Policy Briefing · Spherity Research · 2026

AI-Cyber,
Quantum Risk,
Hybrid Warfare
& Physical AI

A threat escalation and operational readiness model for Germany and Europe

Dr. Carsten Stöcker· Spherity GmbH·2026-05-18

Germany and the EU should treat AI-enabled cyber operations, hybrid warfare, post-quantum cryptography, digital identity, Physical AI, and frontier AI as one coupled escalation problem. The next systemic crisis is unlikely to be caused by one isolated technology — it is more likely to arise from interacting failures across digital trust, public administration, finance, industrial systems, critical infrastructure, and security policy.

Central Thesis · Spherity Research 2026
01 · Five-Layer Escalation Model 02 · Six Coupled Threat Domains 03 · Nine Structured Scenarios 04 · EU Law → Operational Controls 05 · 7-Country Maturity Comparison 06 · 12/24/36-Month Roadmap
Scroll to explore
Section 3 · Integrated Threat Escalation Model

Five-Layer Escalation Stack

Not a ladder — a map of escalation paths and feedback loops. An incident can begin at any layer and propagate to others.

5
Frontier Systemic Risk
Quantum disruption · Frontier AI · Autonomous physical systems · Cascading failure
Can political systems govern high-impact capability shifts in time?
4
Geopolitical & Military Escalation
Hybrid warfare · Sabotage · Espionage · Coercion · Cyber conflict · Autonomous weapons
Can the state respond without escalation loss or democratic harm?
3
Critical Infrastructure & Cyber-Physical Systems
Energy · Transport · Telecom · Healthcare · Water · Manufacturing · Robotics · OT
Can essential services continue during attack or failure?
2
Institutional & Economic Systems
Government services · Hospitals · Banks · Universities · Utilities · Insurers
Can institutions function under digital pressure?
1
Individual & Societal Sovereignty
Privacy · Identity · Autonomy · Public opinion · Democratic discourse · Deepfakes
Can citizens and organisations act freely and trust information?
Section 4 · Threat Domain Analysis

Six Threat Domains

All coupled. None standalone. Separate policy treatment is no longer sufficient.

01 — NOW
○
Human Privacy & Digital Sovereignty
The base layer of democratic resilience. Personal data — behavioural, location, biometric, social graph — is a direct input to cyber operations and hybrid influence campaigns.
Active now
02 — NOW
◈
Hybrid Warfare & Cyber Threats
Combines cyber attacks, disinformation, sabotage, economic pressure, and legal ambiguity. Germany is a high-value target. SMEs — 53% of employment — are the exposed supply-chain nodes.
Active now
03 — NOW
⚙
AI-Enabled Cyber Capabilities
AI reduces the skill barrier, shortens recon cycles, improves phishing credibility. Core risk: asymmetric adoption. Attackers move faster than defenders’ procurement and certification cycles.
Active now
04 — MEDIUM TERM
□
Quantum Threat to Trust Infrastructure
Harvest-now-decrypt-later: adversaries collect encrypted data today for future decryption. PQC migration is not an algorithm swap — it is a multi-year migration of protocols, HSMs, identity, OT, and legal evidence systems.
Medium term — migrate now
05 — HORIZON
△
Frontier AI & Prospective Superintelligent AI
Governance lag: AI capabilities shift in months; political institutions act in years. Treated as a horizon-risk category — not a current claim. Near-term: AI-assisted cyber, synthetic media, agentic systems.
Horizon risk
06 — NOW–MEDIUM
⬢
Physical AI & Cyber-Physical Autonomy
Robots, drones, autonomous vehicles, industrial machines, smart grids. Converts digital compromise into physical harm, production loss, safety risk, and insurance exposure. Germany’s industrial base is structurally exposed.
Now to medium
Section 5 · Interdependencies & Amplification Effects

Why Coupling Changes Everything

Threat domains don’t merely coexist — they amplify each other. This is why separate policy tracks are insufficient.

Source domain
Target domain
Amplification effect
Horizon
Confidence
AI-enabled cyber
Cyber-security
Faster vulnerability discovery, exploit support, phishing, and reconnaissance. Changes economics of attack vs. defence.
Now
High
Cyber-security
Hybrid warfare
A real breach or outage becomes material for disinformation, blame, coercion, or public trust erosion.
Now
High
AI
Hybrid warfare
Synthetic media, bots, multilingual propaganda, micro-targeting reduce cost of influence operations.
Now
High
Privacy loss
Cyber & influence
Data brokers, leaks, and profiling improve targeting for phishing, coercion, and manipulation.
Now
High
Quantum
Trust infrastructure
Cryptographic fragility affects identity, certificates, software updates, secure comms, and legal evidence simultaneously.
Medium
Medium
Physical AI
Physical harm
Digital compromise translates to factory shutdowns, vehicle failures, hospital systems, ports, rail, and energy grids.
Now–Med
Med-High
Supply-chain dependency
All domains
Cloud, chips, models, certificate authorities, code repositories create shared failure modes across sectors.
Now
High
Weak state capacity
All domains
Slow procurement, unclear ownership, limited expert capacity, and fragmented federal structures create delayed response.
Now
High

Five Amplification Mechanisms

01
Speed Amplification
AI shortens time from vulnerability discovery to exploit execution. Governance acts in years; attacks move in hours.
02
Scope Amplification
Shared dependencies — cloud, identity, certificates, chips, models — connect institutions. One compromise can affect entire sectors.
03
Trust Erosion
Real incidents become political events. Breaches plus forged documents. Service outages framed as state collapse. The result: uncertainty and delay.
04
Physical Translation
Robotics, OT, connected vehicles, and medical devices convert software failure into safety risk. EU law is beginning to acknowledge this direction.
05
Response-Time Compression
Cryptographic migration, procurement, retrofitting take years. AI capability shifts and adversarial operations move faster. Governance becomes the bottleneck.
Section 5.1 · The Missing Cross-Cutting Layer

Trust Infrastructure

The shared control layer that cyber, AI, quantum, identity, and physical AI policy all depend on — but rarely name together.

Definition

What trust infrastructure is

The technical and organisational layer that allows a digital society to know who acted, with what authority, on which data, through which system, and under which rules.
IdentityAuthenticationAuthorisationDigital SignaturesCertificatesCredentialsRevocationAudit LogsTimestampsSoftware SigningProvenanceEvidence PreservationAI Agent Authority
Design requirements — non-negotiable
Standards-neutralOpen & interoperableIndependently certifiablePrivacy-preservingAnti-lock-inRights-preserving

Without strong trust infrastructure…

The state cannot authenticate public communication
Crisis channels become unverifiable. Forged government messages cannot be distinguished from authentic ones in adversarial conditions.
Companies cannot verify representation rights
Business wallets and agent delegation lack verifiable authority. Supply-chain transactions cannot be audited. Mandate boundaries are unenforceable.
Software updates cannot be trusted
Without signed provenance, an adversary can push malicious updates across entire software supply chains — from public-sector platforms to industrial OT systems.
AI agents cannot be bounded
Agents acting for people, companies, and machines need verifiable scope, expiry, and revocation. Without these, AI delegation creates unlimited liability and attack surfaces.
Logs cannot support accountability
Without tamper-evident audit trails, legal evidence is contestable. Incident attribution collapses. Democratic oversight loses its technical foundation.
Section 7 · Comparative Analysis

Policy Maturity — 7 Jurisdictions

Score 0–5 · Half-step scale · Structured expert judgement based on public evidence · Law exists ≠ operational readiness

Jurisdiction
AI Safety
AI Cyber
Hybrid
PQC
Phys.AI
Frontier
State Ctrl
Rights DigSov
Impl.
Germany
2.5
2.5
3.0
2.5
2.5
2.0
3.0
3.0
2.5
European Union
3.5
3.0
3.5
3.0
3.5
2.5
3.5
4.0
3.0
United States
3.5
4.0
3.5
4.0
3.0
3.0
3.5
3.0
4.0
China
2.5
3.5
4.0
3.0
4.0
2.0
5.0
1.0
4.0
Japan
3.5
3.5
3.5
3.0
3.0
2.5
3.0
3.0
3.5
South Korea
3.5
3.5
3.5
3.5
3.5
2.5
3.5
2.5
3.5
Canada
3.5
3.0
3.0
4.0
2.5
3.0
3.0
3.5
3.0
0 = No recognition1 = Expert debate2 = Strategy3 = Laws + institutions4 = Funded implementation5 = Operational readiness

Germany and the EU are strong in law. The gap is operational integration and tested readiness.

The EU’s rights-preserving digital sovereignty score (4.0) is among the highest globally. The challenge is converting legal commitment into operational readiness: ownership, metrics, exercises, procurement rules, and feedback from real incidents. The EU legal stack (AI Act, NIS2, CRA, DORA, eIDAS 2.0, DSA, Data Act, Machinery Regulation, CER, Cyber Solidarity Act) is necessary but not sufficient without an operational readiness layer on top.
Section 8 · Germany Case Study

Germany: Recognition High, Execution the Gap

Germany has moved from strategic under-recognition to strategic recognition. It has not yet reached integrated readiness.

Cyber-security
Gap
NIS2 transposed late. Implementation differs across sectors and federal levels. No common minimum controls across Bund, Länder, and municipalities.
First action
Mandate common minimum controls and incident exercises across federal, Länder, municipal, and critical-sector interfaces.
PQC Migration
Gap
No visible national critical-sector PQC migration dashboard. Treated as specialist topic, not national resilience programme.
First action
Require cryptographic inventories and prioritised migration plans for government, finance, trusted services, and critical infrastructure.
AI-Enabled Cyber
Gap
Limited public German frontier AI-cyber evaluation capacity. No equivalent to UK AISI or J-AISI in profile and mandate.
First action
Create or assign a German AI-cyber evaluation function with BSI and EU AI Office interfaces.
Hybrid Warfare
Gap
Limited public evidence of integrated exercises combining cyber, finance, media, platforms, and municipalities in one scenario.
First action
Run annual hybrid exercises with synthetic media, service outage, data leak, and financial-confidence modules.
Physical AI
Gap
Limited cyber-physical stress testing at national scale. AI safety, machinery safety, product liability, and OT not planned together.
First action
Start sector stress tests: automotive, machinery, energy, logistics, healthcare, and ports.
Trust Infrastructure
Gap
Fragmented identity, authorisation, revocation, and logging across citizen, company, machine, and AI agent domains.
First action
Publish a standards-neutral trust infrastructure architecture for persons, legal persons, products, machines, services, and AI agents.

“Germany’s problem is not lack of institutions. It is the difficulty of turning many institutions into one operational response model. The National Security Council, BSI, BMDS, IT Planning Council, defence structures, intelligence agencies, BaFin, Bundesnetza­gentur, Länder, municipalities, and EU interfaces can form the basis of an integrated model — but this requires a practical operating system: scenarios, dashboards, rehearsed escalation paths, procurement rules, authority controls, and public communication protocols.”

Table 1 · Key Policy Messages

9 Policy Messages for German & EU Decision-Makers

Each message includes ownership and a first measurable indicator — implementation, not only recognition.

1
The relevant threat unit is no longer a single technology.
Highest risk comes from coupled escalation across AI-cyber, hybrid warfare, cryptographic fragility, identity failure, cyber-physical systems, and social destabilisation. Indicator: one annual cross-domain exercise covering cyber, comms, finance, public administration, and critical infrastructure.
Primary owner
NSC · Federal Chancellery · EU crisis and cyber bodies
2
Trust infrastructure is national security infrastructure.
Identity, authentication, authorisation, certificates, signatures, credentials, revocation, logs, software signing, and provenance must be treated as critical controls for people, organisations, products, machines, and AI agents.
Primary owner
BMI · BMDS · BSI · sector regulators · ENISA
3
Trust infrastructure must be standards-neutral and rights-preserving.
Open standards, interoperability, procurement neutrality, independent certification, privacy-by-design, and clear redress. Avoids both vendor capture and surveillance creep.
Primary owner
BMDS · BSI · Datenschutzkonferenz · EU standardisation bodies
4
Germany has improved strategic coordination, but execution remains the test.
The NSC created in 2025 is a material improvement. It should now be used for cross-domain exercises, implementation dashboards, and escalation planning — not only strategic coordination.
Primary owner
Federal Chancellery · National Security Council
5
PQC migration is not a cryptography project.
Multi-year migration of protocols, certificates, HSMs, identity systems, software supply chains, embedded systems, OT, procurement rules, and legal evidence systems. Indicator: percentage of critical systems with cryptographic inventory and prioritised migration plan.
Primary owner
BSI · BMDS · BMI · BaFin · Bundesnetza­gentur
6
AI-cyber capability is moving from expert tooling to institutional risk.
Public evaluations show advanced AI systems can assist complex cyber tasks. This changes the economics of vulnerability discovery, phishing, triage, and defence. Indicator: coverage of AI-cyber red-team exercises across public administration, finance, critical infrastructure, and key suppliers.
Primary owner
BSI · German AI evaluation function · ENISA · EU AI Office
7
The EU is stronger in law than in operational integration.
The legal stack is broad and real. It does not yet amount to a single readiness architecture. Indicator: EU control map linking each legal instrument to operational controls, reporting duties, exercises, dashboards, and readiness metrics.
Primary owner
Commission · ENISA · EU AI Office · Member States
8
Digital sovereignty must be split into two dimensions.
State and industrial digital control is not the same as rights-preserving digital sovereignty. A European model must strengthen both operational capacity and individual agency under rule-of-law safeguards.
Primary owner
EU institutions · national governments · data protection authorities · parliaments
9
Germany should move from strategies to tested readiness.
Shift priority from recognising threats to measuring resilience: patch time, PQC migration status, authentication coverage, AI red-team coverage, cyber-physical stress tests, supplier assurance, crisis communication exercises. Indicator: national resilience dashboard with public and classified tiers.
Primary owner
NSC · BSI · BMDS · Länder · municipalities · sector regulators
Sections 9.2 & 9.3 · Implementation Roadmap

Germany & EU Roadmap — 12/24/36 Months

Sequenced by dependency — not political aspiration. Ownership and measurable output for each action.

0–12 Months
Foundations
Stand up the architecture before building on it
Establish Integrated Threat Escalation Cell under NSC — mandate, membership, exercise calendar, dashboard scope
Launch national resilience metrics dashboard with public and classified tiers — patch time, MFA coverage, PQC inventory, backup recovery, cyber-physical tests
Start mandatory PQC inventory for government, finance, trusted services, critical infrastructure, and defence-relevant suppliers
Mandate phishing-resistant MFA and rapid revocation for priority public systems
Run first national AI-enabled hybrid crisis exercise with after-action report, assigned owners, deadlines, and budget implications
12–24 Months
Capability Building
Develop the specialised functions that foundations require
Build or assign German AI-cyber and frontier AI evaluation capacity — secure test environment, evaluation protocols, red-team partner network
Expand PQC migration planning to all critical sectors — public administration, finance, trusted services, software-signing systems
Run sector cyber-physical stress tests: automotive, machinery, energy, logistics, healthcare, ports, and rail
Create SME supplier resilience support — cyber-security, identity assurance, incident response, secure software, PQC-readiness
24–36 Months
Integration & Testing
Connect the architecture and prove it works under pressure
Implement AI agent authority controls in high-risk workflows — verifiable authority, scope, expiry, revocation, audit logs, liability mapping
Connect EUDI Wallet, European Business Wallet, machine identity, software signing, and PQC into one trust infrastructure architecture
Publish annual readiness scores and close after-action findings with score changes, evidence notes, and closure rate
Section 2.3 · Boundary Conditions

Security as a Democratic Resilience Function

A broad threat model can justify over-securitisation. These eight safeguards are immovable boundary conditions — not trade-offs.

01
Clear Legal Basis
Security measures must be grounded in law, with defined purpose, scope, authority, and review process.
02
Proportionality
Controls should match the severity and likelihood of the risk. Routine risks do not justify emergency powers.
03
Parliamentary & Judicial Oversight
Powerful monitoring and crisis-response measures require independent checks outside the executive.
04
Transparency Where Possible
Public metrics, architecture principles, and after-action summaries should be published where this does not create operational risk.
05
Privacy-Preserving Design
Identity, wallet, agent-authority, and crisis-communication systems must minimise data exposure and avoid unnecessary centralisation.
06
Redress & Contestability
Citizens and companies need channels to challenge wrong identity, authority, content, risk, or enforcement decisions.
07
Sunset Clauses
Temporary crisis powers should expire unless renewed through a clear democratic process. No permanent emergency states.
08
Open Standards & Procurement Neutrality
Trust infrastructure must avoid vendor lock-in — interoperable, auditable, standards-based components throughout.

“Europe’s strategic advantage is legitimate governance, not only technical control. A European response must be rights-preserving — with legal basis, proportionality, independent oversight, parliamentary scrutiny, and privacy-preserving technology at every layer.”

Outlook · Section 11

Can current institutions detect and contain escalation before it becomes systemic?

“On the public evidence available, Germany and the EU are improving. They are not yet ready enough. The question for policy makers is not whether every scenario will occur. It is whether current institutions can detect and contain escalation before it becomes systemic.”
I
Treat escalation as cross-domain, not sectoral — the threat unit is no longer one technology
II
Build operational readiness, not just law — legal coverage ≠ tested resilience
III
Create tested, measurable trust infrastructure for people, organisations, machines, and AI agents
IV
Run cross-domain exercises annually — scenario, dashboard, after-action, close findings
V
Protect democratic rights as a design requirement, not an afterthought
Read the full paper