DC EDIC and the Next Eighteen Months of European Infrastructure

The window

Laurent Rojey took over as Director of the Digital Commons EDIC on 1 April 2026. By the time you read this he has been in post for three weeks. He has roughly eighteen months — until the next Multiannual Financial Framework debate begins in earnest, until the Commission reshuffles attention to its next initiative, until every founding minister gets distracted by an election cycle — to convert a legal vehicle into a procurement instrument.

That is the actual deadline. Not the official 2027 work programme milestones. Not the publication of the first State of the Digital Commons Report. The deadline is the moment at which a procurement officer in a German Land, a French ministry, or a Dutch municipality can issue a tender, receive bids that include DC-EDIC-supported components, and award a contract under a framework that already exists.

Either DC-EDIC produces a SKU in eighteen months, or it joins GAIA-X in the museum of European federation projects that talked their way into oblivion.

Why "a catalogue line" is the unit of measure

Procurement officers do not buy infrastructure. They do not buy values. They buy line items on a framework agreement, with a price, a delivery date, a warranty clause, and a vendor that can sign an invoice. Anything else is consultation.

This is the lesson of GAIA-X. The federation had governance, working groups, white papers, and a logo that appeared on conference banners for half a decade. What it did not produce, until very recently and only in narrow national channels, was a contract that a public buyer could actually execute against. The result was that every major hyperscaler kept selling, every European challenger kept fundraising, and every minister kept giving the same speech about strategic autonomy.

The DC-EDIC inherits this as its first failure mode. The "One-Stop-Shop and Expertise Hub" promised in the work programme can be a forum, or it can be a shopping cart. Only one of those is infrastructure. The other is a conference series.

The procurement test is not technical. It is administrative. Can a buyer issue a tender that names a specific DC-EDIC-coordinated component — say, an open-source identity broker, a sovereign object storage stack, an EU-hosted foundation model — and expect to receive responsive bids from vendors who have already done the eIDAS, the GDPR, and the Data Act compliance work? Today, in April 2026, the answer is no. In eighteen months, it has to be yes.

What has to happen

Three things, in three blocks of six months.

Months one to six: convert the 100 Day Challenge into pilot tenders. The 100 Day Challenge — DC-EDIC's first announced initiative to build sovereign and interoperable open source components — is a useful prototyping format and a terrible end product. Its outputs need to land in real procurement workflows, not in another GitHub showcase. That means at least one founding member state (France is the natural candidate; the consortium's statutory seat is in Paris) must commit to running a live tender that names a 100 Day Challenge component and accepts it under existing public procurement rules. Without this step, the Challenge is a hackathon. With it, it is a precedent.

Months six to twelve: open the EU Sovereign Tech Fund as a maintainer-grade funding line. The pilot of the European Sovereign Tech Fund, developed with Germany's Sovereign Tech Agency, is the single most important deliverable in the whole work plan. It is also the one most likely to be smothered by process. The Sovereign Tech Agency model works because it pays individual maintainers — the people who keep cURL, OpenSSL, and Sequoia PGP running — directly, on contracts measured in tens of thousands of euros, with paperwork measured in pages, not binders. If the EU-STF pilot reproduces that operational discipline, it will fund the actual sub-stratum of European digital infrastructure within a year. If it requires consortium membership, multi-party DEP funding applications, and biannual EC reviews, it will fund three Brussels intermediaries and zero packages.

Months twelve to eighteen: cross-border framework agreements. The single proof point that DC-EDIC has become infrastructure is that one ministry in one founding member state can buy a service or component, supported by a maintainer in a second founding member state, using a contract template approved by all five. This is harder than it sounds. It cuts across national procurement codes, intellectual-property regimes, and tax treatment of grants versus services. It is also the only way the consortium delivers anything other than coordination. A cross-border framework agreement is the European Schengen moment for digital procurement. Eighteen months is enough time to sign one. It is not enough time to sign several.

The risks

Three failure modes are visible from here.

The first is captured forum syndrome. Every DC-EDIC observer who has watched a multi-stakeholder body knows the pattern: sufficient governance complexity that nothing controversial ever gets decided, sufficient public output that nobody can quite call it dead. The defence against this is operational: hire fewer policy advisors, hire more contracting officers and procurement lawyers.

The second is national capture. France hosts the seat. Germany funds the Sovereign Tech Agency. Italy runs ACN. The Netherlands has been the most consistent voice for cross-border interoperability. If any one of these begins to treat DC-EDIC as an extension of national industrial policy, the consortium loses its only structural advantage — that it is the multilateral instrument none of the others is — and becomes a duplicative layer.

The third is talent geometry. Rojey is a credentialled choice, with the right policy résumé to negotiate the Brussels-Paris-Berlin triangle. He needs to recruit at least one operator with hyperscaler procurement experience and at least one open source maintainer with credibility outside the EU bubble. Without both, the consortium will be unable to translate between the people writing tenders and the people who would otherwise win them.

Closing

Procurement officers do not read manifestos. They sign purchase orders. The next eighteen months are about whether the Digital Commons EDIC can produce one — a real one, with a vendor name, a price, and a delivery date — that demonstrates the consortium is infrastructure rather than rhetoric.

The window is open. It is also the only one we get for a while. Keep the receipts.

Read the companion essay on why sovereignty has to mean more than residency, or get in touch.