The Lu-177 Capacity Number Nobody Publishes
The industry reports grams of Yb-176, reactor access, production weeks and patient doses. None of those numbers answers the only question a new sponsor actually has: how much qualified, uncommitted, GMP-released Lu-177 can I contract, and when.
Byron Fitzgerald
Founder, ProGen Search
The industry reports grams of Yb-176, reactor access, production weeks and patient doses. None of those numbers answers the only question a new sponsor actually has: how much qualified, uncommitted, GMP-released Lu-177 can I contract, and when. The market is not the capacity that exists. It is the capacity you can qualify and contract.
Six numbers. Six different questions.
Ask how much Lu-177 capacity is available and the public record will happily answer. Here is what it offers, as of August 2026.
| Source | The public number | What kind of number it is |
|---|---|---|
| Kinectrics | Annual Yb-176 capacity above 500 g by end-2025 | Nameplate capacity, mass |
| TMC Group | More than 2 kg of elemental Yb-176 produced in 2025 at above 99.7% isotopic purity | Company-reported output, mass |
| ASP Isotopes | Initial product expected in H2 2026 | Forward guidance, timing |
| ITM | Priority access to half of ILL's available neutron-irradiation capacity for n.c.a. Lu-177 | Contractual share of an undefined pool |
| SHINE | Capacity for 100,000 doses annually, expandable to 200,000 | Company-defined commercial unit |
| MURR | GMP n.c.a. Lu-177 produced 52 weeks a year | Operating continuity |
Every one of those statements is defensible inside its own definition. Not one of them can be added to another. A gram is not a becquerel. A reactor-access percentage is not released product. A dose is not a patient. Fifty-two weeks is not a batch size.
The awkward number is TMC's. Its company-reported 2025 output is roughly four times the far more widely quoted 500 g headline. That is not a like-for-like comparison, and pointing that out is not a defence of the comparison. It is the entire problem. One figure is reported production. The other is announced capacity. They have been circulating in the same conversations, in the same slide decks, as though they measure the same thing.
Figure 1
Six headline numbers, six unit families
Four of the six public numbers cannot be converted into activity at all. The other two convert only if you accept an undisclosed in-house assumption.
The two-kilogram hole in the market map
TMC Group is a private North Vancouver manufacturer that has been enriching stable isotopes since 2004. In May 2026 it stated that it produced more than two kilograms of elemental Yb-176 in 2025, at isotopic purity exceeding 99.7%, and that 2026 volumes remain steady. It also said it is commissioning new lines that will add 50% to total company capacity, a figure that covers Yb-176, Ga-69 and Gd-160 together and should not be read as a 50% uplift in Yb specifically.
That single disclosure does two things. It ends any version of the market map in which North American Yb-176 supply is a story about one or two listed names. And it demonstrates how easily a research process anchored on the loudest disclosures will miss a quieter, larger one. It is the same failure mode we described in the ghost capacity problem, inverted: there, announced capacity overstated what could be scheduled. Here, the loudest announcement understated who was actually producing.
There is a second TMC disclosure that has attracted almost no attention and is arguably more commercially useful than the tonnage. In October 2025 the company stated it had more than 250 grams of Yb-176 in stock and available for purchase.
That is close to the only public number in this entire market that describes uncommitted, merchant-available inventory. Everything else describes what has been built, what has been produced, or what has been announced. This describes what someone could actually buy.
There is a third triangulation available, and it costs nothing. TMC's cumulative shipped figure moved from “more than 10 kilograms” in October 2025 to “more than 13 kilograms” in May 2026, across all products. Both are floor figures, so the delta is soft. But it is a shipment signal against a production claim, and shipment signals are rarer and more informative than production claims. Anyone maintaining a serious supply map should be logging those cumulative figures every time a company restates them, precisely so the deltas can be read later.
What the two-kilogram figure proves: TMC reports commercial-scale current production, at high purity, and the North American supplier universe is wider than the coverage implies.
What it does not prove: that the material is uncommitted, that it is in a chemical form every conversion route accepts, that it has been qualified in any given sponsor's quality system, or that precursor is no longer strategic.
Why grams never become doses in a straight line
There are two established reactor routes to Lu-177, and conflating them is the fastest way to publish something wrong.
The direct route irradiates enriched Lu-176. Target and product are the same element, so processing is comparatively simple and the neutron-capture cross section is high, but the product is carrier-added and carries a lower specific activity plus some long-lived Lu-177m.
The indirect route irradiates enriched Yb-176, which captures a neutron to form Yb-177, which decays to Lu-177. Because ytterbium and lutetium are different elements, the lutetium can be chemically separated from the bulk target. That is what makes no-carrier-added, high-specific-activity material possible, and it is why Yb-176 matters at all.
The cross sections are not close. Published thermal neutron-capture values sit at roughly 2.5 to 2.85 barns for Yb-176, against roughly 2,065 to 2,090 barns for Lu-176. The exact figure depends on the nuclear data source and the spectrum. The strategic point is the order of magnitude, not the decimal: the indirect route converts a very small fraction of a very expensive target, which is exactly why recovery and recycling of the unconverted ytterbium sit at the commercial centre of the process rather than at its edge.
So a gram of Yb-176 is not an input that gets consumed into doses. It is working capital that goes round a loop.
Figure 2
The n.c.a. Lu-177 conversion route, with the recovery loop
Every stage has its own capacity, qualification status, yield, owner and access rights. The binding constraint moves.
The only public grams-to-patients bridge, and what it implies
There is exactly one widely quoted public conversion from mass to clinical output in this market. Kinectrics said that annual Yb-176 capacity above 500 g would support more than 150,000 patient treatments annually.
Run the arithmetic. That is roughly 300 treatments per gram per year. At the labelled 7.4 GBq per administration for Pluvicto and Lutathera, 300 administrations implies about 2,220 GBq, or roughly 60 curies of administered activity, per gram of Yb-176 per year.
Now set that against the production literature. Published campaign data for high-flux indirect production reports on the order of 210 to 230 Ci of Lu-177 per gram of ytterbium at end of irradiation, for a single irradiation of around eight effective days. If a gram of target could be cycled even a handful of times a year, the gross figure would be an order of magnitude above 60 Ci.
We are not saying the Kinectrics conversion is wrong. We have no basis to say that, and it may well be conservative. We are saying that the distance between those two numbers is filled entirely with assumptions that nobody publishes: neutron flux and spectrum, cycle length, how many times a gram is realistically recycled before requalification, separation and recovery yield, hot-cell throughput, quality release pass rate, decay between calibration and administration, and whether “patient treatment” means an administration or a completed course.
That gap is the article in one calculation. The public grams number and the public patients number are separated by six to eight unpublished variables, and the answer swings by an order of magnitude depending on what you assume for them. The tool below lets you move them yourself.
Interactive tool
Route Yield Translator
Take a mass of Yb-176 and watch it collapse into doses a new sponsor could actually contract, one real-world factor at a time. The point is not precision. The point is how far the answer moves on assumptions nobody publishes.
Preset routes
The spread, at your current target mass and dose size
117,420
Optimistic route
9,168
Realistic route
441
Constrained route
Same target mass. Same isotope. Same arithmetic. The three answers differ by roughly 267x, and every input that separates them is commercially confidential.
Assumptions
Doses a new sponsor could theoretically contract per year
9,168
67.8 TBq per year, or 35.3 Ci per week at the stated calibration time.
Where it goes
Gross activity produced
20,000 Ci/yr
Target mass, times yield at end of irradiation, times cycles per year
After separation and recovery
17,600 Ci/yr-12%
88% separation and recovery yield
After quality release
16,720 Ci/yr-5%
95% release pass rate
After decay to administration
12,224 Ci/yr-27%
3 days from calibration, against a 6.64-day half-life
After uncommitted share
1,834 Ci/yr-85%
15% not already captive, exclusive, allocated or contracted
Assumption sensitivity
From where the sliders currently sit, one notch of uncommitted merchant share moves the answer by 33.3%. That is the question to ask first.
This is a sensitivity model, not a plant engineering calculation. It exists to show how far the answer moves on assumptions that are not publicly disclosed. Do not use it to size a supply agreement. Use it to work out which questions to ask before you sign one.
The decay clock turns logistics into manufacturing
Lu-177 has a half-life of about 6.64 days. That single fact converts every logistics decision into a manufacturing decision.
Activity is decaying while the batch is tested, while it waits for release, while it is packaged, while it clears customs, while it sits on a Sunday, and while a patient's appointment slips by two days. A capacity figure quoted without a reference time is not a capacity figure at all. It is a marketing figure. This is the same physics that makes radiopharmaceutical manufacturing an execution problem rather than a capital one, which we set out in execution under decay.
| Elapsed time | Activity remaining | What it costs you |
|---|---|---|
| 1 day | 90% | A routine release delay |
| 2 days | 81% | One transatlantic hop with a customs hold |
| 3 days | 73% | More than a quarter, gone |
| 5 days | 59% | A weekend plus a rescheduled patient |
| 7 days | 48% | Half the batch, on paper and in the vial |
| 10 days | 35% | Two thirds |
Figure 3
What the 6.64-day half-life does to a batch
Two published activity figures are only comparable if they share a calibration time. They rarely do.
There is no single reactor queue
The tempting story, and one we deliberately abandoned in the research, is that precursor supply is easing and the constraint has already migrated downstream into a single crowded irradiation queue. The evidence does not support it.
The downstream network is expanding on multiple, independent clocks.
- ILL and ITM, France and Germany. ITM holds priority access to half of ILL's available neutron-irradiation capacity for this route. Read that wording carefully. “Available” may already exclude research allocations, maintenance and other commitments. “Priority access” is not ownership. The agreement proves high-flux access is worth locking up contractually. It does not prove the residual pool is empty.
- Bruce Power, Isogen and ITM, Canada. Commercial production began in 2022; a second Unit 7 line entered service in 2024; an on-site hot cell completed construction in July 2026 and is moving through licensing and commissioning, with target-carrier removal as its first stated role. That is a paired investment: more irradiation creates more post-irradiation handling demand, and the pairing is more informative than either announcement alone. It is also the clearest example of the shift we described in the reactor that makes electricity first.
- MURR, United States. Made GMP n.c.a. Lu-177 available globally in March 2026 and states production 52 weeks a year. Weekly batch size and uncommitted share are not disclosed.
- SHINE, United States. Commercial n.c.a. Lu-177 through Cassiopeia, with company-stated capacity for 100,000 doses annually and potential expansion to 200,000. SHINE also enriches its own Yb-176 and recycles it internally, which makes it a vertically integrated source rather than an open merchant one. Ilumira received EU marketing authorisation as a precursor in March 2026.
- ITM NOVA, Germany; NTP and SAFARI-1, South Africa; NRG PALLAS with Global Morpho Pharma, Netherlands, targeting first commercial delivery in 2027; and Laurentis at Darlington, Canada, targeting 2027.
Six or seven routes, expanding, on different clocks, under different access models, published in incompatible units. What the evidence supports is not “the bottleneck has moved”. It is something more uncomfortable and more useful: the public record does not reveal where the marginal constraint sits, and the constraint is probably episodic and sponsor-specific rather than universal.
A reactor outage dominates one quarter. A separation train maintenance window dominates the next. A failed release, a packaging limitation, an import delay or an unavailable final-product slot constrains one sponsor while global nameplate looks ample. That is the same structure we mapped for the alpha emitters in when actinium stops being the bottleneck, where the binding layer moves from year to year and the shortage can end without the system having scaled.
Capacity is not the market. Access rights are the market.
This is the assertion the whole analysis rests on, so here it is plainly. The physical existence of capacity tells you almost nothing about whether you can buy it.
| Access class | What it means | What it is worth to a new sponsor |
|---|---|---|
| Captive | Consumed inside an integrated company or dedicated product route | Improves system resilience, creates no merchant supply |
| Exclusive | Tied to one counterparty or field by contract | Exists physically, unavailable commercially |
| Restricted | Government, grant or licence conditions limit resale | Supports research and national capability, not an open market |
| Allocated | Committed under long-term agreements | Headline nameplate overstates addressable supply |
| Merchant, qualified | For sale and accepted in your route and quality system | Actionable, subject to lead time |
| Merchant, unqualified | Available but not yet cleared for your process | An option, not supply |
| Prospective | Under construction, commissioning or validation | A scenario, not a source |
| Unknown | Not disclosed | An evidence gap, and it should be recorded as one |
Two suppliers do not give you two routes if they share the same enrichment source, target design, reactor, hot cell, separator, transport lane or release site. This is where most supply “diversification” quietly fails. A backup that is contractually subordinate during a shortage is not a backup. A backup that shares a critical node is not a backup either. It is the same route with two invoices.
Figure 4
From nameplate to contractable
Every stage of this funnel is a commercial negotiation, not a physical fact.
Interactive tool
What is actually addressable?
Nine real sources, each carrying the evidence label it was published with. Untick the filters that matter commercially and watch the market you thought you had.
Sources in your market: 9 of 9
None of the remaining rows discloses uncommitted weekly activity at a stated calibration time.
In your market
TMC Group
Canada
Yb-176 precursorOperatingMerchantMore than 2 kg produced in 2025 (company-reported). More than 250 g in stock and available for purchase, October 2025 (company-reported).
Kinectrics
Canada
Yb-176 precursorOperatingMerchant, uncommitted share undisclosedAnnual capacity above 500 g by end-2025 (nameplate). No statement of how much of it is unallocated.
SHINE Technologies
United States
Yb-176 and n.c.a. Lu-177OperatingCaptive, vertically integratedCapacity for 100,000 doses annually, expandable to 200,000 (company-defined commercial unit). Enriches and recycles its own Yb-176.
ASP Isotopes
South Africa
Yb-176 precursorCommissioningProspectiveInitial product expected in H2 2026 (forward guidance).
DOE / ORNL National Isotope Development Center
United States
Yb-176 precursorOperatingRestrictedListed by quote. Published conditions can restrict resale.
Atom Mines / Eckert & Ziegler
United States
Yb-176 precursorOperatingExclusiveLong-term exclusive supply relationship announced.
ITM via Institut Laue-Langevin
France and Germany
Irradiation accessOperatingAllocated under priority accessPriority access to half of ILL's available neutron-irradiation capacity for n.c.a. Lu-177. The size of the residual pool is not stated.
University of Missouri Research Reactor
United States
n.c.a. Lu-177OperatingMerchantGMP n.c.a. Lu-177 made available worldwide, produced 52 weeks a year. Weekly batch size and uncommitted share are not disclosed.
Orano and Silex
France and Australia
Yb-176 precursorDevelopmentProspectiveDevelopment-stage enrichment projects.
The point
A supplier count is not resilience. Nine names became one or two the moment you applied the only filters that matter commercially.
Every row is drawn from public statements and retains the evidence class it was published with: company-reported, nameplate, guidance or contractual share. Operating status in this market changes quickly. Recheck before you rely on any of it.
Demand is visible. A shortage is not.
Novartis reported Pluvicto sales of $651 million in Q2 2026, up 43% in constant currencies, with $1.293 billion in the first half, up 55%. Lutathera added $225 million in the quarter, up around 9%. On 31 July 2026 the FDA approved Pluvicto in PSMA-positive metastatic hormone-sensitive prostate cancer, an indication Novartis characterised as nearly doubling the eligible US population. Novartis has expanded its radioligand therapy delivery network to more than 880 sites in the US and more than 650 outside it.
That is unambiguous commercial growth. It is not evidence of a current universal shortage, and this article will not claim one. Sales are translated through diagnostic eligibility, treatment-centre throughput, physician adoption, reimbursement, patient health and manufacturing slots before they become isotope demand.
But there is a demand-side development that deserves more attention than it is getting, and it points directly at the uncommitted capacity nobody publishes.
Generic and 505(b)(2) radioligand therapy entrants need their own supply routes. Lantheus received tentative approval for a Lutathera generic in March 2026. A US district court ruled against the validity of key Lutathera patents in June 2026, and Novartis has appealed. No generic has final FDA approval as of this writing. Set the litigation outcome aside: the structural point is that a generic RLT is not like a generic small molecule. The entrant cannot simply out-source a tablet press. It has to secure qualified n.c.a. Lu-177, at a stated calibration time, on a reliable weekly cadence, through a route it can reference in its own regulatory file, and it has to build or buy the delivery network to get a decaying product to a patient on a scheduled day.
Every credible generic RLT programme is therefore a new bidder for exactly the merchant, uncommitted, qualified capacity that this article is about. Originator label expansions increase demand for capacity. Generic entry increases demand for independent capacity, which is a scarcer thing.
Novartis' own comments on the topic were revealing in a different way. Asked about generic RLT competition, the response leaned on supply and delivery network complexity rather than on price. In a market where the grid is the moat, that is the honest answer, and it is the argument we made about the fixed grid when ITM-11 approached approval.
Four ways this develops
The public evidence does not support a single-point forecast. It supports naming the conditions and watching for them.
| Scenario | Conditions | Consequence |
|---|---|---|
| A. Matched expansion | Precursor, targets, reactor access, separation, release and labelling all expand at similar usable rates | More supplier choice, better resilience, competition shifts to service and reliability |
| B. Stranded precursor | Enrichment output grows faster than qualified conversion and merchant access | Kilograms exist, lead times stay long, value accrues to integrated processing routes |
| C. Captive concentration | Large physical capacity ties itself to internal products and exclusive contracts | Headline supply looks abundant while new sponsors face thin addressable capacity and poor terms |
| D. Release and logistics constraint | Irradiation grows, but hot-cell, separation, QC, packaging, cross-border or final-product capacity lags | Activity losses and scheduling failures rise; regional manufacturing and route redundancy gain value |
The leading indicators are readable if you are watching the right things. For B, watch growing precursor inventories, price softness in enrichment and investment in separators and recovery systems. For C, watch vertical integration, offtake agreements and priority rights. For D, watch carrier-removal projects, validated transport packages, regional GMP sites, and hiring in QA, radiochemistry and time-critical logistics.
That last one is worth dwelling on.
Hiring is the leading indicator nobody indexes
Capacity is talent-constrained as well as asset-constrained, and talent moves before output does.
An announced facility is a press release. A facility that is then sustainedly recruiting isotope separation scientists, target engineers, reactor interface specialists, hot-cell automation and maintenance staff, GMP radiochemists, CMC quality and regulatory leaders, radiation safety officers, EU Qualified Persons and time-critical logistics operators is a facility that intends to run.
The inverse holds too, and it is the more useful signal. Persistent vacancies in release-critical functions will delay utilisation long after construction is complete. A hot cell without a licensed, staffed operating team is a building. We have watched sites take twelve to eighteen months longer than announced to reach routine output, and in a meaningful share of those cases the gap was a staffing and qualification gap, not a construction one. The scarcity itself is not new, and we set it out in why radiopharmaceutical talent is the new gold standard, and in more operational detail in how to hire a head of radiochemistry.
If you are trying to work out whether an announced route will actually be available to you in eighteen months, the org chart is better evidence than the nameplate.
A hot cell without a licensed, staffed operating team is a building.
Figure 5
Two suppliers, one route
Dual sourcing that converges on a shared node is single sourcing with extra paperwork.
The questions a sponsor should actually ask
If you take one thing from this piece, make it the shift from asking “how much capacity exists” to asking “how much of this route is mine, and what happens when one node fails”.
On the material
- What is the chemical and physical form, and is it qualified in my process or only in someone else's?
- What was actually shipped and accepted in the last twelve months, as distinct from produced?
- How much is uncommitted, and for how long does that remain true?
- Who owns the ytterbium after recovery, and who carries the loss on a failed batch?
On the route
- Which reactor positions, on what cycle, with what outage history?
- Where does separation happen, in whose hot cell, and what is the queue?
- At what stage is GMP release claimed: precursor, or finished product?
- What is the stated calibration time, and is the price quoted at calibration, at shipment or at receipt?
On access and resilience
- What is my contractual priority during an allocation event?
- Which nodes does my backup share with my primary?
- How long, realistically, from selecting an alternative source to first qualified receipt?
- What regulatory cross-reference support comes with the supply, and does it survive a change of control?
Interactive tool
Supply route resilience scorecard
Score your own route against the twelve questions above. Nothing is stored, submitted or emailed. Close the tab and it is gone.
Route resilience
0 / 24
You do not have a supply chain, you have a purchase order.
On the material
0 / 8
On the route
0 / 8
On access and resilience
0 / 8
Close this one first, worth 2 points
What is the chemical and physical form, and is it qualified in my process or only in someone else’s?
This is a prompt for internal diligence, not an assessment of your supplier. A low score is a list of questions you have not asked yet, not a verdict on anyone's capability.
What we would need before anyone can answer the original question
The capacity number that would actually settle this is one sentence long.
Merchant-available, GMP-released n.c.a. Lu-177 activity per week, at a stated calibration time, after real irradiation and separation yields, recovery losses, outages, quality release and captive allocations, through a named end-to-end route.
Nobody publishes it. That is not an accusation. Most of those inputs are commercially sensitive, some are genuinely unknowable in advance, and no operator has an incentive to publish a number that a competitor can undercut or a customer can hold them to.
But the absence has a consequence. It means that every public comparison in this market is being made in the wrong units, and that a sponsor's real supply position can only be established by direct enquiry, route by route, node by node. Which is precisely why we are rebuilding the Lu-177 map around routes rather than around suppliers: who owns the material, who controls irradiation access, which route is qualified, what is captive or already contracted, and what is genuinely left for the next programme through the door. The current published state of that work sits in the Isotope Production Map and, for the wider modality, in The State of Radiopharmaceuticals.
A market can have kilograms of precursor, new reactor lines and six-figure dose claims, and still offer almost nothing to the next sponsor trying to secure supply.
The market is not the capacity that exists. It is the capacity you can qualify and contract.
Method and evidence standard
This piece is built on primary sources: company releases and filings, regulator records, facility operator statements and peer-reviewed production literature, with a research cutoff of 19 August 2026.
Every claim is classified before it is written. Confirmed operating fact, company-reported actual, nameplate or designed capacity, forward guidance, and analytical inference are kept distinct, and no claim is moved up that ladder. Constructed is not commissioned. Commissioned is not licensed. Licensed is not producing. Producing is not merchant-available. Merchant-available is not uncommitted.
Three limitations are worth stating openly. Corporate output figures are self-reported and not independently audited. Merchant availability is rarely disclosed at all. And operating status in this market changes quickly, so several statuses referenced here should be rechecked before they are relied on.
The original version of this analysis argued that the Lu-177 bottleneck had already moved downstream. We abandoned that thesis when a contradictory fact turned up. It is a better piece for it.
Frequently asked questions
How much Lu-177 capacity is actually available to a new sponsor?
Nobody publishes the figure that would answer this. The public record offers grams of enriched Yb-176, shares of reactor access, operating weeks and company-defined dose counts, and none of those units convert into one another. The number a sponsor needs is merchant-available, GMP-released no-carrier-added Lu-177 activity per week, at a stated calibration time, through a named end-to-end route, after real irradiation and separation yields, recovery losses, outages, quality release and captive allocations. It can currently only be established by direct enquiry, route by route.
What is the difference between carrier-added and no-carrier-added Lu-177?
The direct route irradiates enriched Lu-176. Target and product are the same element, so processing is comparatively simple and the neutron-capture cross section is high, but the product is carrier-added, carries a lower specific activity and contains some long-lived Lu-177m. The indirect route irradiates enriched Yb-176, which captures a neutron to form Yb-177, which decays to Lu-177. Because ytterbium and lutetium are different elements, the lutetium can be chemically separated from the bulk target, which is what makes no-carrier-added, high-specific-activity material possible.
Why does Yb-176 supply matter for lutetium-177?
Enriched Yb-176 is the target material for the indirect route that produces no-carrier-added Lu-177. Published thermal neutron-capture values sit at roughly 2.5 to 2.85 barns for Yb-176 against roughly 2,065 to 2,090 barns for Lu-176, so the indirect route converts a very small fraction of a very expensive target. That is why recovery and recycling of the unconverted ytterbium sit at the commercial centre of the process. A gram of Yb-176 is not an input consumed into doses. It is working capital that goes round a loop.
How does the 6.64-day half-life affect Lu-177 supply contracts?
It converts every logistics decision into a manufacturing decision. Activity decays while a batch is tested, while it waits for release, while it is packaged, while it clears customs, over a weekend, and while a patient appointment slips. Roughly 27 per cent of the activity is gone at three days and more than half at seven. A capacity or price figure quoted without a stated calibration time is not comparable to any other figure, so calibration time and the point at which price is struck, at calibration, at shipment or at receipt, belong in the contract.
What should a sponsor ask a lutetium-177 supplier before signing?
Twelve questions cover most of the exposure. On the material: chemical and physical form and whether it is qualified in your process, what was shipped and accepted rather than produced, how much is uncommitted and for how long, and who owns the recovered ytterbium. On the route: which reactor positions and outage history, whose hot cell performs separation and what the queue is, whether GMP release is claimed at precursor or finished product, and the stated calibration time. On access: contractual priority during an allocation event, which nodes your backup shares with your primary, realistic time from selecting an alternative to first qualified receipt, and what regulatory cross-reference support survives a change of control.
Is there a lutetium-177 shortage in 2026?
The public evidence does not support a claim of a current universal shortage, and it does not support a claim that supply is comfortable either. Demand is growing clearly: Novartis reported Pluvicto sales of $651 million in Q2 2026, up 43 per cent in constant currencies, and the FDA approved Pluvicto in PSMA-positive metastatic hormone-sensitive prostate cancer on 31 July 2026. Downstream capacity is expanding on several independent clocks. The more defensible reading is that the binding constraint is episodic and sponsor-specific rather than universal, and that the scarce commodity is not capacity but merchant, uncommitted, qualified access to it.
Work with ProGen
We build route-level supply chain intelligence for radiopharmaceutical developers, isotope producers, CDMOs and investors. Not a press release index. Entity and route records carrying operating status, access status, qualification, ownership, common-mode dependencies, evidence class and next verification date. And we run retained search for the roles that decide whether capacity becomes output: Quality, MSAT, Tech Ops, CMC, manufacturing and scale-up. It is the same map. The routes that get qualified are the routes with the teams to qualify them.
Sources
Primary sources, accessed August 2026. Where a primary link exists it is given in preference to any aggregator.
- TMC Group, "New Capacity to Support Next-Generation Radiotheranostics", 21 May 2026
- TMC Group, "Canadian-made Stable Isotopes: TMC Group Attends EANM 2025", 2 October 2025
- Kinectrics, "Kinectrics Significantly Increases Isotope Production", 17 September 2025
- Kinectrics, "Start of Yb-176 Production", 30 May 2024
- ASP Isotopes, Letter to Shareholders, 4 August 2026
- IAEA, "Production and Quality Control of Carrier-Free and No-Carrier-Added Lutetium-177"
- Tarasov et al., "Production of No-Carrier-Added Lutetium-177 by Irradiation of Enriched Ytterbium-176", Current Radiopharmaceuticals, 2015
- Institut Laue-Langevin, "ITM and ILL Extend Collaboration on Lutetium-177", 16 June 2025
- Bruce Power, "On-Site Hot Cell Facility", 14 July 2026
- University of Missouri Research Reactor, "MURR Makes n.c.a. Lu-177 Available Worldwide", 9 March 2026
- SHINE Technologies, Cassiopeia Lu-177 milestone and product specifications
- European Medicines Agency, Ilumira EPAR, EU marketing authorisation 26 March 2026
- European Medicines Agency, EndolucinBeta EPAR
- Novartis, Q2 2026 Results, 21 July 2026
- FDA Prescribing Information, Pluvicto and Lutathera, via Drugs@FDA
- BWXT, Nordic Capital transaction announcement, 3 August 2026, and Q2 2026 Form 10-Q
- US National Nuclear Data Center, Lu-177 evaluated nuclear data (half-life)
- NRG PALLAS and Global Morpho Pharma partnership announcement, 7 April 2026
- Laurentis Energy Partners, Darlington isotope production announcement
This article is independent market intelligence and not investment, legal or regulatory advice. Company names and trademarks are the property of their respective owners. © 2026 ProGen Search Limited.