Overview
The 30-second pitch. Pick a persona below for the deep dive.
State of the Industry
The US data center buildout has hit a power wall — demand is racing ahead of the grid.
Hyperscalers are committing record capital to AI infrastructure, but electricity, transformers and interconnection queues now set the pace. This monitor tracks where capacity is going, what it costs, and what is holding it back.
Dive deeper:
Capital
Capex magnitudes, vacancy, $/MW returns, M&A heatmap. For investors.
Buildout
Interconnection queues, lead times, cost stack, site selection. For builders.
Grid
Demand vs generation, rate impact, regulation, utility actions. For energy + policy.
Tokens
Industry token volume, $/token compression, tokens × energy bridge.
Capital
Where the money goes — capex, vacancy, returns, M&A. For capital allocators, VCs, corp dev, public-market investors.
~$808B of capex is chasing a supply-constrained market. The bottleneck is power, not demand — winners will own the megawatts.
2026 capex by operator
Six largest AI infrastructure operators · USD billions · ~75% AI-attributable
What a megawatt costs
Illustrative $M to build one MW of AI capacity · GPUs dominate
M&A and major deals
Deals shaping the buildout · past 30-60 days
Public-market plays
Listed companies grouped by thesis · not investment advice
Vacancy & pricing power
Primary-market vacancy · year-end · the supply-tightness story
Vacancy compressed from ~9% in 2019 to 1.4% today — landlord pricing power has shifted decisively. Asking rents have roughly doubled since 2022.
Build your IRR scenario
100 MW build · adjust assumptions · IRR recomputes live
Project IRR
—
Tokens →
$/token compression compresses inference revenue per token. Open the unit-economics deep dive →
Buildout
How capacity actually gets built — queues, lead times, cost stack, site selection. For infra leaders and supply-chain planners.
The binding constraint isn't compute — it's transformers, queues and lead times. Where you build matters more than how big you build.
Interconnection queue → construction → operational
US data center grid pipeline · GW · LBNL Queued Up 2025 + CBRE + Goldman Sachs
Where the capacity sits — primary US markets
Bubble size = operational GW · color by status · hover for detail
US Primary Markets
Power-constrained
Major growth
Emerging
Time to deployment by region
Years from greenfield to energized · ranked by lead time
Equipment lead times
Months from order to delivery · 2026 industry estimates
Where every $1M goes
Build cost breakdown per MW · facility vs compute layers
Self-built capacity: operational vs pipeline
Top 5 hyperscalers · MW · modeled from IR + analyst sources
Site-selection scorecard
Composite scores across top US markets · 1-10 scale (higher = better for builds)
AI infrastructure stack — where the binding constraint sits
Each layer of the buildout has its own choke point. As one loosens the constraint cascades upstream. Status reflects current public commentary; every row carries a citation.
Upstream substrate bottlenecks
Supply concentration beneath the transceiver layer
Optical interconnect roadmap
Bandwidth generations + CPO transition
Tokens →
Tokens per MW-hour. Energy efficiency curves for inference. Open the energy bridge →
Grid
The energy + policy story — demand vs generation, rate impact, regulation, utility responses. For utility planners, regulators, energy reporters, policy analysts.
If demand rises another 60 GW by 2030 without commensurate grid expansion, six states face >25% retail rate hikes — and AI buildout becomes politically untenable.
Annual demand growth vs. new firm generation
DC load added per year vs. firm generation committed to DC · GW · 2024-2030
Projected retail rate increase by 2030
Residential rate impact under high-DC-load scenarios · vs. 2024
Transmission constraints by ISO
Large-load interconnection queue depth, by region
Regulatory tracker
State + federal actions shaping the buildout
Utility M&A + capacity additions
Who's adding what, for whom
Demand response & flex load
The grid's pressure-release valve
Tokens →
Tokens × energy: how much grid load does AI inference actually create per unit of usage? Open the adoption-vs-energy curve →
Tokens
The AI economy underneath the buildout — how much is run, what it costs, what it consumes.
At the largest platforms, total tokens processed have scaled ~330× in 24 months (Google: 9.7 T → 3,200 T per month, May 2024 → May 2026), while $/token has fallen ~100×. The AI economy runs on this curve.
Industry token volume by provider
Trillions of externally-billed API tokens per quarter · stacked area · modeled from provider revenue + Epoch AI + SemiAnalysis. Does not include in-product AI inference (Search AI Overviews, YouTube, Workspace) — see Disclosed totals below for the bigger picture.
Disclosed total tokens processed — the bigger picture
When platforms disclose total inference across all surfaces (API + in-product), the numbers are materially larger than monetized API volume alone.
$/token compression — the 100× price drop
$ per 1M output tokens at launch · log Y axis (data spans ~125× from frontier to OSS-hosted — linear would hide the cheap end) · lower-right = newer + cheaper
Tokens × energy bridge — from AI economy to grid load
The link that ties this tab to the rest of the dashboard
Inference vs training split
Where the compute actually goes · late-2025 industry estimates