From Ohio to Nairobi: Mapping the Uneven Surge of Data Center Growth
From Ohio to Nairobi, data center growth is anything but linear. North America builds today. Europe and Asia ramp next. And leapfrog markets like Kenya or Vietnam post the steepest jumps. The winners? They know where to show up, when to move, and what to bring.
What we asked the platform
We asked Starzdata: “Where are new data centers planned — and how does that stack up against what’s already live?”
What we found: A second fleet is already under way
Today North America carries most of the build, but by 2034 Europe and Asia nearly catch up.
Today’s global fleet: ~42 GW installed
Already in the pipeline: ~84 GW → nearly 2× the live fleet.
By 2034, capacity could reach ~173 GW.
Exhibit 1 — Global fleet tripling by 2034 (Installed, Pipeline, Projected)
How we built this view
This global snapshot comes from 3,500+ datapoints across 125 states, provinces, and countries. Starzdata Smart Queries stitch together broker reports (CBRE, Cushman & Wakefield), public inventories (Cloudscene, DataCenter Map), and company announcements — scored and deduplicated by our confidence layer.
See how capacity shifts in real time — not in quarterly PDFs.
👉 Build this view for your markets → Book a 30-min walkthrough
The story in plain words
Think of data centers as the factories of the internet. One fleet is already live. A second one is on the way — but it won’t land evenly.
North America is pouring concrete today.
Europe & Asia will see more of the build in the 2030s.
Momentum markets (Kenya, Vietnam, New Zealand) look small today, but their plans are several times larger than what’s already live.
Anecdotes at country and state level
🇺🇸 United States — Building at scale
Already the biggest fleet: nearly 20 GW installed. Pipeline adds another 65 GW (≈3.3×). Hot spots include Ohio, Virginia, and Texas.
For suppliers: Cooling, power, and commissioning demand is immediate.
For investors & consultants: Sequencing states by permits and power incentives defines the winning play.
Exhibit 2 — USA fleet expansion (Top states by pipeline)
🇰🇪 Kenya — Small base, big leap
Today’s installed base is just 27 MW. Pipeline: 160 MW — almost 6× bigger. It’s like going from a local clinic to a full hospital overnight.
For suppliers: Anchor deals now can lock future expansion.
For investors & consultants: Treat this as option value — small entry, high upside.
🇻🇳 Vietnam — A fast mover
Installed 221 MW, pipeline 729 MW (~3.3×). Once projects clear permits, capacity can triple quickly.
For suppliers: Package turn-key builds ready to launch when permits release.
For investors & consultants: Focus due diligence on grid interconnection and policy clearance.
🇨🇦 Canada — Different games in each province.
Alberta leads with project volume, Ontario and Québec with incentives, and BC with energy pricing.
For suppliers: Adjust go-to-market by province, not just country.
For investors & consultants: Model provinces separately; lumping Canada hides risk and opportunity.
Exhibit 3 — Canada’s uneven build (Top provinces by pipeline)
Connectivity is the hidden signal
Like airports and hotel hubs, the more routes you have, the more traffic flows through. In data, that means cable landings. Our analysis shows a clear signal: where cables land, data centers follow.
Our platform shows a clear link: more subsea cables usually means more new data centers. The correlation isn’t perfect (~0.37, p-value ~0.009), but it gets sharper at the metro level — proving that cables are a growth signal you can’t ignore.
Exhibit 4 — Connectivity as a growth signal (Pipeline vs. Cable Landings)
🇮🇪 Connectivity in Dublin
One of the densest cable landing hubs in Europe. Its interconnectivity underpins Dublin’s hyperscale role — but power permits are now the main brake.
So what: Connectivity pulls demand, but ESG slows delivery.
🇸🇬 Connectivity in Singapore
Despite land and power limits, its subsea hub status makes it unavoidable for cloud and AI deployments. Demand clusters where cables meet.
So what: Regulation caps volume, but connectivity locks in strategic value.
🇺🇸 Connectivity in Virginia (USA)
Ashburn thrives not just on land and power, but because it sits at the intersection of major backbone and subsea routes — reinforcing Northern Virginia as the world’s largest cluster.
So what: Virginia wins by combining all three signals — power, land, and connectivity.
But cables and latency aren’t the only magnets. Increasingly, what decides whether projects get built isn’t where the cables land — it’s how markets handle sustainability and regulation.
ESG as the new gatekeeper
ESG isn’t killing growth — it’s reshaping it.
Our data shows ESG-linked markets account for over 75 GW of planned capacity — outpacing even AI-driven demand. That means growth is now gated by new rules: PPAs, carbon reporting, and efficiency thresholds.
Other drivers matter too — cheap power keeps the U.S. South and Midwest attractive, AI accelerates builds in Asia, and incentives shape selective opportunities in Canada and Europe.
Exhibit 5 — Market drivers reshaping growth (Installed vs. Pipeline)
Anecdotes at country and state level
🇳🇱 Netherlands — ESG as a cap
Strong connectivity, but pipeline is throttled by policy. Spend shifts to retrofits, heat reuse, and compliance.
🇺🇸 United States — Cheap power + ESG
Virginia and Texas still lead on cost, but every hyperscale deal now comes with renewable PPAs attached. ESG is non-optional.
🇸🇪🇳🇴 Nordics — ESG as an accelerator
Abundant renewables and district heating loops make these markets magnets for sustainable hyperscale builds.
🇸🇬 Singapore — Selective builds
ESG caps volume, but operators win permits by bringing liquid cooling and high-density footprints.
Why it matters
Data centers are doubling — but not evenly. The winners will be those who know:
Where to show up — North America today, Europe and Asia in the 2030s, leapfrog markets if you manage risk.
When to show up — The winners ride the right wave at the right time.
What to bring — Cranes and servers in some markets; optimization tools (PUE, heat reuse, carbon reporting) in others.
How we know
This view is built from 3,500+ datapoints across 125 zones — curated, cross-referenced, and scored by our confidence optimization engine. We merge broker reports, regulatory data, inventories, and operator disclosures into a stitched signal you can act on.
The So What
The next wave isn’t a single tide — it’s a set of uneven surges.
Some driven by cables. Others gated by ESG. Or powered by cheap energy and incentives.
If you can read the signal early, you don’t follow the market — you lead it.
That’s the Starzdata edge: from 3,500 datapoints to the 10 next accounts your team should call.