CHINA AI
INFRASTRUCTURE
Home / Evidence & methodology
CHINA AI INFRASTRUCTURE · LIVING SYSTEMS RESEARCH

Evidence & methodology

How claims are classified, tested and carried into the research.

02 • RESEARCH PROTOCOL

The thesis is recursive, not sequential.

Each question produces a provisional answer. Later evidence is required to propagate backwards and amend earlier conclusions. We distinguish global AI trends from genuinely China-specific divergence, and architecture from economic proof.

01

Observe

Record primary evidence, company claims, standards, deployments and measurable operating data without forcing a narrative.

02

Classify

Separate observed fact, company claim, external analysis, inference and thesis-to-test.

03

Connect

Map the evidence across architecture layers and companies. Convergence is not automatically causality.

04

Back-propagate

Ask whether new evidence changes earlier questions. Preserve the revision rather than silently overwriting it.

05

Falsify economically

Engineering scale is not enough. Test utilisation, portability, reliability, watts, capex and ultimately RMB per useful token.

05 • EVIDENCE • UNDER THE GRAPH

Evidence is the substrate, not a competing navigation layer.

Evidence now surfaces primarily inside company and relationship drill-downs. The full ledger remains available for audit, but the main experience starts from the system or an interaction and lets the reader inspect why the edge exists, what it means and what remains unknown.

Evidence states: Confirmed • Corroborated • Emerging • Convergence • Historical • Research gap. Contradictory evidence is preserved rather than overwritten.
06 • RECURSIVE AUDIT • Q1–Q10

What survived when architecture met the economics test?

The architecture survives Q10, but the claim becomes narrower. System optimisation can plausibly mitigate component constraints through utilisation, memory hierarchy, workload placement and abstraction. We still do not have independent evidence that the complete Chinese system generally beats a frontier Nvidia/CUDA system on total cost per useful task.

RING 1 • ESTABLISHEDLarge AI demand • component constraints • domestic compute scale • specialised AIDC • growing importance of memory, interconnect, software and power
RING 2 • STRONGLY EMERGINGHeterogeneous compute pools • cloud abstraction • workload-specialised infrastructure • distributed inference • East/West differentiation • national resource orchestration
RING 3 • PLAUSIBLE / UNPROVENEnergy-aware workload scheduling • economically meaningful national compute marketplace • widespread cross-region arbitrage • material AIDC grid-service value
RING 4 • ULTIMATE TESTDoes whole-system optimisation materially close the economic disadvantage created by frontier-component constraints?
BIGGEST REVISION

Not one giant computer. The more plausible architecture is heterogeneous homogeneous islands connected by an increasingly capable control plane.

BIGGEST ELEVATION

Software + orchestration move toward the centre. Physical capacity matters only insofar as workloads can discover, access and productively use it.

STRONGEST FABRIC

Interconnect links almost every thesis layer: SuperPods, external context/storage, distributed inference, geographic scheduling and national resource pooling.

DISCIPLINE

Policy, standards and engineering demonstrations increase viability. Only utilisation, revenue, reliability and economics establish value.

07 • RESEARCH STATUS

Q1–Q10 now form one thesis. Economics remains the governing test.

COMPLETE • Q6
Software & orchestration

Abstraction is real and improving. Full mixed-vendor fungibility is neither established nor necessary; schedulable resource pools are the more plausible target.

COMPLETE • Q7–Q8
Energy & geography

Compute-energy coordination and workload geography are real directions, but dynamic hyperscaler arbitrage remains economically unproven.

COMPLETE • Q9
Demand reality check

Commercial inference, MaaS, agents and production workloads validate the need for the architecture faster than they validate a completed national fabric.

COMPLETE • Q10
Economics stress test

Architecture can improve utilisation and reduce dependence on scarce resources, but system-level cost superiority is not established. The key comparison is total system cost per useful completed task.