Placement becomes a business variable
As models, data and tasks change faster, a deployment location chosen once can lose its advantage. Latency, transfer, power conditions and recovery arrangements should jointly determine placement.
Energy cloud and edge infrastructure.
The next infrastructure advantage is the freedom to change placement.
Growing global data-centre electricity demand brings compute back into a world of physical constraints. For an energy enterprise, the practical judgement is more specific: not every workload belongs at the centre, and field autonomy does not require a complete platform at every site.
A SIGNAL WORTH EXAMINING950 TWh
Central projection for global data-centre electricity demand in 20301
2026 · Global data centres · 2030 projectionTHE ZHICE INTERPRETATION
Compute decisions are also energy decisions.
Expanding compute meets shared constraints in electricity, networks and field operations. ZHICE believes an energy business must compare more than server prices: which tasks stay near their data, which can move, and who maintains service during disconnection will shape reliability and cost. Architecture must connect the business task, the compute resource and the energy context.
THE ARGUMENT BEHIND THE VIEW
A judgement needs a complete argument.
The next infrastructure advantage is the freedom to change placement.
THE BUSINESS LOGIC OF TECHNOLOGY
THE REASONINGThe business conditions behind the view
Connectivity is not continuous
The stronger the cloud capability, the more important it is to explain operations during disconnection. Collection, condition assessment and safety control have different continuity requirements.
Elasticity is not unlimited expansion
Peaks, model updates and data movement can shift cost from processing to networks and operations. We expect visibility and workload portability to become valuable architectural assets.
IMPLICATIONS FOR LEADERS
What should leaders reconsider?
Preserve room to adjust
For fast-changing analytical and inference tasks, favour replaceable interfaces and portable packaging. For stable critical control, respect field operating boundaries first.
Assess cost with reliability
Compare the complete cost per useful task, including recovery, data transfer and field support. An individual resource price is not the cost of an operating solution.
Allocate redundancy by consequence
Assess recovery within the actual business chain. Different consequences of interruption justify different levels of redundancy and local autonomy.
WHEN TO RECONSIDER
Where workloads are stable, connectivity dependable and central capacity sufficient, central deployment may remain the better option. Distribution is a means, not the objective.
Which activities must continue while disconnected, and which tasks can move to improve cost or flexibility?
FROM PERSPECTIVE TO PRACTICEPut this perspective to work.
ExpandCollapse
Put this perspective to work.
Group analytics, field maintenance and production control have different latency, connectivity and reliability needs. A uniform migration policy can conceal these differences, while fully distributed deployment increases duplication and maintenance effort. Start with operating constraints before choosing platform boundaries.
Classify tasks
Map analytics, inference, collection and control workloads, recording sensitivity, disconnect tolerance, recovery needs and interface dependencies; state production-control boundaries separately.
Design operating boundaries
Compare cloud, central and edge options. Specify identity, network segmentation, configuration, observability and failover, then test failure scenarios for critical paths.
Sequence migration and retirement
Pilot a representative non-control service. Define parallel running, rollback and acceptance conditions, and include legacy retirement and operations handover in the same migration plan.
Work products
Workload-placement assessment
Cloud-edge reference architecture and ownership boundaries
Migration waves, recovery validation and retirement plan
What to measure
- Critical-service recovery
- Use exercises to verify recovery-time and recovery-point objectives set through business-impact analysis.
- Unit service cost
- Allocate cost per useful device connection, inference task or business transaction while retaining reliability requirements.
The consulting design does not replace specialist design, testing or approval of production-control systems. Migration remains subject to field safety requirements.
Data notes and research foundationsExpandCollapse
- 1IEA · Key Questions on Energy and AI, 2026
Central projection for all global data centres, not AI alone or a capacity plan for an individual enterprise.
IEA 2026 · Global data centres · 2030 projection - 2IBM Institute for Business Value · 2025 CDO Study
Self-reported confidence across 27 geographies and 19 industries; not an energy-sector adoption rate or observed project success rate.
IBM 2025 · 1,700 senior data and analytics leaders globally - 3DORA · 2024 Report Infographic
The study uses a broad platform definition; associations with productivity, throughput and stability are mixed.
DORA 2024 · Cross-industry software-delivery survey
THE NEXT CONVERSATIONThe question worth discussing:
Contact ZHICE