Pattern no. 3 — “Paying for the first tank”: Strategy · Financial stake
Translated from the French, which remains the reference text. This translation is under review.
Pattern n°3 · Bootstrapping (primary), Alternation, Metabolism · tensions T11, T4, T7 · Legitimacy · volets I.1, II.1 · ↑ navigation
Schema F3.1 · Context of pattern no. 3
Tension(s) at work: Coopetition; public, private and citizen; uncoordinated and coordinated action — and, implicitly, the two dynamic challenges of momentum and demand. This pattern answers the question every infrastructure commons meets on its first day: an empty network is worth nothing to anyone — who enters it first, and why?
Context: A common infrastructure exists — the protocol is published, the reference implementation runs, the guardian is instituted. What remains is the wall: network effects. The network's value grows with its members, so the empty network is worth nothing, so no one enters it, so it stays empty. Facing the nascent commons stand established platforms whose pull captures each new user by default. The European context adds its own pathology, documented in this report: laws enacted, billions invested in R&D — and alternatives that no one buys, not even the public authority that funded them.
Forces at play: On one side, the individual rationality of waiting: for each actor taken in isolation, the optimal strategy is to let the others take the first blows and to join the network once it is worth something — the free rider of bootstrapping. On the other, the necessity of critical mass: below a threshold, the network produces none of the benefits that would justify joining it. The median compromise — sprinkling subsidies across all actors to encourage each a little — fails systematically: it funds supply without creating demand, multiplies pilots without ever igniting a network, and produces orphaned gems.
The pattern: One actor deliberately takes on, massively and at a single point, the cost of igniting the network — then withdraws. The corpus reveals four ignition strategies, which can be combined: the State as reference user — the public authority is not only the funder or the regulator of the commons, it becomes its first client, committing its own production to it (public procurement as an act of bootstrapping, not of support); the backed service — tying the nascent network to an existing flow one controls (Aadhaar backed by the payment of social benefits: using the rail becomes the path to a service that is already indispensable); the enlisted champion — bringing on board from the outset the actors whose presence makes the network credible (the banks on UPI, brought together by NPCI before the opening to the public); the chosen killer app — concentrating ignition on a single segment where the pain is greatest and the demonstration spectacular (Namma Yatri: mobility, the commissions of ride-hailing platforms, a drivers' union — the “Linux of mobility” did for Beckn what no campaign would have done). The cross-cutting rule: concentrate, never sprinkle — a fire is lit at one point, not by sharing out the matches. And the exit clause is constitutive of the pattern: bootstrapping is a phase, dated, with end criteria — past the threshold, the relay passes to metabolism (“Those who prosper maintain”, n°10): a new phase then opens, that of institutionalisation. Bootstrapping without an exit is called life support.
Operator(s) mobilised: Primary: bootstrapping — this pattern is its constitutive rule. Support: alternation — ignition is a moment in a sequence, not a regime; confusing it with permanent operation is the most common design error. Support: metabolism — the pattern is complete only if the relay is designed before ignition: knowing who will tend the fire before lighting it.
The invariant under test: The pattern's own risk is the gravest in the whole language: the most effective ignition strategy in the corpus — the backed service — is also the one that produced the Aadhaar paradox. Tying the rail to food aid guarantees adoption... because it makes adoption a condition of survival; and algorithmic exclusion then strikes precisely those who cannot afford to be excluded. Associated criterion of legitimacy: legitimate bootstrapping empowers, it does not coerce — and where constraint is unavoidable (regulatory obligations, exclusive public procurement), it must come with real alternative paths and a continuous measurement of the exclusion produced (“Judged by those it serves”, n°12). Corollary on the supply side: the State as first client must not become the State as sole client — a commons with a single buyer is as fragile as a commons with a single supplier.
Schema F3.2 · Paying for the first tank
Cases of embodiment: Indian: the JAM trinity is the most massive bootstrapping in digital history — Aadhaar backed by direct social transfers, UPI launched with the major banks pre-enrolled by NPCI and then violently accelerated by the demonetisation of 2016: India paid for the first tank at a high price, including the ethical price this report documents. Namma Yatri illustrates the virtuous, bottom-up variant: a single, chosen use case, where the network demonstrates its superiority irrefutably before expanding. European: the turn is recent and still to be confirmed — the EOS explicitly sets up the State as reference user (public procurement directives, OSPOs, sovereignty built into investment), the EDIC Digital Commons organises the pooling of public demand between Member States, and eIDAS 2 creates a regulatory bootstrapping for the EUDI wallet: the obligation on large platforms to accept it. Europe, which long funded supply without committing its demand, is learning the pattern — this is perhaps the most important Indian lesson it has received.
Schema F3.3 · The growth loop, and an open question
Counter-example: Two symmetrical failures teach the pattern in the negative. Solid, the project to re-decentralise the web led by Tim Berners-Lee: the most legitimate architecture there could be, the inventor of the web himself — and no bootstrapping. No State as user, no backed service, no champion, no killer app: the conviction that technical rightness would suffice. The network is still waiting for its users. Conversely, the French “sovereign clouds” of the 2010s (Numergy, Cloudwatt): the State paid — but for the wrong tank. It took a stake in the capital of the suppliers instead of committing its demand: it funded factories without becoming their client, and its own administrations went on buying elsewhere. Both entities died within a few years. The theorem, in one line: a network is not bootstrapped by funding supply; it is bootstrapped by guaranteeing demand. Solid paid nothing; Cloudwatt paid at the wrong counter.
Signs of misapplication: The public funder of the commons does not use it itself (the hypocrisy test: look at what the subsidising administrations actually buy). Subsidies are spread across dozens of projects with no commitment to adopt any of them. The bootstrapping budget has neither an exit threshold nor end criteria. The indicators count projects funded, never transactions completed. Pilots have followed one another for years without any ever being generalised (pilotitis is to bootstrapping what life support is to metabolism). Compulsory adoption is rolled out with no alternative path and no measurement of exclusion. Two symptoms warrant an alert; three, a diagnosis.
Related patterns: Upstream: “Running code first” (n°2 — only what runs can be bootstrapped) and “Minimising the surface of the common” (n°4 — only what is adoptable can be bootstrapped). Downstream: “Those who prosper maintain” (n°10 — the necessary relay of any bootstrapping), “Arriving first at the periphery” (n°1 — the choice of the ignition point), and “Judged by those it serves” (n°12).