The Market

Two Infrastructure Crises. One Structure.

The world is trying to build a trillion-dollar compute industry on top of a transport network that is already failing under its own weight. Umpireal doesn't choose a side of that collision. It sits at the exact point the two crises meet, and turns the collision into the asset.

Demand Side 01

The Compute Crisis.

AI infrastructure spending has become one of the largest capital deployments in economic history — and every gigawatt of it needs land, power, grid connection and water that increasingly doesn't exist where it's needed.

82→220 GW
Global DC demand,
2025 vs. 2030 (McKinsey)
$1.7T
Cumulative DC infra
investment through 2030 (McKinsey)
20–25%
Retail colocation equity IRR
(McKinsey, June 2026)
~50%
Of 2026 global DC capacity
at risk of delay (McKinsey)
McKinsey Global Institute (June 2026) found that as much as 50% of global data centre capacity due online in 2026 could be delayed by permitting, grid connection queues (now averaging over four years, and up to a decade in constrained markets), public opposition, and rising local power demand. Retail colocation — smaller, flexible, closer-to-tenant deployments — earns $200–$380/kW/month and equity IRRs of 20–25%, meaningfully above the $150–$200/kW/month and 13–18% IRR typical of large wholesale hyperscaler leases. The bottleneck isn't capital. It's physical siting and delivery speed — and pricing already rewards the smaller, faster-to-deliver model Umpireal is built around.
Demand Side 02

The Junction Crisis.

While the compute industry searches for land, the world's existing transport network is quietly failing at its most basic structural task: getting conflicting movements through the same space without destroying momentum, money, or lives.

270,000+
Signalised intersections
in the US alone (FHWA)
1.1M+
Global road deaths
per year (WHO, 2025)
~26%
Of those deaths are
pedestrians & cyclists (WHO)
$3.6T
Annual global cost of
crash injuries (CDC)
The WHO puts the cost of road traffic crashes at approximately 3% of GDP in most countries — before a single litre of wasted fuel, hour of lost productivity, or dollar of avoidable municipal redesign is counted. Roughly one in three signalised-intersection incidents in the US is fatal or serious. This is the stock of infrastructure Umpireal was built to replace, not manage: not smarter signals, not another roundabout, but the physical removal of the conflict itself. See the full Momentum Tax breakdown for the complete picture.
Case Study · Ireland
Ireland's data centres consumed 23% of national electricity in 2025, up from 22% in 2024 (CSO), with EirGrid projecting a rise toward 30–32% by the early 2030s. In June 2026, a UN academic body — the Institute for Water, Environment and Health — named Ireland a "live cautionary example" in its report on the environmental cost of AI's energy use, noting that if data centres were a country, they would rank as the 11th most energy-hungry in the world, with a projected 2030 land footprint exceeding 14,000 km² and water use sufficient to meet global drinking needs for 1.6 years. This is the exact strain Umpireal's zero-water, zero-incremental-land model was built to avoid.
Demand Side 03

Every Junction Not Yet Built Is a Junction Not Yet Wasted.

The retrofit opportunity is large. The new-build opportunity, over the long run, may be larger — because it carries none of the cost of removing existing infrastructure first.

~60%→66%
World urban population,
2018 vs. UN 2050 projection
400,000+ km/yr
New road construction,
Asia-Pacific (industry est.)
~2%/yr
Typical regional road
network expansion rate
36.7%
Share of road-infra spend
on new construction, 2025
These new-build figures come from industry market-research estimates rather than a single audited dataset, and we present them as directional, not exact — global road-construction reporting is fragmented and inconsistent between sources. What is not in dispute: two-thirds of humanity will live in cities by 2050, up from just over half today, and every one of those cities is laying new junctions right now. A junction designed as a Continuous Flow Digital Junction™ from day one costs nothing extra to "convert" later — it is simply built correctly the first time.
Where Umpireal Fits

Not Every Junction Qualifies. That's the Point.

Umpireal doesn't claim every junction on earth is addressable. Siting criteria filter the total stock down to a smaller, genuinely viable pipeline — which is what makes the number underneath it defensible rather than aspirational.

Tier Site Type Siting Criteria Delivery Path
1 Existing signalised urban junction High traffic volume · adequate footprint · grid proximity · municipal appetite Retrofit — founding sponsorship & pilot nodes
2 Failed roundabout / bypass junction Documented capacity failure or safety retrofit already scheduled Retrofit — replaces a planned rebuild rather than competing with one
3 New residential / commercial development access junction Planning stage not yet finalised · developer or council co-funding available New-build — specified into the masterplan from day one
4 New arterial / expressway junction, fast-urbanising region National road programme · greenfield · no legacy signal infrastructure to remove New-build — territory-scale, institutional capital raise
The Umpireal Market

Built Bottom-Up From One Node's Economics.

Not the size of the problem. The size of what Umpireal can actually build and operate, at the node economics already proven in our own model.

TAM · Full Global Stock + Flow
Millions of qualifying sites
Every existing signalised junction meeting Tier 1–2 criteria, plus every new junction entering design globally each year (Tier 3–4). At an average €6M CAPEX and €3M–€12M NOI per node, this is a multi-trillion-euro long-run addressable market.
SAM · Serviceable, 10-Year Horizon
9 regions · 100 nodes each
Ireland & UK, EU Core Hub, UAE, India & Broader Asia, Canada, California, Texas, Korea, Japan. 900 nodes at full build-out: €2.7bn–€10.8bn aggregate annual NOI, €67.5bn–€271.8bn aggregate asset value at 25× NOI.
SOM · Obtainable Now
9 founding nodes
One proven node per founding region, funded through sponsorship rather than institutional capital — the evidence base that unlocks each region's full €100M territory raise.
Where This Goes Next

Two Trillion-Dollar Markets. One Physical Answer.

See the engineering behind the node, or the revenue architecture behind the raise.