisions.

Engineering has a practical role here: expand the approved vendor list and qualify alternatives before the original component becomes scarce. For passives, that work can keep a cheap shortage from immobilizing a costly build.

Count complete builds

No warehouse has a perfect mix for long. Boards change, customers cancel, and suppliers deliver only part of what they promised. A more useful measure is how many complete builds the available material supports, a concept known as “square sets,” in which all required components are available to complete production. Counting them exposes the missing sockets that an aggregate inventory valuation conceals.

It also changes the conversation about surplus. Material without a production home today may help fund the components missing from another, higher-margin project.

Before selling anything, however, there is a simpler check. Does another division already want it? A company may purchase a component in one region while another division sells excess of the identical part. Both transactions appear reasonable locally. Together, they reveal how much money disappears between disconnected inventory records.

An enterprise-wide overhaul is not the only place to start. One shared list of the 50 most critical manufacturer part numbers, covering plants, contract manufacturers, and material in transit, can reveal worthwhile transfers. Assigning an owner and reviewing that list weekly can influence purchasing.

Visibility also involves relationships. When an original equipment manufacturer (OEM) leaves component procurement solely to a contract manufacturer, it competes for attention alongside that manufacturer’s other customers. Direct supplier relationships give the OEM another way to understand availability and join allocation discussions.

On the shelf or at the hub

Critical material belongs on the customer’s shelf. A promised delivery offers little comfort when the line runs out. For a company serving several plants, maintaining separate inventory buffers at each location can limit flexibility. Inspected stock in a regional hub can move among facilities as their requirements change. Expensive, volatile components may benefit from that arrangement, while inexpensive parts that stop production may belong closer to the line.

Geography can undo either plan. Customs delays, tariffs, and export restrictions affect whether inventory reaches the operation it was purchased to protect. Stock across a border is not necessarily accessible stock.

Nor do all semiconductor categories reward the same buying decision. Memory prices move sharply as supply changes. Processors and graphics processing units (GPUs) carry generation risk: surplus from the wrong generation can become obsolete. Storage substitutions involve controllers, firmware, and density, so an apparently similar product may be unsuitable for an already qualified application.

That is the tension in building a buffer. Protection has value, but buying broadly without regard to the part can create tomorrow’s excess.

Available is not yet usable

Receiving and authentication belong inside inventory planning. A shipment awaiting verification cannot safely be treated as ready for production. Checks include manufacturer and part number, date and lot codes, packaging, moisture sensitivity, and traceability. Inspection can reveal remarking, resurfacing, or mishandling; higher-risk material may require electrical or functional testing. The source and application help determine the level of scrutiny.

Supplier audits also extend upstream. A distributor’s testing capability matters, but so does how it evaluates its suppliers. Rejecting questionable material remains part of securing supply, even during a shortage. Inspection takes time. A replenishment plan that ends at the receiving dock leaves that time unaccounted for.

Build the exit before buying

Finance, operations, and procurement can compare buffer costs with lost margin, downtime, expedited freight, spot-market premiums, and customer penalties. Agreeing on an exit route makes the decision less speculative.

Returns, internal transfers, and redeployment into other products offer possible outlets. A managed program can sell truly surplus material into active demand rather than liquidating it as an undifferentiated lot.

The exit plan also changes when reviewing the surplus. Waiting until an annual write-down can miss buyers with current requirements. Matching authenticated parts to those requirements, testing prices, and selling selectively creates additional opportunities to sell excess inventory and free cash for purchases that support scheduled production and planned customer shipments.

Just-in-time still fits reliably replenished components with qualified alternatives. Elsewhere, deliberate buffers and supply agreements offer protection. The inventory target is to carry enough usable material to finish the right builds and redirect what those builds no longer consume.

Page 1 of 98
Next Page