Ifri's full DefTech memo turns a broad collaboration argument into questions about integration, intellectual property and decision rights. Those choices determine whether a partnership produces a repeatable business.
Ifri's February 2026 memo on Europe's defence-technology industry argues that collaboration between startups and established industrial groups needs explicit governance. Authors Alexandre Papaemmanuel and Laure de Roucy-Rochegonde describe a difficult relationship between companies organised around iterative digital development and established manufacturers working through longer industrial cycles. Ifri's research summary
The full English memo gives that argument a practical dimension. It calls for clearer allocation of component development, integration, testing and qualification, alongside arrangements for intellectual property, liability, market access and revenue sharing. It also advocates common reference architectures with governance for changes. These are the authors' proposals, rather than a new European contracting rule. The full Ifri memo
For a commercial team, the central question is who will turn the combined technology into a product that a customer can accept, support and buy again. A partnership announcement begins that conversation. Its commercial substance appears in the division of work.
A useful partnership supplies something that neither participant can economically provide alone within the required timeframe. That may be an integration environment, a tested component, customer support capacity, manufacturing expertise or a route into an established product.
The reason should be specific enough to guide investment. A startup seeking access to a prime's customer base has a different objective from a prime seeking a specialist software module. Both may describe the arrangement as innovation cooperation, but their expected returns and decision rights differ.
Consider a hypothetical company supplying software that organises maintenance records. An established equipment manufacturer offers access to its installed product base and service organisation. The software company contributes a reusable product, while the manufacturer provides equipment knowledge and customer support relationships.
A credible agreement would explain how that combination improves a maintenance workflow and who pays for the first integration. A vague commitment to explore opportunities leaves the essential commercial decision unresolved.
Interfaces are often treated as a technical detail beneath the commercial agreement. They can instead determine whether the proposed partnership works at all.
In the maintenance example, the software supplier may need access to a documented data export, a representative test environment and an agreed way to identify equipment records. The manufacturer may need evidence that the resulting service works with its supported equipment versions. Neither party can complete its own task if the other leaves these dependencies open.
The agreement should therefore make integration work visible in scope and schedule. Someone must maintain the reference environment, resolve incompatible changes and decide when the joint product is ready for customer acceptance.
This allocation is especially consequential when the first demonstration succeeds through intensive assistance from both founders and senior engineers. That effort can prove feasibility while leaving the cost of ordinary delivery unknown. The next stage should establish how a normal project team reproduces the result.
A startup may retain ownership of its existing software while granting a partner permission to distribute a supported version. The parties may jointly fund a customer-specific extension. These are different assets and commercial arrangements.
The difficult questions concern how the extension affects the core product. Can the startup incorporate general improvements into future releases? Can the manufacturer continue supporting customers if the startup changes strategy? Who may use customer-provided records to improve the product, and under what agreed conditions?
There is no universally optimal answer. Broad exclusivity may be valuable when a partner commits substantial funding and customer access. The same exclusivity can be costly if it restricts the startup's market without a corresponding minimum commitment.
The commercial analysis should compare the rights transferred with the resources received. A prominent partner's brand is relevant, but it does not substitute for a funded workload, access to necessary information or a defined route to customer revenue.
A software supplier may release improvements frequently. A manufacturer may support equipment configurations for years. A joint product needs a workable relationship between these timelines.
In the example, a new software release could alter data fields or the support interface used by the manufacturer's service team. Releasing it immediately to every customer might create avoidable disruption. Delaying every improvement until a major equipment update could undermine the software's value.
A sensible commercial design identifies supported versions, a compatibility process and who funds maintenance of older configurations. The cost belongs somewhere: within the product price, a support agreement or a separately commissioned change.
This is where the governance described by Ifri becomes concrete. The decision-making arrangement must handle routine disagreements about readiness, priority and cost, rather than relying on personal relationships between senior executives.
An introduction to a government team, inclusion in a catalogue and participation in a funded delivery are different forms of access. Partnerships become easier to evaluate when these activities are described accurately.
A startup could gain substantial value from selling through a manufacturer that already provides the relevant support service. It may never need a direct government contract. Alternatively, it may want direct customer discovery while using the manufacturer only for integration or production.
The chosen structure determines who hears user feedback, controls the proposal and owns the renewal discussion. If the startup receives only technical tickets, it may have limited visibility into why customers buy or stop buying. If the manufacturer carries customer responsibility, it needs sufficient influence over changes that affect delivery.
Our analysis of CSIS's new-entrant data and repeat-business evidence explains why participation alone is an incomplete measure of commercial progress. Partnership coverage should follow the same discipline.
The first sale can require exceptional effort: migration, documentation, training and negotiation over unfamiliar responsibilities. The partnership becomes more attractive if subsequent customers can reuse the resulting work.
That does not mean every implementation must be identical. It means the parties should understand which differences are configuration and which create a new engineering obligation. Otherwise a growing customer count can produce an expanding collection of bespoke products.
For the maintenance service, reusable connectors and standard support documentation may improve delivery economics. Repeatedly rebuilding the data model for each customer would suggest a different business. The price and staffing plan should reflect the actual pattern.
The same issue appears in joint procurement and commercial standardisation. Several customers only create production economies to the extent that the work can be shared.
Partnerships must also survive staff changes, acquisitions and shifts in product priorities. A support obligation may last longer than the executives who negotiated it.
A practical continuity arrangement identifies the documentation, product knowledge and access required to keep customers supported. It also explains how responsibilities transfer if one party withdraws from a product line or the relationship ends.
The appropriate mechanism depends on the product and contract. What matters commercially is that the buyer is not left depending on an informal promise that two companies will always remain aligned.
This planning can improve the initial sale. A customer assessing long-lived equipment or information services has reason to ask whether the combined offering remains supportable after a corporate change. An answer grounded in documented responsibilities is stronger than a general assertion of strategic commitment.
Ifri's contribution is to make industrial cooperation a question of organisation and incentives as well as technology. BDI's commercial interpretation is that a useful partnership should make a defined customer outcome easier to purchase and sustain.
Evidence of progress includes a funded integration, a supported joint release, a clearly allocated service obligation or a repeat delivery using the same product foundation. These milestones are more informative than the number of organisations appearing in an announcement.
For a startup or established group considering a new relationship, the decision is therefore concrete: identify the work each party enables, the investment each commits and the commercial rights needed to make that investment worthwhile. Governance becomes valuable when it turns complementary capabilities into a product that can be delivered repeatedly.
Public primary research summaries, introductory material or project descriptions reviewed on 6 September 2026. The text identifies the material reviewed and distinguishes research claims from BDI's commercial analysis. No independent technical validation or review of every full-length research chapter is claimed.
Suggest a correction