MERIDIAN — Business-model what-ifs

Every option in the analysis behind the memo was generated along one axis: what MERIDIAN can make, and who buys it. Fifty-seven candidates, three lines, one participation tier. All of them are the same business with different products in it. MERIDIAN patterns film, sells the film to somebody who owns a clearance, and is paid a few cents of the price the patient's insurer eventually sees.

The other axis is the shape of the business itself. The memo's two decisive findings are both properties of that shape rather than of the products. What MERIDIAN would be paid is a few percent of what the device sells for: about $0.06 of a $0.30 ECG electrode, about $0.60 of a $2.00 return electrode. And the tier MERIDIAN would enter earns less than the tier it already occupies, because Integer at 11.9% and Zhen Ding at 7.6% bracket where a merchant manufacturer lands, whatever it manufactures. Neither finding is about CGM, or about neurostimulation, or about which segment gets prioritised. Both would survive a perfect product choice.

So the question is: what shape could this business take, other than a component supplier selling into the medical device chain?

Six answers follow. Each carries a precedent where one exists, and where the search returned nothing the structural argument stands alone and says so.

The shape. Do not become a supplier to the branded and private-label electrode business. Buy one. Acquire a company that already holds the 510(k) clearances, the hospital and distributor customers, the brand and a live ISO 13485 certificate, and put MERIDIAN's roll-to-roll capacity behind it. The economics change at the first line of the income statement: instead of being paid $0.06 of a $0.30 electrode, the business books the $0.30 and MERIDIAN's manufacturing becomes the cost line rather than the revenue line.

What would have to be true. Three things. Businesses of that description have to be available at a price a $6B company will pay. The brief's boundary against a finished-device OEM strategy has to be readable as a boundary against implantable and therapeutic devices rather than against every product that carries a clearance, because a disposable ECG electrode is a device with a clearance and MERIDIAN would own it. And MERIDIAN's manufacturing has to actually lower the acquired business's unit cost, which is the same untested claim the memo flags as the largest hole in Line A.

Whether it looks true here. The precedent is exact and it is already in the memo, under a different heading. Nissha, a Japanese converting and printing company whose core business was applying those technologies to smartphones and tablets, agreed on 5 August 2016 to acquire Graphic Controls Holdings for USD 135 million. Graphic Controls was a Buffalo business founded in 1909 making patient-monitoring disposable electrodes and surgical consumables, with USD 153.5 million of sales in 2015 and about a thousand employees. Nissha added HeartSync, a defibrillation electrode maker, in May 2018, and a second medical contract manufacturer in June 2018. Nissha Medical Technologies is the company the memo names as the incumbent Line A would compete against, and as a printing and film company that entered medical electrodes from the same kind of starting asset.

The memo reads that as a precedent for the move. It is a precedent for a different move than the one recommended. Nissha did not qualify as a supplier to the electrode industry. It bought the industry position outright, with the clearances and the customers attached, at roughly nine tenths of one year's revenue, and it arrived at scale on the closing date rather than after a two-year certificate calendar.

What was true in that case and would have to be true here: the target's value sat in clearances, brands and customer relationships that a converter could not build, while its manufacturing was the part a converter could improve. That is the same asymmetry MERIDIAN faces. What differs is the market for the asset. In 2016 Nissha was buying from a seller in a quiet category. Medtech contract manufacturing has since become a private-equity category: one industry count records 64 deals completed in the medtech CDMO space in 2024 against 93 private-equity-backed platforms. That does not contradict the memo's finding that strategic acquirers do not compete for component suppliers. It says who does compete for them, and financial buyers set prices off cash flow rather than off strategic fit. Whether anything is available near Nissha's ratio is the first thing to check and it can be checked from a screen.

Set against the recommended path, the comparison is not close on its face. That path consumes $221M of cash, reaches $164.8M of revenue in 2032 at a negative operating margin, and returns about -$99M. A business of Graphic Controls' 2015 scale cost less than that and had the revenue on day one, at a margin a branded consumables business earns rather than a merchant converter's.

Verdict: worth a look, and it is the one worth taking furthest. It uses the asset the memo never spends, which is the balance sheet, and it attacks both decisive findings at once rather than either alone.

The shape. The memo's central obstacle is that the companies worth supplying make the part themselves. Treat that as an asset register rather than as a closed door. Buy the plant. A carve-out of an OEM's in-house electrode or consumables operation delivers the certificate, the qualified process, the trained people, the customer and the revenue in one transaction, because the seller signs a multi-year supply agreement as part of the price.

What would have to be true. There has to be a seller, and a seller only appears when an OEM decides the part is no longer what it competes on.

Whether it looks true here. The mechanism precedent is clean and it is from electronics rather than from medicine. Celestica began as IBM's Canadian manufacturing arm and was sold to Onex for $550 million in 1996; in February and May 2000 it bought further IBM assets at Rochester, Minnesota and at Vimercate and Santa Palomba in Italy for $470 million in total, signing two three-year strategic supply agreements with IBM at the same time. What was true in that case is that IBM had concluded box assembly was not what it competed on and wanted the plants off its balance sheet, so the buyer acquired capability and demand together.

Whether that condition holds here is answered segment by segment by the memo's own sourcing rule, and mostly it does not. Abbott opened a facility specifically to meet CGM demand, which is a company investing in the part rather than exiting it. Dexcom's sensor is its product. Nobody is selling those.

Where it could hold is the commodity end, at branded manufacturers for whom disposable electrodes are a small consumables line inside a large portfolio. Portfolio pruning at that end does happen: Nordson agreed to divest selected product lines within its medical contract manufacturing business, stating the intent to concentrate on proprietary medical components. I searched for a named case of a medical device OEM divesting an electrode or sensor plant with a supply agreement attached and did not find one. So the category is live and the specific instance is unevidenced, which is the honest position.

Verdict: worth a look, as a standing mandate rather than as a plan. It cannot be scheduled, because it depends on somebody else's decision to sell. What it costs to hold open is a screen and a banker relationship, and the same 93 private-equity platforms that raise the price on shape 1 are the ones generating deal flow to see.

The shape. The memo's cash curve is the problem, not the market. $160M of capital and a $14M-to-$36M fixed organisation go in ahead of revenue, capacity is built on a forecast, and free cash flow is negative in all six years. Invert the financing. Sell dedicated capacity rather than parts: the customer funds or shares the fit-out of a line configured to its specification and commits to a volume, MERIDIAN operates the line and keeps the process knowledge, and the capacity risk sits with the party that has the demand forecast.

What would have to be true. Buyers who value security of supply more than unit price, who can fund capital themselves, and who will sign a take-or-pay commitment. And MERIDIAN willing to accept a lower margin ceiling in return for a certain one.

Whether it looks true here. The model exists and is documented. Lonza's Ibex offer builds customer-dedicated suites in which the fit-out cost is borne by the customer or shared depending on the commercial model, which converts what would be the customer's capital expenditure into operating expenditure and lets it avoid building a plant. Moderna signed a ten-year agreement on that basis in May 2020, and Lonza subsequently committed a $415 million expansion of the site.

What was true in that case: the customer faced a capacity emergency, had no plant, and had a product whose volume it could not itself serve. Of MERIDIAN's three lines, only Line B has buyers with the first two properties, since pharmaceutical and diagnostics companies genuinely have no electronics plants and no intention of building one. None of them has the third. Nobody launching an on-body injector is in the position Moderna was in. In private-label electrodes the proposition is weakest of all, because the buyer's alternative is another converter at a lower price rather than no supply at all.

The value of the shape is therefore not that it wins customers. It is that it refuses to build capacity before somebody has paid for it, and the memo already says the case rests on merchant shares nobody has measured. Ten roll-to-roll lines built against a 20% merchant share estimate with no source behind it is precisely the exposure this structure removes.

Verdict: worth a look, as a financing rule laid over any entry rather than as a business. Stated as a rule: no capacity is built that a customer has not committed to fill. It converts the plan from spending $221M and finding out into building against signed volume, which is the same thing the $5.1M test programme is trying to buy more cheaply.

The shape. Stop selling a component into companies that build devices. Sell a complete platform, being the sensor film, the module, the bonded electronics and a reference design, to buyers who want a branded product and have no manufacturing at all: pharmacy chains, distributors, insurers, health systems, pharmaceutical companies. They own the brand and hold the clearance. MERIDIAN owns the platform and every unit made on it. It is the memo's CGM challenger route taken past a component sale to its conclusion.

What would have to be true. The device has to be commoditised far enough that clinical performance no longer decides the sale, so that a channel owner can win on price and distribution. The platform's differentiating content, being the sensor chemistry and the algorithm, has to be something MERIDIAN owns or can buy. And a channel buyer has to be willing to hold a clearance.

Whether it looks true here. In the neighbouring category it has already happened. Blood glucose monitoring is substantially a private-label industry: Sinocare, founded in 2002 and the largest blood glucose meter manufacturer in Asia, offers OEM and ODM supply, and in 2016 acquired Nipro Diagnostics, now Trividia Health, described as the largest US manufacturer of pharmacy-branded diabetes products. The mechanism that made it possible was time and patent expiry: once meters and strips stopped differing in a way a patient could feel, the brand on the box became the retailer's to own.

CGM is not there, and the closest available evidence points the wrong way. i-SENS, a Korean company founded in 2000 with capacity for 2.2 billion test strips a year across two Korean plants and one in China and exports to more than 110 countries, is exactly the profile of a company that would buy a platform if platforms were bought. It built its own instead: the first Korean CGM, approved by the national regulator in June, reimbursed from July, produced at Songdo from August on an automated line rated at two million units a year, with CGM sales of KRW 6.7 billion in the first half of 2025 against KRW 6.0 billion for all of 2024, and a stated ambition of 10% of the global CGM market by the 2030s. A challenger that builds its own line is not a customer for a platform, and this one is in the same city as the acquisition candidate the memo assesses.

That is the finding on the shape, and it does not close the file. i-SENS has a capacity problem rather than a platform problem: a two-million-unit line against a ten-percent-of-world ambition is two orders of magnitude short. The memo's test programme already calls for approaching three CGM challengers on second-source supply, and this names one worth calling.

Verdict: needs something we do not have. The sensor chemistry and the algorithm are the platform, and MERIDIAN owns neither. Revisit when first-generation CGM patents lapse and the category starts behaving as blood glucose monitoring did. Until then the useful residue is a phone call, not a business.

The shape. Never hold a customer relationship at all. Be the volume manufacturer standing behind the specialists: Cirtec, MicroConnex, NeuroOne and the thin-film startups win the design and own the account, and hand MERIDIAN the units when a product graduates from thousands to millions. The position that blocks entry, being the regulatory record and the design-in history, is one somebody else has already paid for.

What would have to be true. A meaningful population of designs graduating from specialist scale to volume scale, and specialists willing to hand volume to a company large enough to displace them.

Whether it looks true here. The demand-side observation is the memo's own and it is a good one: the module layer is empty because that work needs manufacturing scale a startup does not have, and Cirtec and MicroConnex are built for low-volume implantable and catheter work rather than for hundreds of millions of units. That is a real structural gap.

What is missing is anything flowing through it. No implantable BCI holds a therapeutic marketing authorisation anywhere. Thin-film neural volumes are in the thousands. NeuroOne is the type specimen and it is small. The population of designs waiting to graduate is a handful, and each graduation is years away.

I searched for a merchant roll-to-roll house that built a business manufacturing other suppliers' medical designs at volume, and found only small service printers, such as Eastprint, offering sheet-fed and roll-to-roll printing of electrodes and biosensors with design assistance and prototyping. That is the low end of the same idea, occupied and small. No precedent at scale was found, so the argument here is structural, and the pure-play semiconductor foundry is an analogy rather than a precedent. What was true in that case was a large population of design houses that could not afford a fabrication plant. That population does not exist in medical thin film, and a business model that requires one to appear is a forecast rather than a plan.

Verdict: no, as a business. It is a sensible way to price spare capacity on the interventional line, and nothing more, and it should not be dressed up as a position.

The shape. Change the reason to switch. The memo establishes that being cheaper moves nobody on a part the buyer already makes, and that the entry barrier is the certificate rather than the process. Neither of those blocks a buyer who has to move a part out of a country. Sell the same films MERIDIAN already knows how to make, into supply-chain relocation rather than into cost reduction, and let the certificate be the only thing being bought.

What would have to be true. Buyers actually relocating; the parts they are relocating being ones MERIDIAN makes; and the move happening on a timescale longer than MERIDIAN's certificate calendar, since nothing can be quoted before the certificate exists.

Whether it looks true here. The pressure is documented and it is large. Medical devices and components imported from China faced a US tariff of 54% including the base rate and the penalty layers, with the 2025 peak reported at 145%, and Johnson and Johnson told investors to expect roughly $400 million of tariff cost in 2025, concentrated in its medtech division. The China Plus One posture, meaning diversification into a second country rather than exit, is the standard response.

Two facts cut against it. GlobalData found that 13% of overseas-manufactured medical devices originate in China, so the exposed base is a good deal smaller than the volume of commentary implies. And the United States and China reached a partial agreement in early May reducing tariffs across a broad range of goods, which is the shape of risk that matters most here: the policy can reverse faster than a supplier qualification completes. A buyer who begins moving in 2026 and qualifies MERIDIAN in 2029 may no longer have a reason to have moved.

What this does change is who is worth calling and what is said to them. A buyer already relocating has decided to pay a switching cost that MERIDIAN would otherwise have to overcome, which is the single most expensive obstacle in private-label selling.

Verdict: worth a look, at no incremental cost. It is a question to add to the private-label approaches already in the test programme rather than a shape of its own: ask each buyer whether any current supply is being moved out of China, because that answer identifies which conversations start with the switching cost already paid.

Shape 1. The memo's two decisive findings are that MERIDIAN would capture a few percent of the device price and would earn less than it earns today. Every one of the three recommended lines accepts both. Buying a branded or private-label consumables position rejects both at once, and there is a company that made exactly that move from exactly this starting asset, at a price below what the recommended path burns in cash, arriving at scale on the closing date rather than after a two-year certificate calendar.

It is a sketch and it is not free of the memo's constraints. It requires the brief's finished-device boundary to be read as a boundary against implants rather than against clearances, which is a question for leadership rather than for analysis. It buys a hospital and distributor sales channel MERIDIAN does not have and would have to keep. It carries product liability MERIDIAN currently avoids by never holding a clearance. And it is priced today by financial buyers rather than by the quiet market Nissha bought into, so the ratio that made that transaction work may not survive.

None of that is settled here. What is worth saying is that the first test costs a screen: how many businesses of that description exist in North America and Europe, at what revenue, at what multiple, and how many are already owned by a platform that will not sell. That is a fortnight of work, it is smaller than any item in the memo's test programme, and it addresses a larger part of the gap than all of them together.

• Nissha's acquisition of Graphic Controls Holdings, including price, Graphic Controls' 2015 sales, and Nissha's own description of its core converting and printing business. https://www.nissha.com/english/news/2016/08/5th_1.html
• Nissha's subsequent acquisitions of a defibrillation electrode manufacturer and a medical device contract manufacturer in 2018. https://www.nissha.com/english/news/2018/05/21yi_1.html and https://www.nissha.com/english/news/2018/06/27th_1.html
• Celestica's origin as IBM's manufacturing arm, the 1996 sale to Onex, and the 2000 acquisition of IBM plants with three-year supply agreements attached. https://www.theglobeandmail.com/report-on-business/celestica-buys-three-ibm-plants/article25453263/ and https://www.sec.gov/Archives/edgar/data/0001030894/000091205702012827/a2074563zex-99_4.txt
• Nordson's divestiture of selected product lines within its medical contract manufacturing business. https://www.sec.gov/Archives/edgar/data/72331/000007233125000057/ndsn-20250430.htm
• Medtech CDMO transaction volume and private-equity platform count. https://pharmasource.global/content/guides/category-guide/medical-devices-cdmo-contract-services-market-report-oem-outsourcing-guide-2026/
• Lonza's customer-dedicated suite model and the funding of fit-out by the customer. https://www.lonza.com/news/2020-12-02-06-00 and https://www.bioprocessintl.com/facilities-capacity/lonza-targets-both-early-and-late-stage-biotech-in-415m-ibex-expansion
• Moderna's ten-year manufacturing agreement with Lonza. https://www.bioprocessintl.com/facilities-capacity/moderna-contracts-lonza-to-help-scale-up-covid-19-mrna-vaccine-candidate
• Sinocare's OEM and ODM position and its acquisition of Nipro Diagnostics, now Trividia Health. https://en.sinocare.com/ and https://www.trividiahealth.com/blood-glucose-monitoring-system/
• i-SENS test strip capacity, export reach, and the approval, reimbursement and Songdo production of the first Korean CGM. https://www.koreabiomed.com/news/articleView.html?idxno=21390 and https://www.mobihealthnews.com/news/asia/south-korea-approves-first-local-cgm-device
• Merchant roll-to-roll and screen printing services for electrodes and biosensors. https://www.eastprint.com/electrodes-biosensors/
• Tariff rates on medical devices and components from China, and the reported cost to Johnson and Johnson. https://meddeviceguide.com/blog/medical-device-tariffs-trade-war-impact-2026-guide and https://www.fiercebiotech.com/medtech/tariff-driven-uncertainty-rattles-medtech-industry
• Share of overseas-manufactured medical devices originating in China, and the partial tariff agreement. https://www.supplychaindive.com/news/one-year-in-how-medtech-companies-are-coping-with-tariff-challenges/817194/

Searches that returned no precedent, recorded because their absence is part of the argument: a named medical device OEM divesting an electrode or sensor plant with a supply agreement attached, for shape 2; and a merchant roll-to-roll manufacturer operating at volume on other medical suppliers' designs, for shape

1. Both arguments above are structural where the precedent is missing, and say

so at the point of use.