Skip to content

Systemic dependence: the architecture of the digital stack

European organisations buy their digital technology through a stack of complementary layers, from lithography and chips to cloud infrastructure, software, AI models and the payment rails every digital sale runs on. This programme reads the stack as one architecture of dependence: which positions others depend on, which of them earn above the cost of their capital, how control spreads from one position to the next, and what that architecture means for the value Europe sends abroad and the capability it builds at home. Each layer is mapped from rated public sources: competition authorities' findings, audited filings, analysts' figures and the lab's own measurements.

The Architecture of Dependence: Control Points, Value Capture and Capability in the Digital Stack

Paper page PDF

The paper maps the stack's control points, the positions many others depend on, costly to substitute and held by few, and measures the rent each earns on the capital the UK competition authority counts: 48 cents of each euro at the leading accelerator maker, 25 at Microsoft's productivity suite, 22 at the leading foundry, 6 to 11 in cloud infrastructure and 32 to 39 at the card schemes, against a median of 2 where several suppliers compete. The rent is also the channel through which control spreads: the holders of lower control points earn their margin on what the positions above them buy, so they back those positions, and backing runs up the stack against the payments. A census of 88 annual filings checks the map's edges, and every documented financing is tested against that prediction. Opening a control point helps buyers and moves value to the next control points. Europe's four holds earn value, largely from outside Europe, without anchoring capability in the layers that use them. Cloud infrastructure is the worked example throughout. Status: draft working paper (v0.3).