Chapter 2 — Research and development procurementProposed — COM(2026) 567
Article 22 — Multiple sourcing
In brief
This article makes multiple sourcing the default in R&D procurement: public buyers are to award several contracts to several contractors within the same procedure. It requires buyers to state the minimum number of contracts expected, to award them by ranking, and to weigh the benefits of spreading work across suppliers. A limited set of justified cases allows a single award instead.
Key points
Buyers must apply multiple sourcing and indicate in the documents the minimum number of contracts expected to be awarded.
Contracts are awarded and signed in order of the ranked list, highest to lowest score, unless there are duly justified reasons to proceed otherwise, all within the same procedure.
In deciding how many contracts to award, buyers weigh benefits such as avoiding over-reliance on a single supplier, reducing technology and innovation risk, addressing weak competition or lock-in, ensuring redundancy, and, where justified, reserving contracts for operators originating in Member States.
A single contract may be awarded, with reasons stated in the documents, for certain R&D consultancy services (defined by CPV codes) or where, for non-pre-commercial procurements, the market is too small, the benefits do not outweigh the extra costs, or security and public-safety interests so require.
The Commission may adopt delegated acts to amend the list of single-award cases against defined criteria.
What it means in practice
Public buyers will need to design R&D procedures around parallel awards, set out the expected minimum number of contracts, and justify any departure from strict ranking or any move to a single supplier. For innovative firms, start-ups and SMEs, multiple sourcing is intended to widen access, sustain competition and reduce dependence on incumbents, since more than one contractor can be selected from the same procedure.
Anthony Bochon’s analysis
Making multiple sourcing the default — awarding several contracts to several contractors within one procedure — is a structural policy choice, and reading it against the current climate I see resilience and anti-lock-in written into procurement law itself. The rationale in paragraph 4 is candid about its objectives: avoiding over-reliance for critical technologies and raw materials, reducing innovation risk by funding parallel approaches, stimulating competition and building technical redundancy, with an express hook to reserve contracts for Member-State operators on security grounds under Article 28(2). This is the same strategic-autonomy thread that runs through the IP provisions, now expressed as market structure.
In my view the litigation and audit interest will concentrate on the paragraph 5 derogations permitting a single award, particularly the “market too small” and “benefits do not outweigh the extra costs” tests, which are inherently evaluative and must be justified in the procurement documents. Because the Commission may reshape that derogation list by delegated act against explicit commercialisation, market-access and strategic-autonomy criteria, addressees should expect the boundary between single and multiple sourcing to keep moving, and buyers should document their cost-benefit reasoning as if it will be reviewed.
Official text — Article 22 (COM(2026) 567)
1. Public buyers shall apply multiple sourcing in R&D procurement procedures and
shall award multiple R&D procurement contracts to multiple contractors.
2. Public buyers shall indicate in the procurement documents the minimum number of
R&D procurement contracts that are expected to be awarded in accordance with
paragraph 1.
3. R&D procurement contracts under multiple sourcing shall be awarded and signed in
the order of the ranked list of tenders from the highest to the lowest score based on
their evaluation in accordance with Article 11 unless there are duly justified reasons
to proceed otherwise, in full com pliance with the principles of equal treatment and
transparency. All R&D procurement contracts shall be awarded within the same
R&D procurement procedure.
4. To determine the number of R&D procurement contracts to be awarded in
accordance with paragraph 1, public buyers shall take into account the benefits that
multiple sourcing can generate to:
(a) avoid the over -reliance on a single or a limited number of contractors for
products, works or services that are essential to the public buyer’ inter nal or
external operations, or for products, works or services that are based on critical
technologies or critical raw materials;
(b) reduce technology or innovation risks including by stimulating the development,
testing or integration of alternative results from different contractors;
(c) tackle situations where there is not sufficient competition on the market or
prevent or reduce supplier lock -in by stimulating more competition on the
market;
(d) mitigate service disruptions and single point of failure by ensuring technical
redundancy in the public buyer’s infrastructure, network or service delivery
system;
(e) where necessary and justified for protection of the security and public safety
interests of the Union or one or several of its Member States with in the meaning
of Article 28(2), reserve a minimum number of R&D procurement contracts for
economic operators originating in Member States in accordance with Article
11(6).
5. By way of derogation from paragraph 1, where justified by any of the following
reasons, which shall be specified in the procurement documents, public buyers may
award a single R&D procurement contract without applying multiple sourcing in
accordance with this Article:
(a) R&D procurement procedures which have as their only subject th e
procurement of research and development consultancy services covered by
CPV codes 73200000-4, 73210000-7 and 73220000-0;
(b) in R&D procurements that are not pre -commercial procurements, where any of
the following conditions are fulfilled:
(i) the resu lts of the outcomes of the market consultation referred to in
Article 9 and market research indicate that the market is too small for
multiple contractors to participate in the R&D procurement procedure;
(ii) the long-term benefits of multiple sourcing as listed in paragraph 4 of this
Article do not outweigh the extra costs for the public buyer applying
multiple sourcing;
(iii) the award of a single R&D procurement contract is necessary and
justified for ensuring the protection of the security and public sa fety
interests of the Union or one or several of its Member States within the
meaning of Article 28(2);
The Commission is empowered to adopt delegated acts in accordance with Article 39
to amend the list of cases set out in the first subparagraph of this p aragraph taking
into account the following criteria:
(a) the impact on increasing the commercialisation of R&D results;
(b) the impact on improving the access of innovative enterprises, including in
particular innovative startups and scaleups, and new entrants to the market;
(c) the impact on socio-economic or technological development in the Union;
(d) the impact on the Union’s strategic autonomy and economic security;
(e) the impact on protecting public interests;
Source: European Commission, proposal for a Regulation establishing the European Innovation Act, COM(2026) 567 final, 9 September 2026. Read the official proposal (PDF). Text may change during the legislative process.
Anthony Bochon’s analysis
Making multiple sourcing the default — awarding several contracts to several contractors within one procedure — is a structural policy choice, and reading it against the current climate I see resilience and anti-lock-in written into procurement law itself. The rationale in paragraph 4 is candid about its objectives: avoiding over-reliance for critical technologies and raw materials, reducing innovation risk by funding parallel approaches, stimulating competition and building technical redundancy, with an express hook to reserve contracts for Member-State operators on security grounds under Article 28(2). This is the same strategic-autonomy thread that runs through the IP provisions, now expressed as market structure.
In my view the litigation and audit interest will concentrate on the paragraph 5 derogations permitting a single award, particularly the “market too small” and “benefits do not outweigh the extra costs” tests, which are inherently evaluative and must be justified in the procurement documents. Because the Commission may reshape that derogation list by delegated act against explicit commercialisation, market-access and strategic-autonomy criteria, addressees should expect the boundary between single and multiple sourcing to keep moving, and buyers should document their cost-benefit reasoning as if it will be reviewed.