Why Your Supply Chain Is Hiding a 5-Year Shortage

A drug shortage used to last about two years. It now lasts five and a half.

That is the figure that should stop you, and it comes from the organisation that tracks it most carefully: 5.3 years on average, the longest since anyone began counting. But the number underneath it is worse. Of the 75 medicines in shortage at the end of last year, only four of them started last year. The other seventy-one were already there. The list is not turning over. It is setting.

Headline counts fell, 98 down to 75, and almost every report led with that. Read it the other way. Fewer new shortages, and the old ones will not clear. That is not a market recovering. That is a backlog hardening.

And the reason is not printed anywhere on the box.

Take one vial. Cefotaxime: an injectable antibiotic, a small glass vial, nothing exotic about it. It has been in shortage for more than ten years.

Look at the carton and you will find the country where the active ingredient is made. That is accurate. The plant is real, the inspection record is real, and if you audited that facility you would find a competent operation with its paperwork in order.

Now walk to the other end of that plant, to the goods-in gate. Drums arrive there. Inside the drums is the starting chemistry: the powder that plant buys so it can make the molecule you eventually put in the vial. Almost all of it is made somewhere else entirely.

Two gates. Two maps. One box.

Here is the part I want you to sit with, because I have seen the same shape on factory floors that have nothing to do with medicine. Open the quality folder. The certificate of analysis names the maker of the active ingredient. It is signed, it is dated, it is legitimate. And there is nothing in that folder, not one page, that names who made what went *into* it. The document set ends exactly where your risk begins.

You did not skip a step. You completed every step that exists.

Here is what most Western buyers still miss.

There are three layers, and it is worth naming them once because the whole thing turns on telling them apart.

The **finished dose** is the thing with a label: the pill, the vial, the box with a lot number.

The **API** is the active pharmaceutical ingredient, the molecule that actually does the work.

The **KSM** is the key starting material, the foundational chemical the API gets synthesised from. It is the input to the input.

Your visibility usually stops at the first layer. If you are diligent, it reaches the second. Your exposure lives in the third.

Now the numbers, in the order that matters. Fifty-eight percent of the key starting materials behind medicines approved in the United States come from a single country only. Half of those approved active ingredients depend on one origin for at least one starting material. Seventeen percent depend on one origin for every single one of them.

Then look at the shortage list itself. Forty-four percent, 33 of 75, have at least one starting material from a single source. Six of them depend on one source for all of it.

The most useful thing in that report is not any single percentage. It is that 2026 was the first year anyone mapped this layer at all. The analysts' own words: concentration "often begins far earlier in the supply chain than previously understood." Nobody was hiding it. Nobody had looked.

And this is the part that has nothing to do with medicine. Swap "API" for cell, casting, pigment, bearing, control board, resin, magnet. The structure does not change. You audited the assembler and you inherited the input. A chain that looks diversified at the product level can be a single thread at the chemistry level. And diversification you cannot see is not diversification.

The 7-Factor Supplier Checklist: the checks I run before a deposit moves.

Free. Plus the Monday briefing from Shanghai.

Now let me be fair, because the obvious reading of this is the wrong one.

If you assume concentration is something that happens in one part of the world, the data does not support you. Among the medicines actually in shortage, the European Union is the primary source of the active ingredient for fifteen of them, more than any other origin. India is primary for seven. The origin most readers arrive expecting to find is primary for none of them.

Concentration is not a national characteristic. It is a shape supply chains take under sustained price pressure, and it takes that shape everywhere.

Which brings me to the cause, and it is not geography at all. Discontinuations rose sixty percent in a single year, the highest since 2019. Sixty-five percent of the products discontinued were priced under one dollar a unit. That is the whole mechanism in one line: a product that sells for eighty cents cannot fund a second supplier. Redundancy is a cost, somebody has to carry it, and at that price nobody can. So the second source quietly closes, and what is left is a single thread that looks fine right up until the day it doesn't.

Nobody decided this. It was priced into existence.

I want to be straight about the limits of what this data says, because the one thing a well-informed reader would catch is an overclaim.

The report establishes that cefotaxime's starting chemistry is concentrated, and it establishes that cefotaxime has been short for a decade. It does not say the first caused the second. Neither will I. The same caution applies to the 44%. That is a co-occurrence across the shortage list, not a proven cause of it. What concentration gives you is **exposure**: a map of where a chain has one thread instead of two. Whether that thread breaks is a separate question, and anyone who tells you they can predict it is guessing.

There is also a scale figure worth knowing. A separate list of a hundred vulnerable medicines found that forty-eight of them depend on starting materials from at least one single origin. And only thirty of that hundred were actually in shortage. The other seventy were listed purely because of what sits upstream. Nearly three-quarters are classified as essential medicines.

That gap, 48 exposed, 30 in trouble, is the honest picture. Exposure is not failure. It is the set of places failure would come from if it came.

So the question is not whether your chain will break. It is whether you could name the thread if it did.

So here is the move, and neither part of it requires you to relocate anything or predict anything. The advice is not "source somewhere else." It is "see what you already have."

One. Pick your three highest-volume inputs and ask each supplier one question, in writing: who makes the material you buy to make this, and where? Not "where is this made". You already know that, and it is the answer that hides the problem. You want the tier below them. Put it in an email so the answer exists on paper. If they cannot answer it in writing, that silence is the finding: they do not know their own chain, which means neither do you. The 7-Factor Supplier Vetting System is the checklist I run to pressure-test a supplier before I rely on one.

Two. Write the answer into the contract so it stays true after the first order. A supplier who changes their input source mid-programme has changed your risk without telling you, and most agreements give them every right to. The clause you want names the input, requires disclosure of the sub-supplier, and requires notice before it changes. That is standard practice in an OEM supply agreement. And if you are buying to your own specification, it belongs in the manufacturing agreement as well.

If you would rather I ran it (put the supplier, the quote and the sub-tier through the full system and hand you a written verdict), that is the Supplier Risk Diagnostic, $497. And when the input matters enough that you want eyes on the line, I go to the factory and check it myself: Factory Verification, from $1,500.

You cannot re-engineer a supply chain from your desk, and you should be suspicious of anyone who tells you to try. But you can find out, this month, which of your inputs sit on one thread. And that is the entire difference between the operators who get surprised and the ones who don't.

The shortage was never really the surprise. The surprise is always the same: it came from a tier nobody had a document for.