Last-Time-Buy vs Second-Source: Managing Discontinued Fab Parts

A last-time buy (LTB) is a single final purchase placed before an OEM stops producing a part, sized to cover every unit a fab will need for the rest of that tool’s service life. A second source is a qualified equivalent part from an alternative manufacturer. The right choice depends on remaining demand, how confidently it can be forecast, and whether a form-fit-function equivalent already exists. For most 200mm and legacy LAM, TEL and Applied Materials toolsets, the two routes work best in combination, not in isolation.

When an OEM issues an end-of-life or product discontinuance notice, procurement faces a clock and a question. The clock is the last-order-date window. The question is whether to bridge the gap by stockpiling original parts or by qualifying a second-source alternative. This article gives the forecast horizon, the order-quantity formula, the carrying-cost maths and the decision logic to answer it for discontinued fab parts, then shows where each route fits an obsolescence-management programme aligned to IEC 62402.

What is a last-time buy?

A last-time buy is the final order a customer places against a part the original component manufacturer has scheduled for discontinuation. The OEM publishes a last-order-date, after which no further orders are accepted, and a last-ship-date by which outstanding orders are fulfilled. The buyer’s job is to forecast total remaining demand across the supported life of the affected tools and place one order, the bridge buy, that covers it.

The LTB exists because semiconductor capital equipment outlives its component supply chain. A 200mm etch or deposition platform can run productively for 20 to 30 years, while the controllers, RF generators, mass-flow controllers, valves and circuit boards inside it follow much shorter commercial lifecycles. When a sub-component reaches end-of-life, the tool does not. The LTB is the mechanism that keeps an otherwise healthy tool supplied until the fab is ready to retire, upgrade or second-source it.

An LTB is distinct from routine spares replenishment in three ways. It is a one-time, non-repeatable order. It carries the full obsolescence exposure of the part for the rest of the tool’s life. And it forces a forecast over a horizon far longer than a normal reorder cycle, often a decade or more, where forecast error is unavoidable.

Last-time buy vs second source: the core trade-off

Both routes solve the same problem: keeping a discontinued part available. They differ in where the risk sits and when the cost lands.

Dimension	Last-time buy (LTB)	Second source
What you buy	Original OEM part, in bulk, now	Qualified form-fit-function equivalent, as needed
Capital timing	Large up-front outlay	Spread over time, pay-as-you-consume
Carrying cost	High, inventory held for years	Low, minimal stock
Forecast risk	Over-buy (write-off) or under-buy (stranded tool)	Forecast risk largely removed
Qualification effort	None, identical part	Up-front qualification required
Shelf-life risk	Real, parts degrade in storage	Minimal, built to order
Best when	Demand is modest, predictable and short-horizon; no equivalent exists yet	Demand is high, long-horizon or uncertain; an equivalent is available

The trade-off is not binary. A common pattern is a modest LTB to cover the qualification window, run in parallel with a second-source programme that takes over once the alternative part is fab-qualified. The LTB buys time; the second source removes the long tail of risk. The sections below quantify each side so the split can be set deliberately rather than by guess.

How do you calculate last-time-buy quantity?

The order quantity is the single most consequential number in the decision. Buy too little and the tool is stranded when stock runs out before a second source is ready. Buy too much and capital is sunk into inventory that is written off at end of life. A defensible LTB quantity is built from a demand forecast, a safety buffer for forecast error, and an adjustment for incoming yield, then netted against stock already on hand.

A repeatable formula, adapted from standard obsolescence-management practice, works as follows:

• Projected demand = production/PM demand + field-failure replacements + repair-loop attrition + qualification spares + training/engineering spares, summed across the remaining support horizon.
• Required quantity = Projected demand × (1 + safety buffer), where the safety buffer reflects forecast uncertainty.
• Procurement quantity = Required quantity ÷ incoming acceptance yield, rounded up.
• Final LTB = Procurement quantity − usable stock on hand, rounded to the OEM minimum order quantity.

Each input carries a fab-specific nuance. Production and PM demand is the most reliable element because it is driven by scheduled preventive-maintenance intervals and historical wafer throughput. Field-failure and repair-loop demand depends on the part’s failure mode: a wear item such as an O-ring set or a process kit is consumed predictably, whereas a controller board fails sporadically and is harder to forecast. The safety buffer typically sits in the 10–30% range, widening as the support horizon lengthens and the failure history thins.

Worked example: an RF generator control board

Consider a fab running eight legacy etch chambers, each with one discontinued RF generator control board, and a planned support horizon of eight years before the toolset is upgraded.

• Installed base: 8 boards in service.
• Historical field-failure rate: roughly 1 board every 2 years across the fleet, giving 4 failures over 8 years.
• Repair-loop attrition: 1 board scrapped during board-level repair over the period.
• Qualification and engineering spares: 1 board.
• Projected demand: 4 + 1 + 1 = 6 boards.
• Safety buffer at 25% (long horizon, sparse failure data): 6 × 1.25 = 7.5, rounded to 8.
• Incoming acceptance yield 95%: 8 ÷ 0.95 = 8.4, rounded up to 9.
• Usable stock on hand: 1 board. Final LTB = 9 − 1 = 8 boards.

The figure is illustrative, not a benchmark, every fleet’s failure history differs, but the method is the point. It forces each demand driver to be named, quantified and defended, which is exactly what an auditor or a finance reviewer will ask for when the LTB requisition crosses their desk.

The carrying cost of a last-time buy

Carrying cost is the annual expense of holding inventory: capital tied up, storage, insurance, handling, and the risk of obsolescence or degradation while the part sits on a shelf. For an LTB, carrying cost is not a footnote, it is often the deciding factor against a large stockpile, because the parts may be held for the better part of a decade.

Industry practice puts annual inventory carrying cost in the region of 15–30% of the inventory’s value, depending on the storage regime and the part’s sensitivity. Over an eight-year hold, that compounds into a material fraction of the original purchase price. A simple model makes the exposure visible:

• Annual carrying cost = LTB inventory value × carrying-cost rate.
• Lifetime carrying cost ≈ average inventory value over the hold × carrying-cost rate × years held.

Because inventory is consumed over the hold rather than sitting full the whole time, the average value is lower than the purchase value, but the multiplier still bites. A £100, 000 LTB held with an average inventory value of £55, 000 at a 20% carrying rate over eight years carries roughly £88, 000 in lifetime holding cost on top of the purchase. That total cost of ownership is the number to compare against the up-front qualification cost of a second source, framed through the SEMI E35 Cost of Ownership methodology, which formalises how acquisition, operating and obsolescence costs are aggregated for capital-equipment decisions.

Shelf-life and degradation risk

Shelf-life is the period a part remains fit for installation while in storage. It is the quiet failure mode of an over-sized LTB: parts bought to defeat obsolescence can themselves expire before they are used. Several categories common to fab spares are vulnerable.

• Elastomers and seals, O-rings and gaskets harden and lose compression set; manufacturer shelf-life ratings often run a few years from cure date.
• Electrolytic capacitors on control boards dry out and lose capacitance over long unpowered storage, and may need reforming before use.
• Moisture-sensitive electronics require controlled humidity and dry-pack storage to avoid solder-joint damage on later reflow.
• Lubricated assemblies and pumps can suffer grease separation or seal set if stored for years without rotation.

The practical implications are concrete. Shelf-life caps how large an LTB can sensibly be, because anything beyond the usable storage window is pre-committed scrap. It demands a storage regime matched to the part, climate control, dry-pack, first-in-first-out rotation and date-of-manufacture tracking. And it tilts long-horizon, shelf-life-limited demand towards a second source, which is built to order and never sits ageing on a shelf. A part with a five-year shelf life is a poor candidate for an eight-year stockpile.

Obsolescence risk: which route reduces exposure

The obsolescence risk of an LTB is concentrated and one-shot. Once the last-order-date passes, the forecast is locked. If demand was under-estimated, or a tool’s failure rate rises, or the support horizon extends, the fab is left with a stranded tool and no further supply, the worst outcome obsolescence management exists to prevent. An LTB therefore does not remove obsolescence risk; it converts it from a supply problem into a forecasting problem, and concentrates the consequences of a bad forecast at the far end of the hold.

A qualified second source attacks the risk at its root. By establishing an alternative, ongoing supply of a form-fit-function equivalent, it removes the dependency on a single discontinued part and on a single forecast. This is why obsolescence-management frameworks such as IEC 62402, the international standard for obsolescence management, treat proactive second-sourcing and design resilience as primary mitigations and position last-time buys as a tactical, time-bounded bridge rather than a permanent solution. The deeper context for that framework is covered in the guide to semiconductor equipment obsolescence.

Second-sourcing carries its own risk, the equivalent must be genuinely qualified, not merely similar. Qualification evidence, copy-exact discipline and form-fit-function verification are what separate a legitimate engineering second source from an unverified open-market substitute. Treated as a qualified engineering route, second-sourcing consolidates the supply chain and improves uptime without disparaging the OEM design it replaces.

A decision tree: LTB, second source, or both

The two routes resolve into a small set of decisions driven by remaining demand, the forecast horizon, and the availability of a qualified equivalent.

1. Is a fab-qualified form-fit-function equivalent already available? If yes, default to the second source and hold only a minimal bridge stock. The forecast risk and carrying cost of a large LTB are unnecessary when a qualified alternative exists.
2. If not, is remaining demand modest and the horizon short (roughly under three years)? If yes, an LTB is usually the cleaner answer. Forecast error over a short horizon is small, carrying cost is limited, and qualification effort is avoided.
3. Is remaining demand high, or the horizon long (five years or more), or the part shelf-life-limited? If yes, favour qualifying a second source. A large, long-held LTB stacks forecast risk, carrying cost and shelf-life risk against you simultaneously.
4. Is demand uncertain or the part critical with no equivalent yet? Run both: size a modest LTB to cover the qualification window, and start second-source qualification in parallel. The LTB de-risks the gap; the second source removes the long tail.

This logic mirrors the broader obsolescence-management literature, which consistently recommends matching the response to remaining demand and part criticality rather than defaulting to a single tactic. The route chosen should be recorded in the tool’s obsolescence risk register so the decision, its forecast basis and its review date are auditable.

Frequently asked questions

What is a last-time buy in semiconductor procurement?

A last-time buy is the final order placed against a part before its manufacturer discontinues it, sized to cover all remaining demand for the rest of the affected tool’s service life. The manufacturer sets a last-order-date and last-ship-date; the buyer must forecast lifetime demand and place one bridge order to cover it.

How do you calculate last-time-buy quantity?

Sum projected demand across the support horizon, production and PM use, field-failure replacements, repair-loop attrition, and qualification and training spares. Multiply by a safety buffer for forecast uncertainty (commonly 10–30%), divide by the incoming acceptance yield, then subtract usable stock on hand and round to the supplier’s minimum order quantity.

Is a last-time buy cheaper than buying second-source parts later?

At the unit level, parts bought during an LTB window are typically far cheaper than sourcing the same item years later on the secondary market. But the LTB also carries years of inventory carrying cost (around 15–30% of value per year) and the risk of over- or under-buying. A like-for-like comparison must weigh the up-front purchase plus lifetime carrying cost against the qualification cost of a second source.

What is the difference between a last-time buy and a second source?

A last-time buy stockpiles the original OEM part before it is discontinued. A second source is a qualified form-fit-function equivalent from an alternative manufacturer, supplied on an ongoing basis. The LTB removes qualification effort but concentrates forecast and shelf-life risk; the second source requires qualification but removes the long-term dependency on a single discontinued part.

How long should a last-time buy cover?

It should cover the period from the last-ship-date until the tool is retired, upgraded, or supplied by a qualified second source. For many 200mm and legacy tools this is several years, which is why forecast error and shelf-life become central concerns. Where the horizon is long, a smaller LTB paired with a second-source programme usually beats a single large stockpile.

What are the risks of a large last-time buy?

The main risks are over-buying (capital written off as inventory expires unused), under-buying (the tool is stranded with no further supply), and shelf-life degradation of stored parts such as elastomers, electrolytic capacitors and moisture-sensitive electronics. Sizing the order with a defensible formula and matching storage to the part’s shelf-life mitigates all three.

Managing discontinued fab parts with ESCT

Deciding between a last-time buy and a second source is rarely a one-tool problem, it is a programme that runs across a whole legacy install base. ESCT Fab Support helps 200mm and legacy LAM, TEL and Applied Materials owners size bridge buys, qualify form-fit-function second-source parts, and consolidate supply for discontinued components so tools stay productive past OEM end-of-life. To review the obsolescence exposure on your toolset or scope a second-source qualification, explore our obsolescence management resources or contact the ESCT team for a quote.