Bayh–Dole Across Borders: Government-funded inventions in the United States, EU/Horizon Europe, Germany, Japan, China, and South Korea—organized by coverage, allocation, and continuing

Executive Summary: Bayh–Dole is best understood not as an automatic university-ownership rule, but as a conditional government–contractor bargain that permits a contractor to retain title to qualifying federally funded inventions while preserving disclosure and patenting duties, utilization reporting, a paid-up government license, domestic-manufacturing safeguards, nonprofit-specific restrictions, and discretionary march-in authority tied to practical application. A comparative analysis of the United States, EU/Horizon Europe and selected national-law examples, Japan, China, and South Korea demonstrates that institutional ownership alone is a poor measure of commercialization freedom: each regime differs in its funding trigger, method of allocating rights, exploitation requirements, government intervention powers, transfer controls, and inventor-remuneration rules. Japan provides the closest structural analogue among the systems discussed at the level of the covered government–contractor bargain, although Article 17 applies more narrowly than Bayh–Dole. The analysis also explains that a Bayh–Dole patent may, in principle, be sold to a nonpracticing or patent-assertion entity, but nonprofit sellers generally require agency approval unless the purchaser satisfies the invention-management exception, and a transfer does not eliminate the government license, march-in exposure, reporting obligations, domestic-manufacturing conditions, or other award-specific constraints. These distinctions provide a practical framework for cross-border prosecution, licensing, acquisition, financing, and enforcement diligence.

I.                   Introduction

For foreign patent attorneys, the Bayh–Dole Act is often summarized as a rule that gives U.S. universities ownership of federally funded inventions. That summary is incomplete. Bayh–Dole does not itself transfer an inventor’s title, guarantee a commercial launch, or impose a uniform sales deadline. It permits a qualifying contractor that has acquired a “subject invention” to elect to retain title, subject to disclosure and patenting duties, agency-requested utilization reporting, a standing government license, domestic-manufacturing safeguards, nonprofit assignment and income restrictions, and discretionary march-in authority. Its operative utilization concept is “practical application”: actual use under conditions that make the invention’s benefits available to the public on reasonable terms.

For U.S. patent attorneys, the reverse mistake is to ask whether another jurisdiction “has Bayh–Dole” and then infer the familiar U.S. package from institutional ownership. Japan’s Article 17 supplies a close structural comparison only for specified national-government-entrusted R&D and contracted software. China and South Korea use broader state safeguards, utilization measures, foreign-disposition controls, or revenue-sharing rules. EU/Horizon Europe and German national law illustrate a layered approach: Horizon governs funded beneficiaries and contractual “results,” while inventor entitlement and employee-invention rights remain national.

For both audiences, three questions provide a common vocabulary: how institutional control is allocated; what funding or legal trigger brings an invention or result within the regime; and which public constraints continue after private control is established. Those constraints include utilization or exploitation standards, monitoring, government-use or intervention rights, inventor remuneration, and transfer restrictions. Applied to cross-border prosecution counseling, licensing, acquisition, financing, and enforcement, that framework shows why record ownership is a poor proxy for commercialization authority.

II.                Translating Publicly Funded IP Across Borders

Cross-border patent work creates a recurring translation problem. A foreign attorney may encounter a U.S. patent carrying a federal-funding statement and need to determine whether the government owns the patent, whether the public may use it, whether the owner was required to commercialize it, and whether it can be sold to an assertion-focused purchaser. A U.S. attorney may review a foreign university or research-institution portfolio and assume that institutional ownership carries the same rights and limitations as Bayh–Dole. Neither assumption is safe.

Publicly funded research presents a common policy problem. If the state owns every resulting patent and licenses cautiously, promising inventions may remain unused. If private parties receive unqualified exclusivity, the public may finance research without receiving timely products, access, or other public benefit. Bayh–Dole’s response was not simply to “privatize” government-funded inventions. It coupled qualifying contractor control with disclosure, monitoring, retained public rights, and contingent remedies intended to encourage practical application without guaranteeing commercial success.

That model has influenced policy discussions well beyond the United States. The phrase “Bayh–Dole model,” however, is frequently used too loosely. A regime may promote university patenting without giving the government a paid-up license. It may allow institutional ownership while reserving far broader state intervention than U.S. march-in. Or it may reach a similar commercial result through employee-invention law and grant conditions rather than through a public-funding patent statute.

Publicly funded IP should be classified along three dimensions: (1) its allocation mechanism—who may control the result and how that party obtains title; (2) its scope or trigger—which awards, contracts, beneficiaries, activities, and outputs the regime reaches; and (3) its continuing public constraints—including disclosure, exploitation or utilization standards, performance monitoring, government-use rights, nonuse remedies, remuneration, and transfer rules. The economically meaningful entitlement is therefore better understood as conditional commercialization authority than as record ownership alone. Within the systems discussed, Japan is the closest structural analogue at the level of the government–contractor bargain, but that statement is not an overall similarity ranking: Article 17’s scope and continuing constraints differ materially from the U.S. regime.

Foreign counsel first need a Bayh–Dole primer, including the distinction among inventor title, contractor ownership, election to retain title, practical application, and march-in. U.S. counsel can use the same questions to evaluate EU/Horizon Europe, a German national-law illustration, Japan, China, and South Korea. A common vocabulary then supports cross-border licensing, acquisition, financing, and enforcement diligence.

The selected jurisdictions are major innovation systems and illustrate distinct legal architectures: a U.S. federal patent-rights regime, an EU program layered over national entitlement law, Japanese government nonacquisition subject to commitments, Chinese project-entity control with strong state safeguards, and Korean institutional succession supplemented by utilization and dissemination rules. The focus is formal civilian legal architecture, not relative commercialization rates or the frequency with which every public power is exercised. The comparison also does not claim that Bayh–Dole alone caused the growth of U.S. university patenting or that contractor ownership is normatively superior. Those propositions remain contested. (See David C. Mowery et al., The Growth of Patenting and Licensing by U.S. Universities: An Assessment of the Effects of the Bayh–Dole Act of 1980, 30 Research Policy 99, 99–119 (2001); Bhaven N. Sampat, Patenting and U.S. Academic Research in the 20th Century: The World Before and After Bayh–Dole, 35 Research Policy 772, 772–89 (2006); Rebecca S. Eisenberg, Public Research and Private Development: Patents and Technology Transfer in Government-Sponsored Research, 82 Va. L. Rev. 1663, 1663–1727 (1996).)

Primary statutes, regulations, and current grant terms supply the principal authorities. “Operational reach” refers to legal coverage unless administrative practice is specifically discussed. Defense, classified research, public-health emergency powers, competition law, foreign-investment review, export controls, state-aid rules, and sector-specific funding conditions may alter the analysis. English translations of Japanese, Chinese, and Korean authorities are reference materials; the controlling local-language text and executed funding documents should be confirmed for a legal opinion.

III.             Bayh–Dole Explained for Foreign Counsel: The U.S. Benchmark

For foreign counsel, Bayh–Dole is best understood as a sequence rather than as a rule that federal funding automatically places title in a university or company. Counsel should ask: Was there a covered federal funding agreement? Did the invention arise in performance of the funded work? Did the contractor acquire the inventor’s rights? Did it timely elect to retain title? May the patent be assigned, and to whom? What federal rights and continuing obligations remain after a sale or exclusive license?

A.                 Coverage, Title, and Election

Bayh–Dole’s policy section begins broadly by promoting utilization of inventions arising from federally supported research or development. A separate clause promotes commercialization and public availability of inventions made in the United States by United States industry and labor. Neither clause promises that every subject invention will become a product. Section 202 permits nonprofit organizations and small business firms, subject to exceptions and compliance conditions, to elect to retain title to a “subject invention,” and the standard federal patent-rights clause extends the operative framework to other business firms. The system encourages utilization through monitoring and contingent remedies rather than a universal sales quota or fixed market-entry deadline. (See 35 U.S.C. §§ 200–202; 37 C.F.R. §§ 401.2(b), 401.14(a)(8), (b) (2025).)

The statutory vocabulary supplies the first three steps. A “funding agreement” is a federal contract, grant, or cooperative agreement for experimental, developmental, or research work funded at least in part by the Federal Government. A “contractor” is a party to that agreement. A “subject invention” is an invention “of the contractor” that was conceived or first actually reduced to practice in performance of the funded work. Federal support, timing, or use of federally purchased equipment alone does not replace the required connection to the work undertaken under the agreement. (See 35 U.S.C. § 201(b), (c), (e); 37 C.F.R. § 401.1.)

The familiar statement that “Bayh–Dole gives the university title” is therefore imprecise. In Board of Trustees of the Leland Stanford Junior University v. Roche Molecular Systems, Inc., the Supreme Court held that the Act does not automatically vest an employee-inventor’s rights in the contractor. Bayh–Dole operates after the contractor has acquired the invention under ordinary assignment principles. “Election to retain title” preserves qualifying contractor-owned title against the Government; it does not create title that the contractor never obtained from the inventor. (See 35 U.S.C. §§ 201(e), 202(a); 563 U.S. 776, 786–90 (2011); 37 C.F.R. § 401.14(f)(2) (2025).) Defective assignments, conflicting consulting agreements, and gaps in institutional policy can therefore prevent an invention from entering the contractor’s Bayh–Dole portfolio.

Once the contractor owns a subject invention, the familiar compliance architecture applies. The contractor must disclose the invention, elect whether to retain title, and file patent applications within the applicable periods, or risk the agency taking title. The funding agreement must permit periodic reporting on utilization or efforts to obtain utilization by the contractor, assignees, and licensees. Under the standard clause, the contractor submits those reports on agency request, no more frequently than annually, with information that may include development status, the date of first commercial sale or use, gross royalties received, and other reasonably specified information. Nonprofit contractors must also share royalties with inventors, make reasonable efforts to attract qualifying small-business licensees and prefer them when appropriate, and generally devote remaining net income to scientific research or education. (See 35 U.S.C. § 202(a), (c)(1)–(5), (7); 37 C.F.R. § 401.14(c)–(h), (k) (2025).)

Foreign counsel often first encounters this framework through the government-interest statement in a U.S. patent. The statement gives notice that the invention was made with federal support and that the Government has specified rights. It does not mean that the Government owns the patent, that the public may practice the invention, that the contractor properly acquired title, or that every compliance issue has been resolved. (See 35 U.S.C. § 202(c)(6); 37 C.F.R. § 401.14(f)(4) (2025).)

B.                 Commercialization as Practical Application

The operative utilization benchmark is “practical application,” not commercialization in the abstract. Section 201(f) defines practical application as manufacture of a product or composition, practice of a process or method, or operation of a machine or system under conditions establishing both use and—within applicable law and regulation—public availability of the invention’s benefits on reasonable terms. A commercial sale may supply evidence of practical application, but it is not the exclusive statutory route. The statute does not require the contractor itself to manufacture or sell, and patenting or signing a license does not mechanically establish that effective development steps exist. (See 35 U.S.C. §§ 200, 201(f); 37 C.F.R. § 401.14(a)(3), (h) (2025).)

Section 203(a)(1) does not demand immediate achievement of practical application. An agency may march in only if it determines that licensing action is necessary because the contractor or assignee has not taken, or is not expected to take within a reasonable time, effective steps to achieve practical application in a particular field of use. The standard is contextual and field-specific. The remedy is a required nonexclusive, partially exclusive, or exclusive license to a responsible applicant—or an agency-granted license if the relevant private party refuses—after the determination and review process prescribed by statute and regulation. March-in is neither self-executing nor automatic divestiture of title. (See 35 U.S.C. § 203(a)(1), (b); 37 C.F.R. §§ 401.6, 401.14(j)(1) (2025).)

The more express development commitment in Chapter 18 applies in a different setting. Under § 209(f), an applicant for a license under a federally owned patent or patent application must submit a development or marketing plan. An applicant for an exclusive or partially exclusive license must also commit to achieve practical application within a reasonable time. The license must permit termination, in whole or in part, if the licensee is not executing its submitted commitment and cannot otherwise demonstrate effective steps toward practical application within a reasonable time. Those requirements for licensing federally owned inventions should not be imported into the ordinary contractor-retained-title regime under §§ 202–204. (See 35 U.S.C. § 209(a)(3), (d)(3)(A), (f).)

Formal authority and administrative practice must also be separated. In February 2026, the Government Accountability Office reported that federal agencies had never exercised Bayh–Dole march-in authority and had declined approximately a dozen requests. GAO also reported that, as of December 2025, NIST had not finalized its 2023 draft framework proposing that product price could inform analysis under the practical-application and health-or-safety criteria. NIST’s public materials continued to identify the framework as a draft under review on July 17, 2026. The statutory power is therefore real, but its existence should not be confused with an accrued title defect or a routine licensing event; price-based march-in remains a policy proposal rather than an operative rule. (U.S. Gov’t Accountability Off., GAO-26-107885, Intellectual Property: Information on Draft Guidance to Assert Government Rights Based on Price 1–2, 14–22 (2026); Nat’l Inst. of Standards & Tech., Bayh–Dole Regulations for Federally Funded Inventions (last visited July 17, 2026).)

Private title also carries standing public safeguards. The United States receives a nonexclusive, nontransferable, irrevocable, paid-up worldwide license to practice, or have practiced, the invention for or on behalf of the Government. Beyond inadequate effective steps toward practical application, § 203 permits march-in for unmet health or safety needs, unsatisfied federal public-use requirements, or the absence or breach of the domestic-manufacturing agreement addressed by § 204. Section 204, subject to waiver, conditions specified exclusive U.S. rights on substantial domestic manufacture; it does not require the contractor to manufacture or bring a product to market. (See 35 U.S.C. §§ 202(c)(4), 203(a)(2)–(4), 204; 37 C.F.R. § 401.14(i)–(j) (2025).)

C.                 Assignment, Sale, and the NPE Question

A subject-invention patent may be transferred, but it cannot be sold free of the federal bargain that permitted private retention of title. Patents ordinarily are assignable by a written instrument. A sale changes the owner; it does not change the invention’s status as a subject invention or extinguish federal rights. The downstream overlay takes three forms: the Government’s existing license is a retained property interest; §§ 203 and 204 expressly reach assignees or exclusive licensees in specified circumstances; and reporting, patent-management, and approval duties arise through the funding agreement and standard patent-rights clause. Those distinct legal forms should not be collapsed into a generic statement that every Bayh–Dole duty “runs with the patent” in exactly the same way. (See 35 U.S.C. §§ 202(c)(4)–(8), 203–04, 261; 37 C.F.R. § 401.14(b), (h)–(k) (2025).)

A nonprofit contractor faces a purchaser restriction. A nonprofit may not assign U.S. rights in a subject invention without approval of the funding agency, except to an organization having as one of its primary functions the management of inventions. An assignee using that exception must be subject to the same provisions as the nonprofit contractor, including the applicable inventor-sharing, income-use, and small-business-licensing provisions. A buyer that does not satisfy the invention-management exception—whether an operating company, investment vehicle, foreign purchaser, or assertion-focused entity—requires agency approval. The statute and standard clause do not define “management of inventions” or create a safe harbor for any business-model label. NIST’s iEdison process accordingly provides an assignment-request workflow in which the recipient accepts ownership and associated reporting responsibility and the primary funding agency approves or denies the request. For an agency-approved purchaser outside the statutory exception, counsel should examine the approval, award, and assignment documents rather than assume that every nonprofit-specific economic provision passes through automatically. (See 35 U.S.C. § 202(c)(7)(A)–(D); 37 C.F.R. § 401.14(k) (2025); Nat’l Inst. of Standards & Tech., Receiving an Invention Assignment Request (updated July 1, 2022), https://www.nist.gov/iedison/iedison-organization-user-guide/invention-reports/receiving-invention-assignment-request.)

The nonprofit seller’s proceeds remain independently important. Section 202(c)(7)(C) reaches the balance of “royalties or income earned” with respect to subject inventions and generally requires its use for scientific research or education after permitted expenses. Assignment consideration therefore should not be treated as unrestricted proceeds without analysis. Whether a lump-sum sale payment also triggers a particular inventor share may depend on the governing institutional policy, employment arrangement, and transaction documents because the separate statutory inventor-sharing clause speaks in terms of royalties. (See 35 U.S.C. § 202(c)(7)(B)–(C); 37 C.F.R. § 401.14(k)(2)–(3) (2025).)

A for-profit contractor follows a different default rule. The nonprofit assignment restriction in § 202(c)(7)(A) and paragraph (k)(1) has no direct counterpart for a small-business or other for-profit contractor. Bayh–Dole therefore does not categorically disqualify an operating company, affiliate, investment fund, foreign buyer, licensing company, or nonpracticing entity from purchasing a for-profit contractor’s subject-invention patent. The actual funding agreement, agency supplements, exceptional-circumstances provisions, security restrictions, export controls, foreign-investment rules, and other applicable law may nevertheless require notice, consent, or a different allocation. A transaction that also assigns or substitutes a party under the funding agreement requires separate analysis; buying the patent is not necessarily the same as becoming a party to the funded work. (See 35 U.S.C. §§ 201(b)–(c), 202(c)(7); 37 C.F.R. §§ 401.2(a)–(b), 401.3, 401.14(k) (2025).)

An NPE is not a statutory category. The term can describe a university or technology-transfer organization that licenses rather than manufactures, as well as a patent assertion entity that acquires rights principally for licensing or enforcement. Bayh–Dole does not categorically prohibit either model. For a nonprofit seller, however, an assertion-focused buyer should not assume that acquiring and enforcing patents automatically satisfies the invention-management exception; written agency approval or confirmation is the safer course. For a for-profit seller, no comparable blanket approval rule appears in the standard clause. The NPE label therefore answers neither who may buy nor whether the invention is being utilized.

NPE ownership also does not itself establish or defeat practical application. The patent owner need not manufacture: operating licensees may manufacture, practice, or operate the invention and make its benefits publicly available. Conversely, purchasing a patent, offering licenses, or filing infringement suits does not by itself establish the statutory manufacture, practice, operation, and public availability. A portfolio strategy that leaves a field unused without effective development steps may create a weaker record under § 203(a)(1), but march-in still requires a field-specific agency determination that action is necessary. The purchaser’s identity, the filing of a lawsuit, and the amount of a royalty demand are not independent march-in grounds. (See 35 U.S.C. §§ 201(f), 203(a); 37 C.F.R. § 401.14(j) (2025).)

An assignee may enforce the patent against private infringers because Bayh–Dole does not create a general public license. Ordinary noncompliance likewise does not automatically void title or allow an accused private infringer to exercise the Government’s remedies. Practice by or for the United States within the paid-up license presents a different enforcement market: the assignee cannot exclude conduct that the United States already has authority to practice or have practiced for or on its behalf. (See 35 U.S.C. § 202(c)(4); Central Admixture Pharmacy Services, Inc. v. Advanced Cardiac Solutions, P.C., 482 F.3d 1347, 1352–53 (Fed. Cir. 2007); Campbell Plastics Engineering & Manufacturing, Inc. v. Brownlee, 389 F.3d 1243, 1249–50 (Fed. Cir. 2004).)

Calling the transaction an exclusive license may not avoid the assignment question. Ordinary patent law examines the substance of the rights transferred rather than the agreement’s label; a conveyance of all substantial patent rights can operate as an assignment. No reported decision squarely applies that test to § 202(c)(7)(A), but a nonprofit should not assume that retaining nominal title while transferring enforcement control, sublicensing authority, and substantially all commercial rights eliminates the need for agency review. In any event, § 203 expressly reaches an exclusive licensee, utilization reporting addresses licensee activity, and § 204 governs a downstream grant of an exclusive U.S. right to use or sell. (See Waterman v. Mackenzie, 138 U.S. 252, 255–56 (1891); Lone Star Silicon Innovations LLC v. Nanya Technology Corp., 925 F.3d 1225, 1229–30 (Fed. Cir. 2019); 35 U.S.C. §§ 202(c)(5), 203(a), 204.)

The buyer’s later licensing and patent-management choices therefore matter. Section 204 does not state that an outright purchaser must manufacture; it requires the contractor or assignee to obtain a substantial-U.S.-manufacture agreement before granting another person the exclusive U.S. right to use or sell, unless the agency grants a waiver. The standard clause also requires advance notice of specified abandonment, maintenance, reexamination, opposition, Patent Trial and Appeal Board, and post-grant decisions, and it may permit the agency to request title if patent protection is not maintained or defended. Assignment documents should condition closing on any required approval, preserve the Government’s license, allocate iEdison and utilization reporting, require access to development data, address § 204 in downstream exclusive licenses, allocate nonprofit proceeds and inventor payments, and require cooperation before patent abandonment or post-grant action. (See 35 U.S.C. § 204; 37 C.F.R. § 401.14(d)(1)(iii), (f)(3), (h)–(i) (2025).)

For foreign counsel. The useful U.S. comparison point is therefore a package: contractor-acquired title and election; disclosure and patenting duties; utilization monitoring; a standing government license; a practical-application standard backed by discretionary march-in; downstream transfer and licensing constraints; and domestic-industry and nonprofit safeguards. Bayh–Dole creates commercialization incentives, transparency, and contingent public leverage, not a guarantee that every subject invention will reach market. A patent may be sold, including in principle to an NPE, but purchaser eligibility, approval, continuing federal rights, and post-closing use depend on the seller’s status and the governing award. The remaining systems should be compared against that package, not against ownership—or the word “commercialization”—alone.

IV.            EU/Horizon Europe and Germany: A Layered System, Not a European Bayh–Dole Act

For U.S. counsel, the first point is institutional: “Europe” is not a substantive patent-law jurisdiction with one publicly funded-invention statute. The European Patent Organisation is not the European Union, and the European Patent Convention is not an EU research-funding code. EPC Article 60 provides that the right to a European patent belongs to the inventor or successor in title. For employee inventions, entitlement is determined by the law of the state in which the employee is mainly employed, with a fallback to the employer’s relevant place of business. Article 60(3) merely directs the EPO to deem the applicant entitled in EPO proceedings. It creates neither institutional ownership of publicly funded inventions nor a government license or march-in right. (Convention on the Grant of European Patents (European Patent Convention) art. 60(1), (3), Oct. 5, 1973, 1065 U.N.T.S. 199, as amended by Act Revising the Convention on the Grant of European Patents, Nov. 29, 2000 (entered into force Dec. 13, 2007).)

A.                 Horizon Europe: The Program Layer

At EU level, Horizon Europe supplies the most relevant program-specific comparison. In Horizon terminology, an “action” is the funded project, a “beneficiary” is a party to the grant, and “results” are outputs generated in the action—a category broader than patentable inventions and capable of including data, know-how, software, and other material. Regulation (EU) 2021/695 provides that beneficiaries own the results they generate. Beneficiaries must align employee and third-party rights with the grant. Jointly generated and inseparable results may be jointly owned. Beneficiaries must adequately protect results where protection is possible and justified, and use best efforts to exploit them. Best efforts is a conduct obligation, not a guarantee that every result will become a commercial product. These rules govern funded beneficiaries and contractual results; they do not displace the national law that determines inventor entitlement. (Regulation (EU) 2021/695 of the European Parliament and of the Council of 28 April 2021 Establishing Horizon Europe—The Framework Programme for Research and Innovation arts. 2(21), 38–39, 2021 O.J. (L 170) 1, 15, 37–39; European Patent Convention art. 60(1).)

Horizon’s public rights are narrower and more differentiated than Bayh–Dole’s. EU institutions and bodies receive royalty-free access for developing, implementing, and monitoring Union policies and programs, but ordinary access is limited to noncommercial and noncompetitive use. When the grant activates the right, the Commission or funding body may object to a transfer or exclusive license to an entity in a non-associated third country if the transaction is contrary to Union interests. That is a strategic transfer control, not a domestic-manufacturing preference. (Regulation (EU) 2021/695, arts. 40(4), 41(9), 2021 O.J. (L 170) 1, 39–40.)

The Horizon Europe General Model Grant Agreement, Multi & Mono, version 1.2 dated June 1, 2026, confirms that the granting authority does not obtain ownership. Annex 5 generally continues best-efforts exploitation duties for four years after the action and directs beneficiaries to place specified unexploited results on the Horizon Results Platform, the Commission’s mechanism for publicizing results and connecting them with potential users. Two stronger safeguards are call-conditional, meaning they operate only when the solicitation and grant activate them. A public-emergency clause may require time-limited, nonexclusive licenses on fair and reasonable terms to entities able to address the emergency and commit to rapid and broad exploitation. A call may also activate a four-year authority to object to specified transfers or exclusive licenses to non-EU entities in countries not associated with Horizon Europe. Neither mechanism is a universal statutory march-in right. (European Commission, Horizon Europe Programme & Euratom Research and Training Programme: General Model Grant Agreement—Multi & Mono, version 1.2, art. 16.2, at 47 & annex 5, at 97–100 (June 1, 2026).)

A narrower Horizon mechanism comes closer. In pre-commercial procurement, Article 26(3) leaves at least result-related IP with the generating contractor, grants the contracting authority royalty-free own-use access and an ability to grant or require nonexclusive third-party licenses on fair and reasonable terms, and permits transfer of ownership to the authority if the contractor fails to commercialize within the contract period. This is genuinely march-in-like, but it applies to a procurement mechanism rather than to ordinary research grants. (Regulation (EU) 2021/695 art. 26(3), 2021 O.J. (L 170) 1, 32.)

EU soft law supplies policy rather than title. The current Code of Practice on intellectual-asset management recommends, for predominantly publicly funded collaborations, ownership in the participating public research organization, preferential industry access commensurate with contribution, Union-benefiting valorization, and fair-and-reasonable emergency licensing where a project was specifically funded for the emergency. It also recommends contractual rights allowing a public-research joint owner to license third parties if an industrial joint owner does not exploit within an agreed period. These are influential drafting norms, not statutory entitlements. (Commission Recommendation (EU) 2023/499, § 3.1, paras. 33–41, 2023 O.J. (L 69) 75, 82–84; Council Recommendation (EU) 2022/2415, 2022 O.J. (L 317) 141.)

B.                 Germany: A National-Law Illustration

Germany illustrates one national-law layer; it does not stand in for a unitary European rule. The Employee Inventions Act allocates rights because of employment, not public sponsorship. A “service invention” is an employee invention connected to assigned duties or substantially based on the employer’s activity or experience. The employee must report it. The employer may “claim” the invention—a statutory transfer step, not merely the execution of a conventional U.S. assignment—and the claim is generally deemed made unless the employer releases the invention within four months after proper reporting. Economic rights then pass to the employer, subject to reasonable inventor remuneration. Section 42 modifies the regime for university inventions, preserving limited academic publication and research interests while permitting the university employer to claim the invention. If the employer exploits it, the university inventor receives 30 percent of gross exploitation revenue. (Gesetz über Arbeitnehmererfindungen [ArbnErfG] [Employee Inventions Act], §§ 4–7, 9, 42 (Ger.).)

German federal grant terms can add a separate public-interest layer. The NKBF 2017 conditions are program terms for specified research-ministry cost-basis grants to commercial enterprises; they are not the rules for all German federally supported R&D. When incorporated, the recipient generally owns and exclusively exploits its results. But certain exploitation outside the European Economic Area and Switzerland may require consent; the ministry may demand nonexclusive rights in the public interest or exclusive rights for public security under the stated conditions; and it may require exploitation within a reasonable period. Noncompliance may support revocation or repayment. Those are Bayh–Dole-like functions embedded in program terms, not a universal patent statute. (Bundesministerium für Bildung und Forschung, Nebenbestimmungen für Zuwendungen auf Kostenbasis des Bundesministeriums für Bildung und Forschung an gewerbliche Unternehmen für Forschungs- und Entwicklungsvorhaben (NKBF 2017), version 2017, cls. 3.1–3.8 (Ger.).)

For U.S. counsel. The analysis proceeds in layers. National law determines inventor-to-employer entitlement; the Horizon grant identifies the beneficiary that owns each project result; and the regulation, call, grant agreement, and consortium agreement add exploitation, access, and transfer conditions. For a German asset, employee-invention law supplies the inventor-to-employer title and compensation rules, while incorporated grant conditions may add a separate public-funding layer. No single instrument performs all of those functions, and the German illustration does not describe “Europe” as a whole.

V.               Japan: A Familiar Government–Contractor Bargain with a Narrower Trigger

Article 17 of Japan’s Industrial Technology Enhancement Act may look familiar to U.S. counsel because it permits a contractor to retain specified rights while preserving government-use and nonuse safeguards. The resemblance must not obscure the trigger. “Entrusted R&D” refers here to technology R&D that the national government commissions or entrusts to an outside performer. Article 17 does not apply to all Japanese government-supported research and does not give every funded contractor a right to elect title. It permits the national government to decide not to acquire specified IP arising from covered entrusted technology R&D or government-contracted software development, and only if the contractor accepts four statutory commitments. (Sangyō Gijutsuryoku Kyōka Hō [Industrial Technology Enhancement Act], Act No. 44 of 2000, art. 17(1) (Japan).)

First, the contractor must promptly report the resulting research or development outcome. Second, if the government determines that use is particularly necessary in the public interest and states its reasons, the contractor must give the government a royalty-free right to use the IP. Third, if the contractor has failed to use the right for a considerable period without legitimate reason, it must license a third party when the government determines—and explains—that the license is particularly necessary to promote utilization. Fourth, the contractor generally must obtain advance government approval before assigning the right or establishing or transferring specified exclusive rights, subject to exceptions for mergers, corporate divisions, and circumstances prescribed as unlikely to impede use. (Industrial Technology Enhancement Act art. 17(1)(i)–(iv), (2)–(3) (Japan).)

The third commitment is a genuine march-in-like mechanism. It is narrower than the complete set of U.S. statutory grounds because it centers on prolonged nonuse without legitimate justification. Conversely, Japan’s government-use right is more contingent than Bayh–Dole’s standard paid-up license: the government must identify a particular public-interest need and state its reasons. Article 17 contains no direct counterpart to Bayh–Dole’s U.S.-manufacturing preference, nonprofit assignment limitations, or statutory rule for sharing licensing income with inventors.

The covered rights extend beyond patents. The implementing order reaches patent and patent-application rights, utility-model, design, copyright, semiconductor-circuit-layout, and plant-variety rights, and it identifies specified exclusive rights for which establishment or transfer ordinarily requires approval. (Industrial Technology Enhancement Act Enforcement Order, Cabinet Order No. 206 of 2000, art. 2 (Japan).) Article 17 also includes a pass-through rule where a government-funded entity subcontracts the R&D or software work. It should therefore be understood as a conditional retention regime for entrusted and contracted R&D—not as an automatic title rule for every Japanese research grant.

Inventor compensation comes from Japan’s general employee-invention law. Patent Act Article 35 gives the employer a nonexclusive license in an employee invention and permits agreements or work rules to provide that the right to obtain a patent belongs to the employer from the invention’s creation. When an employer acquires the right or an exclusive license, the employee is entitled to a reasonable monetary or other economic benefit, evaluated in light of consultation, disclosure, the opportunity to express views, expected employer profit, and the parties’ respective contributions. (Tokkyo-hō [Patent Act], Act No. 121 of 1959, art. 35(1), (3)–(7) (Japan).) Japan thus couples Article 17 contractor retention with a separate employee-invention compensation framework rather than a Bayh–Dole-style percentage rule.

For U.S. counsel. Article 17 becomes Bayh–Dole-like only after its narrow contract and subject-matter trigger is satisfied. The contractor retains rights because the government agrees not to acquire them, not because the contractor makes a U.S.-style election covering all publicly supported research. Review the contract type first, then confirm the four commitments, any disposition approval, and the separate Patent Act Article 35 employee-invention arrangement. Japan is the closest structural analogue within the systems discussed at the covered government–contractor level, not the regime with the broadest operational reach.

VI.            China: Institutional Control with Broader State Safeguards

China begins from a premise familiar to U.S. counsel—the performing entity ordinarily controls publicly funded results—but applies it through a different vocabulary and stronger state safeguards. China does not have a single Bayh–Dole statute. It combines public-research rules, employee-invention law, technology-transfer legislation, and program requirements. In this context, “fiscal funds” are public-budget funds; a “science and technology plan project” is a government-organized or funded R&D project; and the “project undertaking entity” is the organization designated to perform or administer that project. Article 32 of the Law on Progress of Science and Technology authorizes that entity to obtain resulting IP, provided that doing so does not harm national security, national interests, or major social public interests. The entity may implement the result, transfer it, license it, commercialize it jointly, or contribute it as an equity investment. (Zhonghua Renmin Gongheguo Kexue Jishu Jinbu Fa [Law of the People’s Republic of China on Progress of Science and Technology] arts. 32, 34 (rev. Dec. 24, 2021, effective Jan. 1, 2022) (China).)

The statutory wording—authorizing the undertaking entity to “obtain” the IP—is not identical to a right to “elect to retain title.” The practical outcome, however, is ordinarily institutional ownership. Under China’s Patent Law, the right to apply for an employee invention made in performing the entity’s tasks or primarily using its material and technical resources generally belongs to the entity. The project entity may be a university, research institute, enterprise, or other qualified organization; the publicly funded results rule is not textually limited to nonprofits and small businesses. (Zhonghua Renmin Gongheguo Zhuanli Fa [Patent Law of the People’s Republic of China] arts. 6, 15 (amended Oct. 17, 2020, effective June 1, 2021) (China).)

Institutional owners receive substantial commercialization autonomy. In Chinese legislation, “transformation of scientific and technological achievements” refers broadly to putting R&D outputs into productive or commercial use through self-use, transfer, licensing, collaboration, or equity contribution. The Law on Promoting that transformation permits state-established universities and research institutions to decide independently whether to transfer, license, or contribute their achievements as investment, subject to market-pricing and disclosure procedures. Commercialization revenue generally remains with the institution. After required rewards and remuneration, the remaining income is used principally for R&D and further technology transfer. (Zhonghua Renmin Gongheguo Cujin Keji Chengguo Zhuanhua Fa [Law of the People’s Republic of China on Promoting the Transformation of Scientific and Technological Achievements] arts. 16–18, 43 (amended Aug. 29, 2015, effective Oct. 1, 2015) (China).)

China regulates researcher participation more prescriptively than Bayh–Dole. In the absence of a governing policy or agreement, the Transformation Law requires specified minimum rewards for important contributors: at least 50 percent of net transfer or licensing income; at least 50 percent of the equity or capital interest received when the result is invested; or, for self-implementation or joint implementation, at least 5 percent of annual operating profit for three to five consecutive years. For state-established universities and research institutions, internal policies or agreements may not reduce those floors. Separate Patent Law rules require rewards and reasonable remuneration for employee inventors. (Law on Promoting the Transformation of Scientific and Technological Achievements arts. 44–45 (China); State Council, Several Provisions for Implementing the Law on Promoting the Transformation of Scientific and Technological Achievements § 2(6), Guo Fa [2016] No. 16 (China); Regulations for the Implementation of the Patent Law of the People’s Republic of China arts. 92–94, State Council Order No. 769 (rev. Dec. 11, 2023, effective Jan. 20, 2024) (China).)

The state retains powerful nonuse and public-interest authority. The undertaking entity must protect and implement the IP and report annually to the project-management institution. If it fails, without legitimate reason, to implement within a reasonable period, the state may implement the right without compensation or authorize another party to implement it on a paid or unpaid basis. Independently of nonuse, the state may exercise similar powers for national security, national interests, or major social public interests. (Law on Progress of Science and Technology art. 32 (China).) These are China’s closest equivalents to march-in, although the purposes are framed more broadly and the statute does not reproduce Bayh–Dole’s detailed administrative procedure.

A 2024 patent-declaration plan supplies a concrete nonuse benchmark. A patent resulting from a fiscally funded research project that remains unimplemented without legitimate reason for more than five years after grant may be implemented by the state without compensation or licensed to another party on a paid or unpaid basis. The plan also requires applicants from specified centrally funded programs to declare the associated project after the patent application is accepted. The funding information is maintained for government management rather than printed as a public federal-rights legend. Failure to declare prevents the patent from counting as a project-completion result; the plan does not state that title automatically forfeits. (China National Intellectual Property Administration et al., Implementation Plan for Establishing a Declaration System for Patents Formed by Fiscally Funded Scientific Research Projects §§ 2, 4(3), Guo Zhi Fa Yun Zi [2024] No. 3 (Jan. 22, 2024) (China).)

China also exercises stronger control over foreign disposition. Article 34 encourages first use in China and requires project-management-institution approval before transfer of publicly funded project IP to a foreign organization or individual or an exclusive foreign license, unless another law identifies a different approval route. (Law on Progress of Science and Technology art. 34 (China).) This is more direct than Bayh–Dole’s domestic-manufacturing preference.

For U.S. counsel. Record title in the project entity does not exhaust the inquiry. The same asset may carry annual project-reporting obligations, nonuse and public-interest intervention exposure, approval requirements for a foreign transfer or exclusive foreign license, and mandatory researcher-reward rules. Article 34 encourages first use in China, while separately requiring approval for specified foreign dispositions; counsel should not collapse the encouragement and the approval requirement into the same legal obligation. China is therefore best understood as a Bayh–Dole-type system with materially broader public and inventor constraints, not an equivalent copy.

VII.         South Korea: Institutional Succession and Public-Technology Dissemination

South Korea shows why a U.S. lawyer should look beyond a single statute. Title succession, utilization, dissemination, commercialization returns, and public intervention are distributed across related regimes, principally the National Research and Development Innovation Act and the Technology Transfer and Commercialization Promotion Act, their enforcement decrees, and general employee-invention law.

Article 16 of the National R&D Innovation Act supplies the principal title rule. The “performing research institution” is the participating organization carrying out the national R&D project. In general, that institution succeeds to the researcher’s rights in the R&D outcome. “Succession” matters: rights move from the researcher to the institution through the statutory structure rather than arising originally in the institution and bypassing the researcher. Depending on the type of outcome and the nature and weight of participation, the researcher may own the result or multiple institutions may own it jointly. (Gukga Yeongu Gaebal Hyeoksinbeop [National Research and Development Innovation Act], Act No. 17343, June 9, 2020, art. 16(1)–(3) (S. Kor.) (version effective June 11, 2026).)

The government may require state ownership where necessary for national security, use in the public interest, participation by a foreign-located research institution, or another prescribed circumstance making institutional ownership inappropriate. Current rules ordinarily require that possibility to be disclosed at the solicitation stage. In other words, state ownership may be reserved when the project is solicited or awarded; this is principally an ex ante allocation power, not a later march-in remedy. (National Research and Development Innovation Act art. 16(3) (S. Kor.); Gukga Yeongu Gaebal Hyeoksinbeop Sihaengnyeong [Enforcement Decree of the National Research and Development Innovation Act], Presidential Decree No. 36291, May 6, 2026, art. 32 (effective June 11, 2026) (S. Kor.).)

Articles 17 and 18 address utilization and commercialization. An owning institution must maintain and manage outcomes, facilitate shared use, connect and disclose relevant information, and undertake additional R&D or other measures needed to promote broad utilization. The ministry may conduct follow-up investigations. The institution may license an outcome and ordinarily collects royalties; prescribed owners may owe a portion of commercialization returns to the funding authority, while remaining income supports participating-researcher compensation, technology transfer and IP expenses, reinvestment, and other authorized purposes. Act No. 21421 took effect in stages: amendments to Articles 17(1) and 18 became effective June 11, 2026, while later Article 17 amendments concerning deposit and disclosure take effect September 11, 2026. (National Research and Development Innovation Act arts. 17–18 (S. Kor.); Act No. 21421, Mar. 10, 2026, supp. arts. 1–2 (S. Kor.) (amendments to arts. 2, 17(1), 18, and 32 effective June 11, 2026; remaining amendments effective Sept. 11, 2026); Enforcement Decree of the National Research and Development Innovation Act arts. 34–41, Presidential Decree No. 36291, May 6, 2026, as effective June 11, 2026 (S. Kor.).)

The companion Technology Transfer and Commercialization Promotion Act applies more broadly to “public technology,” a statutory category that adds dissemination and licensing rules to specified technology produced or supported by public institutions or funds. It permits the state, local governments, and public institutions to vest publicly supported R&D outcomes in participating institutions subject to utilization conditions. Public research institutions generally must make vested technology available to enterprises on equal terms. Under the enforcement decree, nonexclusive licensing is the default; exclusivity is permitted when another law or agreement requires it, no party seeks a nonexclusive license during the prescribed period, or the technology’s nature makes exclusivity unavoidable. If a public research institution fails to take the required utilization measures, the state must take them directly. (Technology Transfer and Commercialization Promotion Act arts. 19, 24; Enforcement Decree of the Technology Transfer and Commercialization Promotion Act arts. 24, 26 (S. Kor.).)

That governmental step-in is Korea’s closest analogue to march-in, but it should not be overstated. It promotes dissemination; it does not recreate Bayh–Dole’s detailed petition, findings, and compelled-license structure. Nor do the National R&D Innovation Act provisions reserve a general royalty-free government license comparable to 35 U.S.C. § 202(c)(4). Separate Korean patent-law provisions authorize government use or compulsory licensing for specified emergencies, public-interest needs, and nonworking, but those powers arise independently of public funding. (Patent Act arts. 106-2, 107 (S. Kor.).)

Korea also places substantial emphasis on revenue sharing. The Technology Transfer Act requires public research institutions to distribute an appropriate share of transfer income to researchers and technology-transfer contributors; the decree fixes minimum shares for those groups. The National R&D regime separately specifies uses and allocations of royalty income, while general employee-invention law protects fair compensation. The exact calculation can therefore depend on the identity of the institution, project terms, funding source, and nature of the revenue. (Technology Transfer and Commercialization Promotion Act art. 19(2), (4); Enforcement Decree of the Technology Transfer and Commercialization Promotion Act art. 24; National Research and Development Innovation Act art. 18; Enforcement Decree of the National Research and Development Innovation Act art. 41 (S. Kor.).)

For U.S. counsel. Analyze two overlapping tracks. The National R&D regime addresses ownership, utilization, and commercialization returns; the public-technology regime adds dissemination and licensing rules. Neither alone is a Korean Bayh–Dole Act. Institutional control coexists with succession from researchers, possible state ownership specified in advance, affirmative utilization measures, detailed revenue allocation, and public-technology dissemination. The result is functionally comparable but structurally dispersed.

VIII.       A Comparative Guide for U.S. and Foreign Counsel

Table 1 states the common inquiry as three questions: When does the regime apply? How does the institution obtain or control rights? What government, public, or inventor constraints continue afterward?

Table 1. Scope, allocation, and continuing constraints across publicly funded IP regimes

System	When does it apply?	How does the institution obtain or control rights?	What obligations and public or inventor rights continue?
United States	Contractor-owned inventions that are or may be patentable, plus protectable plant varieties, conceived or first actually reduced to practice in performance of work under a federal funding agreement.	Contractor first acquires inventor title, then elects to retain a qualifying subject invention.	No fixed sales deadline; utilization reporting; paid-up government license; field-specific march-in after an agency determination; nonprofit assignment approval unless an invention-management exception applies; no parallel default approval rule for for-profit sellers; domestic-manufacturing and nonprofit-income safeguards.
EU / Horizon Europe	Results generated by beneficiaries in a Horizon Europe action; call- or procurement-specific provisions may activate additional controls.	The generating beneficiary owns the contractual result; inventor-to-beneficiary entitlement remains governed by national law.	Noncommercial EU policy access; call-specific emergency and third-country transfer controls; procurement-specific commercialization remedies.
Germany (national example)	Employee-invention law applies because of employment; NKBF 2017 adds a separate layer only when incorporated into specified federal cost-basis grants to commercial enterprises.	Employee reports a service invention; an employer claim—ordinarily deemed made unless the invention is released—transfers economic rights, with a university-specific variant.	Reasonable inventor remuneration; 30% of gross exploitation revenue for a claimed university invention; covered grant terms may restrict specified foreign exploitation, reserve public or security rights, require timely exploitation, and support repayment.
Japan	Specified IP from national-government-entrusted technology R&D or contracted software, when the contractor accepts four statutory commitments.	The national government may decline to acquire covered rights; Article 17 supplies no general contractor-election rule.	Conditional government use; licensing after unjustified nonuse; approval for specified transfers and exclusive rights.
China	Results of science-and-technology-plan projects established with fiscal funds, subject to national-security, national-interest, and major-public-interest limits.	The project undertaking entity is authorized to obtain and commercialize resulting IP, alongside employee-invention law.	State or third-party use after nonimplementation or for public interests; foreign-disposition approval; statutory reward and remuneration duties, including qualified default floors.
South Korea	Outcomes of national R&D projects; qualifying "public technology" also activates the Technology Transfer Act layer.	The performing institution generally succeeds to researcher rights; exceptions permit researcher, joint, or state ownership.	Utilization and dissemination duties; ex ante state-ownership exceptions; administrative step-in and revenue sharing.

Source: Primary authorities cited inline in Sections 2–6. Local terms are compared by function and are not treated as legal equivalents.

Assume a U.S. company seeks an exclusive worldwide license from a research institution identified as the owner of a patent. For a U.S. patent, counsel asks whether it is a subject invention, whether the contractor properly acquired and elected title, what federal license and utilization duties remain, and whether § 204 applies. If the proposed transaction instead transfers title to an assertion-focused affiliate, counsel must also determine whether the seller is a nonprofit, whether agency approval or the invention-management exception applies, which nonprofit provisions pass to the buyer, and how reporting and practical-application evidence will be maintained. For a Horizon result, counsel starts with the generating beneficiary, national inventor-entitlement law, and the call, grant, and consortium terms. For a German invention, counsel adds the employee-invention claim and remuneration rules and any incorporated grant conditions. In Japan, the threshold is covered Article 17 entrusted work; in China, a fiscally funded plan project and possible foreign-disposition approval; and in Korea, the overlapping National R&D and public-technology layers. The patent register supplies the same owner answer in each example but not the same commercialization authority.

Six propositions follow.

First, scope and trigger are outcome-determinative. The U.S. framework attaches to contractor-owned subject inventions conceived or first actually reduced to practice in performance of work under a federal funding agreement. Japan’s Article 17 reaches only specified rights from national-government-entrusted technology R&D and contracted software, and only when the contractor accepts four commitments. Horizon Europe governs results generated by funded beneficiaries, with stronger controls depending on the call or procurement instrument; German grant conditions add a distinct layer only when incorporated into a covered award. China and South Korea likewise key their rules to defined fiscally funded or national-R&D projects, with an additional public-technology layer in South Korea. Similar allocation mechanisms therefore do not imply similar legal coverage.

Second, institutional ownership is not a single legal mechanism. In the United States, Bayh–Dole operates after the contractor acquires inventor title and then permits election to retain it. Japan authorizes government nonacquisition when prescribed conditions are accepted. China authorizes the project entity to obtain IP, working alongside employee-invention law. South Korea generally requires succession from researcher to performing institution. Horizon Europe allocates results generated by funded beneficiaries, while inventor entitlement remains governed by national law; Germany illustrates a statutory employer-claim mechanism within that national layer.

Third, commercialization policy should be separated from the legal test for intervention. Bayh–Dole states commercialization and public availability as objectives, but operationalizes them for contractor-retained inventions through a defined practical-application standard, utilization reporting on agency request, and discretionary march-in when effective steps are lacking in a field of use. Horizon Europe instead uses a best-efforts exploitation duty; Japan and China use more direct nonuse triggers; and South Korea imposes affirmative utilization and dissemination measures. The useful comparison is therefore not whether a statute uses the word “commercialization,” but what conduct it expects, how performance is monitored, and what remedy follows nonperformance.

Fourth, formal authority and administrative practice are different comparative dimensions. The United States supplies the clearest example: march-in appears prominently in the statutory architecture, but GAO reported in 2026 that no agency had exercised it and that agencies had declined approximately a dozen requests. Formal legal powers should therefore be separated from evidence of implementation; the breadth of a statutory power does not establish the frequency of its use. (U.S. Gov’t Accountability Off., GAO-26-107885, Intellectual Property: Information on Draft Guidance to Assert Government Rights Based on Price 1–2 (2026).)

Fifth, “march-in” covers unlike powers. U.S. march-in is a defined, discretionary licensing mechanism attached to subject inventions; it is not automatic title forfeiture. Japan requires third-party licensing after unjustified prolonged nonuse when the government makes a reasoned utilization determination. China permits state or third-party use after unjustified nonimplementation and for broad national or public interests. Korea relies on public-technology utilization duties and administrative step-in, supplemented by generally applicable compulsory-use law. Ordinary Horizon grants use narrower and often call-conditional licensing and transfer controls, although pre-commercial procurement contains a more recognizable commercialization-default remedy.

Sixth, disposition and inventor economics do not follow a single U.S.-versus-foreign pattern. A U.S. nonprofit generally needs agency approval to assign U.S. subject-invention rights unless the buyer has invention management as a primary function, while a for-profit contractor does not face the same default approval rule. China requires approval for foreign transfers and exclusive foreign licenses of relevant publicly funded IP. Japan ordinarily requires approval for specified transfers and exclusive rights. Horizon calls can activate objections to certain non-associated-country transactions. Germany, China, Korea, and Japan each use general or special employee-invention remuneration regimes that can materially affect licensing proceeds, transaction liabilities, and institutional incentives.

IX.             Conclusion

For foreign patent attorneys, the principal U.S. lesson is that Bayh–Dole is neither an automatic university-ownership rule nor a blanket command to commercialize every federally supported invention. It applies to a contractor-owned subject invention conceived or first actually reduced to practice in performance of work under a federal funding agreement. It permits the contractor to elect to retain title while preserving disclosure and patenting duties, utilization reporting, a paid-up government license, nonprofit-specific obligations, discretionary march-in, and the domestic-manufacturing condition applicable to specified exclusive U.S. rights. “Practical application” supplies the operative utilization benchmark. Section 209’s more express development-plan requirements govern the different situation in which an agency licenses a federally owned invention.

A sale does not cleanse a subject-invention patent of the federal bargain. Bayh–Dole does not categorically prohibit ownership by an NPE or patent assertion entity, but the seller’s status controls the transfer path: nonprofit assignments generally require agency approval unless the purchaser has invention management as a primary function, while for-profit contractors face no parallel default purchaser restriction. The Government’s license remains; march-in and § 204 expressly reach downstream actors in defined circumstances; and reporting, approval, nonprofit-income, and patent-management duties must be traced through the award and transaction documents. NPE ownership is neither a march-in ground nor proof of practical application.

For U.S. patent attorneys, the principal foreign-law lesson is that searching for another country’s “Bayh–Dole Act” can conceal more than it reveals. Japan provides the closest structural analogue among the systems discussed at the level of its covered government–contractor bargain, but Article 17 has a narrower trigger. China combines institutional control with broader public-interest intervention, foreign-disposition controls, and researcher-reward rules. South Korea relies on institutional succession, affirmative utilization measures, revenue allocation, and public-technology dissemination. EU/Horizon Europe is program-specific and operates alongside national inventor-entitlement law; Germany illustrates one national and grant-term layer rather than a European rule.

Both audiences can use the same method. Counsel should ask what mechanism allocated institutional control, what funding or legal trigger brought the invention or other output within the regime, and what continuing public constraints follow it. Those questions reveal matters that a patent register cannot: whether title was properly acquired, whether reporting or exploitation duties remain open, whether government or third-party use can be required, whether a foreign license or transfer needs approval, and whether commercialization proceeds must be shared.

The useful cross-border question is therefore not simply, “Does this jurisdiction have Bayh–Dole?” It is: “What public-funding bargain governs this result, and what does that bargain mean for ownership, commercialization, licensing, transfer, and enforcement?”