Data Center Portfolio Planning Execution Lead

Data Center Portfolio Planning & Execution Lead

When to Use

• Build or refresh a multi-site data center capacity roadmap (3–5+ years)
• Prioritize competing DC investments under a fixed capex envelope
• Run portfolio governance: stage gates, RAID, dependencies between DC projects
• Align regional strategy (latency, sovereignty, DR, cloud burst)
• Consolidate business cases from site leads into one portfolio narrative
• Track execution across concurrent builds, expansions, refreshes, and exits
• Prepare steering or board-level DC portfolio status
• Decide build vs colo vs extend vs cloud at portfolio level

When NOT to Use

• Single facility design, MEP, commissioning → data-center-design-execution-lead
• Server/GPU utilization, consolidation, refresh tactics → data-center-compute-supply-efficiency
• Multi-team software program (no DC portfolio) → technical-program-manager
• Corporate strategy, operating model, issue trees → business-consultant
• Terraform, VPC, K8s foundation → infrastructure-engineer
• K8s cluster operations → cluster-deployment-engineer
• Launch messaging only → communication-lead
• MW/rack delivery program and readiness gates → senior-data-center-capacity-delivery-manager

Related skills

Core Workflows

1. Portfolio charter and scope

Define:

• Sites in scope — owned, colo, edge, planned
• Horizon — planning years and refresh cycles
• Constraints — capex ceiling, power at utility, sustainability targets
• Decision rights — steering, stage gates, who owns site DRIs

See `references/portfolio_charter.md`.

2. Capacity and demand rollup

1. Aggregate demand by region (kW, racks, GPUs) from engineering and product 2. Map to supply per site (committed, available, planned) 3. Expose gaps by quarter and region 4. Flag single points of failure (one site, one carrier, one cooling plant)

See `references/capacity_roadmap.md`.

3. Investment prioritization

Score initiatives: strategic fit, risk reduction, NPV/TCO, dependency unlock, time criticality.

Output: ranked backlog, funded / deferred / kill, and tradeoff narrative.

See `references/investment_prioritization.md`.

4. Execution governance

• Stage gates — concept → approved → execute → operate
• Portfolio RAID — risks that span sites (power, lead time, vendor)
• Dependency map — e.g. network build before hall acceptance
• Integrate site-level status from data-center-design-execution-lead DRIs

See `references/execution_governance.md`.

5. Hybrid and exit strategy

• Portfolio rules for cloud burst vs committed metal
• Site exit or consolidation when utilization and TCO favor move
• Contract and migration dependencies

See `references/hybrid_exit_strategy.md`.

6. Steering reporting

Monthly/quarterly: portfolio health (RAG), spend vs plan, milestone slip, top risks, decisions needed.

See `references/steering_reporting.md`.

Output standards

• One-page portfolio summary + detailed initiative register
• Roadmap by region (kW/racks/GPUs) with assumptions stated
• Prioritization scorecard with explicit tradeoffs
• Decision log for steering (what was decided, what was deferred)
• No duplicate site engineering detail — link to site workstreams

When to load references

• Charter and roles → references/portfolio_charter.md
• Multi-site roadmap → references/capacity_roadmap.md
• Scoring and funding → references/investment_prioritization.md
• Gates, RAID, dependencies → references/execution_governance.md
• Cloud, colo, exit → references/hybrid_exit_strategy.md
• Steering pack → references/steering_reporting.md

Capacity Roadmap (Multi-Site)

Rollup structure

Enterprise demand (by region/workload class)
    → Site supply (current + planned)
    → Gap → Initiative (build / expand / cloud / efficiency)

Demand inputs (aggregate)

Supply per site

Gap analysis

Flag:

• Concentration risk — >X% capacity in one metro
• Lead time risk — gap inside GPU or power lead time
• Stranded investment — built capacity with no demand forecast

Roadmap views

1. By region — exec-friendly 2. By initiative — program office 3. By capability — GPU, general compute, storage-heavy

Sync cadence

• Monthly: demand/supply refresh from site and engineering
• Quarterly: rebaseline roadmap and steering decisions

Execution Governance

Stage gates (portfolio)

Align site gates with data-center-design-execution-lead phases—do not duplicate two conflicting gate models.

Portfolio RAID

Track cross-cutting items only; site RAID stays with site DRI.

Dependency map

flowchart LR
  A[Network backbone] --> B[Hall 2 fit-out]
  B --> C[GPU cluster live]
  D[Power upgrade] --> B

Rules:

• One critical path per portfolio release
• Flag circular dependencies before steering

Milestone tracking

Escalation

Hybrid Cloud and Site Exit

Portfolio placement rules

Document default and exceptions:

Rules prevent every team negotiating ad hoc colo vs cloud.

Cloud burst vs committed metal

Portfolio lead arbitrates enterprise view; infrastructure-engineer implements hybrid technical design.

Site exit or consolidation

1. Trigger — lease end, TCO, utilization, disaster risk 2. Inventory — workloads, dependencies, data residency 3. Target — other site or cloud 4. Sequence — migrate, decomm, reclaim contracts 5. Financial — termination fees, write-offs

Coordinate decomm efficiency with data-center-compute-supply-efficiency.

Colo contract portfolio

M&A (light touch)

• Capacity and contract liabilities in target DC footprint
• Integration into portfolio roadmap—not full due diligence alone

Investment Prioritization

Initiative register fields

Scoring dimensions (customize weights)

Use weighted score; document sensitivity if weights disputed.

Funding bands

Tradeoff narrative

Steering needs explicit trades, e.g.:

• "Defer Site B expansion to fund Site A GPU hall"
• "Accept cloud burst cost for 2 quarters vs early build"

Business case rollup

Site leads provide TCO assumptions; portfolio lead:

• Normalizes horizon and discount assumptions
• Surfaces inconsistent utilizations across sites
• Totals portfolio capex and run-rate impact

Deep single-site engineering → data-center-design-execution-lead.

Portfolio Charter

Charter elements

Roles

Delegate site design depth to data-center-design-execution-lead DRIs.

Decision rights

Out of scope (document explicitly)

• Application feature roadmaps
• Cloud account structure (unless tied to DC exit)
• Generic engineering hiring

Handoff from strategy

If business-consultant produced options, translate into funded initiatives with owners and dates—not restate strategy slides.

Steering Reporting

Cadence

One-page portfolio summary

1. Headline — on track / at risk / off track + why 2. Capacity — top gaps by region 3. Spend — capex YTD vs plan 4. Milestones — next 90 days (portfolio-level) 5. Decisions needed — numbered, with recommendation 6. Top 3 risks — mitigation owner

Initiative RAG criteria

Decision log (append each steering)

Metrics dashboard (examples)

• Portfolio capex forecast vs actual
• kW and GPU supply vs demand by region
• Initiatives by stage gate
• PUE / sustainability index (portfolio average)
• Major contract renewals in next 12 mo

Comms quality

• Use communication-lead for external or board-facing narrative polish
• Separate facts from asks — decisions in bullet 5, not buried in status

Anti-patterns

• Site-level MEP detail in steering deck
• Re-prioritizing every month without new data
• No recorded decision when initiatives defer