When geopolitics becomes an architecture variable
There's a question few organizations ask before the answer becomes urgent: what does our IT system actually depend on? Not what it does. What it rests on. The undersea cables carrying 97% of global internet traffic. The American hyperscalers whose servers are subject to the Cloud Act. The Taiwanese chips without which no infrastructure gets renewed. The SaaS vendors headquartered in San Francisco with data stored in Dublin — under a jurisdiction that answers to Washington.
As of March 2026, three shifts structure the environment of every French enterprise IT system. They aren't new. They are now permanent and cumulative.
Energy and maritime instability. Tensions in the Gulf and attacks on desalination infrastructure and oil tankers are disrupting global supply chains. The consequence for IT systems: datacenter energy costs rise, equipment delivery times stretch out, and hosting decisions become full-blown geographic decisions.
Sino-American technological bipolarization. The US-China standoff forces companies to pick a technological side, often without realizing it. Using TikTok for communications, deploying Huawei equipment in factories, and hosting data on AWS — these are three distinct geopolitical bets, made implicitly, without deliberation.
European stalemate. Europe is negotiating its defense, absorbing American tariffs, and struggling to produce a sovereign technological response equal to the stakes. Initiatives like SecNumCloud and EUCS are moving forward, but at industrial scale, dependence on non-European hyperscalers remains massive.
So the question is no longer: how do we protect our IT system? It is: how do we architect an organization capable of absorbing shocks it hasn't anticipated, while continuing to deliver value — and remaining able to decide?
What traditional architectures don't see
Classic enterprise architecture approaches were designed for a linear world: plan three years out, produce a map, define a target, build a roadmap. This framework has one virtue: it gives the illusion of control. It has one massive flaw: it is blind to rupture.
The real problem isn't technical. It's structurally political, in the noble sense: in nearly every large organization there is a gap between those who decide strategy and those who run the systems. This gap goes by many names — silos, technical debt, lack of alignment — but at its core it is a translation failure. No one is in charge of connecting the world's upheavals to the reality of the systems.
→ The executive committee approves a strategic partnership without knowing the underlying technical dependencies — and discovers six months later that the partner is subject to the Cloud Act.
→ The infrastructure team learns the implications of NIS2 or DORA through a memo from the risk department, with no translation into concrete architecture constraints.
→ The cloud vendor map doesn't mention second-tier subcontractors — the ones who, in the event of OFAC sanctions, become invisible risks.
→ Architecture decisions get contradicted mid-execution because the strategic context has changed, and no one updated the underlying assumptions.
These symptoms aren't individual malfunctions. They are the effects of an architecture with no living update mechanism. An architecture whose assumptions are never revisited. An architecture that documents yesterday's world.
This is exactly what Continuous Architecture sets out to fix — and it's where the architect's role shifts.
The strategist-architect's value loop
The enterprise architect is not an upgraded technician. Nor is he a strategy consultant who ignores execution. He is something rarer and more valuable: the only actor in the organization able to make the full journey — from the boardroom to a product team's backlog, and back. In a loop. Continuously.
This loop has four beats. Each beat produces a distinct kind of value. Together, they make up what I call the strategist-architect's value demonstration. Click on each step.
This isn't intellectual curiosity. It's an organizational function — keeping the map of assumptions the IT system rests on up to date.
This translation is the heart of the job. It requires simultaneous mastery of strategic language (risk, sovereignty, competitiveness) and technical language (dependencies, coupling, reversibility). Few profiles master both. That's where the scarcity lies — and the value.
This isn't compliance. It's tooled autonomy.
This regular review — not opportunistic, but ritualized — is what distinguishes a living architecture from a dead repository. It's also what gives the architect his strategic credibility: not being right all the time, but having a method for detecting when he's wrong — and acting accordingly.
This loop is what distinguishes the strategist architect from the builder architect. The builder produces diagrams. The strategist produces decision-making capacity. In an unstable world, it's the latter that has value.
The dialogue zone: Continuous Architecture in action
Continuous Architecture, as formalized in the work of Gregor Hohpe and on continuous-architecture.org, doesn't propose a new framework. It proposes a change of posture: architecture as a living practice, rooted in permanent dialogue between strategy and execution.
This dialogue plays out in what's called the dialogue zone — the space where architecture decisions emerge not from a fixed plan, but from a structured conversation between the organization's layers. Click on each layer.
Intent & Real-World Constraints
- Digital sovereignty vision
- Geopolitical risk tolerance
- NIS2 / DORA / EUCS obligations
- Competitiveness imperatives
- Budget constraints
The Reality of the Systems
- US/China hyperscaler dependencies
- Single points of failure
- Actual network topology
- Tested degraded capacity
- Exposed attack vectors
Six principles for managing complexity without denying it
Continuous Architecture doesn't promise simplicity. It promises legibility within complexity. Here are its six core principles — and their particular resonance in a polycrisis context. Click to expand.
Decisions are dated, contextualized, revisable artifacts. Not truths carved in stone.
Every technical choice is a political choice. Make the trade-offs visible.
Defining what must stay true even as everything changes.
Architecture doesn't sit above delivery. It lives inside delivery.
In a polycrisis, precursors exist. You still have to catch them.
Architecting to preserve the capacity to act in the face of the unexpected.
Maturity radar: are you ready for the unforeseeable?
Readiness for the unforeseeable isn't measured by the number of business continuity plans on file. It's measured in real organizational capacity. These figures are orders of magnitude drawn from field observations — their purpose is to name the discomfort, not to measure it with illusory precision.
14% of companies have built geostrategic monitoring into their architecture practice. 11% have operational geostrategic fitness functions. The rest discover the problem when the crisis hits — and scramble to handle in an emergency what could have been calmly deliberated.
Geopolitical risks × Continuous Architecture levers
| 2026 Risk | Type | Concrete IT impact | Continuous Architecture lever |
|---|---|---|---|
| Cloud Act / Extraterritoriality | Legal | Unconsented access to data by foreign authorities | Sovereignty ADR · Exit strategies · Data-localization fitness function |
| Undersea cable cuts | Infra | Loss of international connectivity, SaaS outages | Degraded-mode tests · Edge computing · Network dependency mapping |
| Targeted state cyberattack | Security | Long-term infiltration, sabotage, data theft | Zero Trust by design · Distributed architecture · Built-in security reviews |
| Semiconductor supply chain disruption | Industrial | Inability to renew physical infrastructure | Critical hardware inventory · Virtualization · Renewal-cycle assumptions |
| Regulatory fragmentation | Compliance | Non-compliance on cross-border flows and localized data | Dated compliance ADRs · Data-localization architecture · Quarterly review |
| Hostile AI & disinformation | Cognitive | Compromised decision-making processes, business deepfakes | AI governance built into architecture · AI-decision traceability · Sovereign models |
| Datacenter energy tensions | Energy | Rising operating costs, partial unavailability | IT energy-footprint mapping · Geographic redundancy · Application-level frugality |
What this means for decision-making bodies
Continuous Architecture only makes sense if it's anchored in a governance that gives it purchase over real decisions. That requires three difficult — and necessary — shifts.
1. Move architecture upward and to the ground, simultaneously
The architect must be present at executive committee meetings to translate geostrategic issues into IT constraints, and present within product teams to make sure implementation choices stay consistent with those constraints. This isn't a job title. It's a distributed function — with relays in every domain, driven by a shared practice.
2. Make architectural decisions reversible by default
In an unstable world, reversibility isn't a luxury — it's an economic requirement. Every infrastructure, platform, or integration choice must be evaluated by its exit cost, not just its entry cost. The cost of reversibility is always lower than the cost of a lock-in discovered too late.
3. Create a mixed geostrategic monitoring unit
General Management, CIO office, Risk Department, Enterprise Architecture. Not to "watch the world," but to maintain a shared reading of the signals that alter the assumptions underlying the IT system's architecture. The key word is mixed: without the technical dimension, monitoring stays abstract. Without the strategic dimension, it stays invisible.
Security vs Performance: building security into the architecture reduces friction without reducing protection — but requires an upfront investment that's hard to justify outside a crisis.
Cost vs Resilience: the real cost of a dependency includes the probability and impact of a rupture. That calculation is rarely made. The strategist-architect makes it visible.
Autonomy vs Coherence: architectural guardrails aren't constraints imposed on product teams — they're the conditions for their real autonomy.
Seven questions for your next executive committee meeting
Continuous Architecture doesn't produce ready-made answers. It produces the right questions — the ones that make decisions possible and accountable.
The architect as political translator of reality
The 2026 polycrisis doesn't ask organizations to become intelligence agencies. It asks them to build an institutional capacity to connect the world's upheavals to their architecture choices — before those upheavals turn into operational emergencies.
Continuous Architecture is that capacity. Not because it predicts the future. Because it keeps alive the dialogue between those who decide and those who build — the dialogue without which any resilience strategy remains a PDF document in a shared folder no one opens when the house is on fire.
And the architect in all this? He's neither the guardian of the technology temple, nor the technical translator of leadership's PowerPoints. He is something harder to sustain over time and more valuable in periods of instability: the professional who moves across the organization's layers so that complexity stays thinkable, decisions stay possible, and the organization stays able to act rightly — even amid the impossible.
The less you depend on a fixed plan, the more capable you are of acting rightly. Complexity isn't managed from a distance. It's crossed — with solid principles, sensitive sensors, and the humility to know that the next surprise won't look like the last one.