____________ ITW 2026 | DAY 1 READ-THROUGH
FROM AI CAPACITY
ambition
TO AI CAPACITY
pROOF.
ITW Day 1 reinforced that AI infrastructure planning is entering a more disciplined execution phase. Announced capacity is not the same as deployable capacity. Four signals from the floor — power, fiber, inference, and forecast discipline — are reshaping how the market will price the next build cycle.
The market is
moving from
ambition to
proof.
"signals from itw point to
constraint-led market
power availability is now the first screen for AI-ready site selection.”
ITW 2026 reinforced that AI infrastructure is no longer defined by projected demand, GPU appetite, or announced data center pipelines alone. The decisive question is becoming whether capacity can be powered, connected, permitted, equipped, and delivered within customer deployment windows.
Power availability is now functioning as the first screen for AI-ready site selection. Fiber readiness is emerging as a parallel gating factor. Inference workloads are moving from a future concept into the 2026–2027 planning window. At the same time, headline forecasts near 100 GW of new global capacity by 2030 remain directionally important — but must be tested against real-world execution constraints.
For executives, the implication is clear: announced capacity is not the same as deployable capacity. The market will increasingly reward operators, developers, utilities, network providers, and capital partners that can prove a credible path to energized and connected AI infrastructure.
WHAT THE FLOOR IS ACTUALLY SAYING.
Each signal is cross-validated against the broader DCS Signals Engine and tagged for strength (emerging · strengthening · recurring · established). Day 1 surfaced four strengthening signals across power, network, workload, and forecast discipline.
01 pOWER BECOMES THE FIRST SCREEN FOR ai cAPACITY
AI infrastructure demand is pushing site selection away from traditional metro-first logic and toward power-first evaluation. For AI-ready capacity — especially medium-sized 10–50 MW blocks — available and actionable power is becoming the critical filter. Interest is shifting to secondary markets with better utility alignment, lower land pressure, or faster paths to energization.
DCS InterpretationPower certainty is becoming a strategic asset. The best-positioned players will be those that can convert power access into commercially usable capacity ahead of customer demand windows.
02 FIBER IS A PARALLEL
BOTTLENECK
Power alone does not make a site AI-ready. ITW commentary reinforced that fiber availability, route diversity, and network lead times are becoming material constraints in their own right. New fiber deployment timelines of 12–36 months create timing risk for projects targeting 2027–2028 service dates.
DCS InterpretationStop treating fiber as a downstream detail. For AI infrastructure, power and fiber must be assessed together as part of the same deployment-readiness screen.
03 INFERENCE PLANNING MOVES INTO THE 2026-2027 WINDOW
The cycle has been dominated by training-led demand: large GPU clusters, high-density racks, hyperscale campuses. ITW signals suggest inference is now moving into the active planning window for 2026–2027 — introducing different requirements around persistent workloads, regional placement, latency, and utilization discipline.
DCS InterpretationWorkload mix is starting to reshape site selection, rack design, network strategy, and regional capacity planning. The next build cycle will be shaped by workload reality, not just GPU demand.
04 GLOBAL FORECAST NEED EXECUTION TESTS
Headline estimates near 100 GW of new global capacity by 2030 indicate the scale of ambition — but should not be treated as automatically deliverable. Forecasts must be reconciled against utility queues, grid constraints, transformer supply, cooling, permitting, fiber, capital discipline, and regional delivery risk.
DCS InterpretationThe headline demand case is strong, but the investable question is deliverability. Treat very large capacity forecasts as directional until validated against physical and regional execution constraints.
what to do with this - by role.
Immediate priorites lifted from the Day 1 read-through, sliced by the three audiences making decisions against ITW signals this week.
+ Re-rank development pipeline by power certainty and energization date
+ Identify which sites have
both power and fiber readiness
+ Separate speculative capacity from commercially actionable capacity
+ Build secondary-market strategies around utility and network feasibility
+ Evaluate whether designs support both training and inference workloads
+ Model inference demand separately from training demand
+ Secure fiber options and route diversity earlier
+ Pressure-test supplier timelines against utility and equipment constraints
+ Evaluate whether regional capacity plans match enterprise inference needs
+ Avoid relying on headline announcements without delivery validation
+ Discount speculative megawatt pipelines with unclear energization paths
+ Prioritize assets with near-term utility commitments
+ Assess exposure to transformer, cooling, and fiber bottlenecks
+ Compare announced growth against actual deliverable capacity
+ Track secondary markets with improving power and network fundamental
Couldn't make all the sessions?
DCS identifies market shifts, emerging constraints, and where infrastructure decisions are moving next so your business can act with greater clarity. Check out Day 01 Instant Signals.
