Innovation in defence: how start-ups are making their mark

Fifty years ago, many of the tech­no­lo­gies used in the civil­ian sec­tor came from the defence sec­tor. Today we are wit­ness­ing the oppos­ite move­ment, with the emer­gence of innov­a­tions in defence com­ing from the civil sec­tor. Why is this so?

There are sev­er­al reas­ons for this phe­nomen­on. Firstly, tech­no­lo­gic­al cycles in the civil­ian sec­tor are much faster than the long cycles of arm­a­ment pro­grammes, which can last from ten to forty years. Secondly, R&D invest­ments in the civil­ian sec­tor some­times greatly exceed the defence budgets of most West­ern nations. For example, Hua­wei’s R&D budget (€21 bil­lion) exceeds Israel’s mil­it­ary budget by one bil­lion, even though Israel is a major play­er in terms of defence innov­a­tion. But we should also be sure what we mean by innov­a­tion here.

The Min­istry of the Armed Forces has been invest­ing for a long time in so-called “planned innov­a­tion”, in the sense of long-term R&D, which aims to anti­cip­ate tech­no­lo­gic­al break­throughs and ensure man­age­ment of emer­ging tech­no­lo­gies of a stra­tegic nature. But there are also innov­a­tions in their use, too. For example, the diver­sion of tech­no­lo­gies from the civil­ian sec­tor (e.g. drones), not fore­seen in the tech­no­lo­gic­al roadmaps, can respond to oper­a­tion­al chal­lenges and be developed more rap­idly by com­pan­ies tar­get­ing civil­ian mar­kets. The Defence Innov­a­tion Agency (DIA), cre­ated in 2018, has the ambi­tion, among oth­ers, to detect, mon­it­or and cap­ture oppor­tun­it­ies from the civil­ian world.

The arrival of these civil­ian play­ers in the world of defence (for example Par­rot, to take the case of drones) raises new ques­tions. The issue of con­fid­en­ti­al­ity is not the most dif­fi­cult to man­age. The real issues, which we encounter in the con­text of open innov­a­tion with large defence groups, are how to organ­ise the co-innov­a­tion pro­cess, by cor­rectly reg­u­lat­ing the intel­lec­tu­al prop­erty rights. Col­lab­or­a­tion between start-ups and large groups can some­times lead to a phe­nomen­on of pred­a­tion, wheth­er vol­un­tary or not. Yet, con­trac­tu­al solu­tions do exist. 

Fur­ther­more, the DIA, through its open innov­a­tion approach, has the mis­sion of detect­ing and cap­tur­ing the “right pro­ject at the right time”: that is to say, a pro­ject that responds to the chal­lenges expressed by the armed forces, cap­able of reach­ing levels of matur­ity not only tech­no­lo­gic­al but also mar­ket and user. And often the busi­ness mod­el of these start-ups can­not be based solely on the defence mar­ket, which is both too small and marked by excess­ively long acquis­i­tion cycles.

The prob­lem of dual­ity must be observed on sev­er­al levels of ana­lys­is: strategy, pro­jects, and busi­ness mod­el. The poten­tial for dual­ity often decreases as tech­no­logy devel­ops, but also as uses, pro­cesses and organ­isa­tion­al choices change. The devel­op­ment costs for defens­ive spe­cific­a­tions are usu­ally so large for the scale of a start-up’s resources that it can rarely focus on sev­er­al mar­kets at the same time, at least in the begin­ning. The lin­ear mod­el of tech­no­logy devel­op­ment that has been used as a ref­er­ence to define dual­ity is some­what out­dated, as we notice today that some (rare) defence start-ups have demon­strated the abil­ity to absorb arti­fi­cial intel­li­gence devel­op­ments in the civil­ian world and adapt them to the defence context.

This is the case of Pre­li­gens: a start-up which devel­ops soft­ware using AI to auto­mat­ic­ally ana­lyse mass data from mul­tiple sources, not­ably satel­lite optic­al imagery, and alerts intel­li­gence agents when an abnor­mal situ­ation is detec­ted on a site of stra­tegic interest.

In prac­tice, the dual­ity is very dif­fi­cult to man­age, espe­cially from a stra­tegic point of view and on the ques­tion of resource alloc­a­tion. There are dual tech­no­lo­gies, but real dual SMEs are rare (we can nev­er­the­less cite the case of MC2 Tech­no­lo­gies, on anti-drone combat).

Talk­ing about a defence eco­sys­tem con­trasts some­what with a concept that is still very reg­u­larly used by the play­ers in the defence mar­ket: that of the defence indus­tri­al and tech­no­lo­gic­al base (DITB), made up of about ten indus­tri­al prime con­tract­ors (some of which were State ser­vices before being privat­ised) and about 4,000 SMEs.

Open innov­a­tion rep­res­ents less than a quarter of the DIA’s budget. There is no revolu­tion. Nev­er­the­less, open innov­a­tion is rede­fin­ing the roles of play­ers in the defence world and their rela­tion­ships, espe­cially in cross-cut­ting areas such as arti­fi­cial intel­li­gence, where there is no his­tor­ic­al play­er that imposes itself on this vertical.

The major defence groups are increas­ingly sys­tems integ­rat­ors: they integ­rate all the sys­tems and com­pon­ents, a skill that only the very large play­ers pos­sess. But they no longer have full tech­no­lo­gic­al con­trol. SMEs are not just sub­con­tract­ors; they provide com­plete tech­no­lo­gic­al build­ing blocks. Open innov­a­tion also encour­ages the emer­gence of inter­me­di­ar­ies to align play­ers in an innov­a­tion situ­ation, for example the DIA’s Innov­a­tion Defense Lab, the innov­a­tion labs of the Armed Forces such as the Battle Lab Terre, or even cap­it­al investors.

The shift towards strategies of hyper-com­pet­i­tion and intens­ive innov­a­tion has been described for more than thirty years by research­ers in man­age­ment sci­ences, and the defence sec­tor is no excep­tion. This con­text of strong com­pet­i­tion has had a sig­ni­fic­ant impact, push­ing com­pan­ies to ration­al­ise the organ­isa­tion of devel­op­ment pro­jects. In the early 2000s, major tech­no­lo­gic­al dis­rup­tions made the sys­tem­at­ic design and stand­ard­isa­tion of devel­op­ment pro­cesses obsolete. 

This is per­haps where the defence mar­ket can shed some inter­est­ing light. Indeed, faced with the increas­ing com­plex­ity of sys­tems, the con­duct of arm­a­ment oper­a­tions through major pro­grammes is find­ing it increas­ingly dif­fi­cult to cohab­it with explor­at­ory approaches. These explor­at­ory pro­jects, char­ac­ter­ised by their uncer­tainty, must allow learn­ing loops to be car­ried out and sev­er­al con­cur­rent devel­op­ments to be car­ried out in order to avoid being “killed” by the qual­ity-cost-delay logic of oper­a­tions. The chal­lenge of the defence innov­a­tion eco­sys­tem today is to allow real break­throughs to emerge from these explor­at­ory projects.

The human factor is of course cru­cial for the armed forces as well as for the act­ors of the defence world. If the ques­tion of man­power is always import­ant, espe­cially in the cur­rent geo­stra­tegic con­text, the ques­tion of a qual­it­at­ive rise in pro­files is cent­ral. Accul­tur­a­tion to new tech­no­lo­gies is essen­tial to face the evol­u­tion of pro­fes­sions, equip­ment, and organ­isa­tions. In the fields of arti­fi­cial intel­li­gence, cyber­se­cur­ity, or nan­o­tech­no­lo­gies, many pro­fes­sions are in per­man­ent ten­sion. Cyber defence is one of the pri­or­it­ies of the armed forces in terms of recruit­ment. And it is here again that the power of an open eco­sys­tem can be illus­trated, as is the case in Bruz near Rennes, where pub­lic act­ors (the Dir­ect­or­ate Gen­er­al of Arma­ments, the Com­Cy­ber) and private act­ors (indus­tri­al­ists and start-ups) work closely togeth­er, includ­ing to attract talent. 

Interview by Richard Robert