The $945 million agreement by GE HealthCare to acquire SOFIE Biosciences is easy to read as another large company buying its way deeper into radiopharmaceuticals. But GE's own explanation of the transaction points to something more specific. The company repeatedly emphasizes SOFIE's position in the time-critical “final mile” of PET radiopharmaceutical supply, including a network of 15 manufacturing sites operating 21 cyclotrons across the United States. That geographic footprint is not incidental to the acquisition. It is a substantial part of what GE is paying for.
We saw another version of the same economics in federal procurement documents surrounding the Department of Veterans Affairs' nuclear medicine programs. In seeking Illuccix and Netspot for facilities in Pittsburgh and Lebanon, Pennsylvania, the VA specified that the distributor had to be positioned within the delivery window imposed by gallium-68's roughly 68-minute half-life. Cardinal Health was identified as the supplier capable of meeting those geographic and regulatory requirements. The government wasn't simply buying radiopharmaceuticals; it was buying radiopharmaceuticals that could reliably exist at the right place at the right time.
Those two developments look very different on the surface, with one involving a nearly $1 billion corporate acquisition and the other a federal purchasing requirement. But both reveal the same underlying feature of the nuclear medicine economy: unlike most pharmaceutical markets, distance can determine whether a product is commercially usable at all. That makes geography more than a logistical consideration. In many cases, it becomes part of the commercial model.
Traditional pharmaceuticals can be manufactured at enormous centralized plants, moved into warehouses and distributed through supply chains that operate over days or weeks. Radiopharmaceuticals impose another clock on that system because every minute between production and administration can represent declining activity, reduced flexibility and, depending on the isotope, a smaller economically viable delivery radius. That changes what infrastructure means and creates a business in which location can directly influence whether capacity has value.
A cyclotron located near several large hospital systems is not economically equivalent to the same cyclotron positioned far from clinical demand. A nuclear pharmacy inside the usable radius of a major metropolitan market can have strategic value that cannot necessarily be reproduced by adding more production capacity somewhere else. Manufacturing capacity matters, but capacity without geographic access to patients can become stranded capacity, particularly when the isotope involved allows little margin for delay.
Gallium-68 provides an obvious example because of its approximately 68-minute half-life, but the principle extends beyond one isotope. F-18 provides more geographic flexibility with its roughly 110-minute half-life, yet commercial distribution still depends heavily on where the radiopharmaceutical is manufactured relative to the imaging center. Therapeutic radiopharmaceuticals can provide substantially longer delivery windows, but they introduce their own requirements around manufacturing schedules, dose calibration, transportation, treatment appointments and radioactive-material handling. In each case, the supply chain is not separate from the product; it helps determine whether the product can actually reach the patient.
The result is a market in which geography becomes part of product strategy. Drug developers cannot think only about whether a radiopharmaceutical can be manufactured at scale. They increasingly have to ask whether it can be manufactured at scale in the places required to reach the patients who are supposed to receive it, and whether the supporting network can do that consistently enough to sustain commercial growth.
That helps explain why the SOFIE acquisition deserves to be viewed differently from the purchase of another manufacturing plant. GE HealthCare is not acquiring one cyclotron or one radiopharmacy. It is buying a distributed operating network with 15 U.S. manufacturing locations, 21 cyclotrons, an established customer base and a theranostics-focused CDMO operation. GE says that network will allow it to participate across more of the radiopharmaceutical value chain while strengthening the reliability of supply.
The word “network” is doing a lot of work there because 21 cyclotrons concentrated in one location would create substantial production capacity, but they would not create the same commercial capability as cyclotrons distributed across multiple markets. PET radiopharmaceuticals force manufacturing infrastructure to follow population centers, hospitals and imaging demand in a way that few other pharmaceutical products require. The commercial value does not come only from what a site can make, but from where that site is located and how many patients it can realistically reach.
That also means the value of an established network is not simply the replacement cost of its equipment. A functioning radiopharmaceutical site includes licenses, quality systems, trained personnel, validated production processes, delivery routes, hospital relationships and the operational experience required to repeatedly manufacture and release a product on a clock. Recreating that footprint requires more than capital, and building it one market at a time can take years.
GE is effectively buying years of geographic positioning along with the physical assets. The company already participates elsewhere in the nuclear medicine pathway through imaging systems, cyclotrons, pharmaceutical diagnostics, software and other infrastructure. Adding SOFIE moves it closer to the point where radiopharmaceuticals are physically produced and transferred to the patient, which helps explain why GE describes the acquisition in terms of the final mile rather than simply additional manufacturing capacity.
This has implications well beyond GE. Radiopharmaceutical companies deciding where to build plants, nuclear pharmacies deciding where to expand and CDMOs deciding where to add capacity may increasingly need to think less like conventional pharmaceutical manufacturers and more like network planners. Where are the treatment centers? Where are new PET programs opening? Where will radioligand therapy volumes grow? Which markets are already well served, and where are hospitals sitting outside efficient manufacturing and distribution corridors?
Those questions could become particularly important as the number of radiopharmaceuticals increases. A regional production network that today distributes FDG and a handful of specialty PET tracers may eventually support a much wider portfolio of oncology, neurology and cardiovascular products. Once the infrastructure is established, every additional radiopharmaceutical moving through that network can potentially increase the economic value of the footprint because the same licenses, personnel, equipment and distribution routes can support a broader commercial base.
The reverse is also true. A promising tracer can face an entirely different commercial trajectory if there are not enough sites capable of making it and delivering it reliably. Approval does not automatically create access in a business where the product begins disappearing almost as soon as it is manufactured. That means development teams may eventually need to think about geographic manufacturing strategy much earlier, particularly for products intended to scale nationally.
The VA procurement language makes that reality unusually explicit. Its requirement for Ga-68 products did not stop with FDA approval, product specifications or pricing. Geography itself became part of the qualification for supplying the contract because the laws of radioactive decay effectively created a boundary around which distributors could realistically serve the facilities. What looks like a procurement detail is actually a compact description of how the radiopharmaceutical market works.
For years, radiopharmaceutical competitive advantage has largely been discussed in terms of isotope supply, target biology, intellectual property and manufacturing scale. All remain critical, but as more products move from development into routine clinical use, the industry is beginning to discover that distribution infrastructure can create another form of defensibility. A competitor can license another target, secure another isotope source or build another manufacturing facility, but replicating a mature network positioned near major clinical markets is harder because the value comes from the combination of geography, regulation, operations and existing customer relationships.
That is why the GE-SOFIE transaction matters beyond its $945 million price tag. GE is acquiring radiopharmaceutical assets, but it is also acquiring locations on a map. Those locations shorten the distance between production and administration, create access to hospital markets and give GE infrastructure through which future products, both its own and those of other companies, can move. In a sector where time and distance can determine whether a dose reaches a patient at all, that kind of footprint can become strategically difficult to replicate.
As nuclear medicine scales, we should probably start looking at industry maps differently. They are no longer just illustrations showing where cyclotrons, pharmacies, manufacturing plants and treatment centers happen to be located. Increasingly, those maps show where commercial advantage exists, where supply gaps remain and which companies are positioned to connect production with clinical demand.
In most pharmaceutical markets, geography is largely a logistics problem to be optimized after the drug is developed. In nuclear medicine, radioactive decay makes geography part of the product itself, which means proximity can become a competitive advantage long before anyone talks about market share. The companies that understand that distinction may discover that one of the industry's strongest moats is not something they can patent, but something they can build only by being in the right places.
