Researched by BEAST Library Curator | Verified by Harper | Quality: 8.3/10
BLUF: National Grid's new digital-twin platform, Triton, just cut UK network-reinforcement planning time by 70% — turning a utility's internal software choice into a de facto timeline variable for anyone building or expanding a large electrified building. Meanwhile PJM, the largest US grid operator, is running its interconnection queue on spreadsheets and case-by-case studies: 220 GW of new generation applied in a single 2026 cycle against roughly 3 GW that actually came online in 2025. If you are scoping a new build, a major electrification retrofit, or a data-center-adjacent expansion, "does this utility run its interconnection assessments on a digital twin or a queue" is now a real site-selection question — and almost no CRE due-diligence checklist asks it yet.
Two Grids, Two Speeds
What Triton Actually Does
Triton, built by National Grid with systems integrator Atos and launched in February 2025 after a two-year build, is a digital replica of the UK transmission network combined with demand forecasts (National Grid). It lets planners run "what happens if we connect this data center / this large load here" scenarios against the live network model instead of commissioning a fresh study each time. National Grid's own number: it cuts the time to analyze options and decide where to reinforce the network by 70% (New Civil Engineer). Verdantix frames the direct application: Triton "can support assessment of future customer connections, such as data centres, by helping to evaluate local and regional network capacity" (Verdantix). That 70% isn't an operations-team vanity metric. For a developer, "time to power" is now sitting alongside land cost, entitlements, and construction schedule as a delivery-risk line item — and it just got materially shorter in one grid, for reasons that have nothing to do with the building itself.
The PJM Counter-Case
Compare that to PJM Interconnection, which covers 13 US states plus DC and is the largest wholesale electricity market in North America. In the first cycle of its reformed interconnection process, 811 new generation projects representing 220 GW applied to connect — against roughly 3 GW that actually reached commercial operation across all of 2025 (Positive Current). PJM's own Northern Virginia corridor — the densest data-center concentration in the world — faces a projected capacity shortfall of up to 15 GW by 2030 as new load additions keep outpacing new generation (EnkiAI). PJM's response so far is a $6.7B transmission-backbone build, not a network-scenario digital twin (EnkiAI). The gap between "the utility can simulate your connection in an afternoon" and "the utility is fielding 220 GW of applications by hand" is now large enough to change project timelines by years, not months.
| Metric | National Grid (UK, w/ Triton) | PJM (US, no equivalent tool) |
|---|---|---|
| Connection-assessment method | Live digital twin + demand model, scenario runs | Case-by-case interconnection studies, sequential queue |
| Planning-time reduction cited | 70% faster network reinforcement decisions | No comparable figure published |
| 2025-26 queue pressure | ~100 new data-center sites expected by 2030 | 220 GW applied vs. ~3 GW delivered in 2025 |
| Projected shortfall | Grid-constrained, actively modeled | Up to 15 GW by 2030 (Northern Virginia corridor) |
Sources: National Grid, New Civil Engineer, Verdantix, EnkiAI, Positive Current — see links above and below.
Why This Belongs on Your Site-Selection Checklist
Every CRE due-diligence checklist already screens for flood zone, seismic risk, and title. Almost none screen for "how does this utility model new connections." That's a blind spot, because the digital-twin-vs-queue gap changes three things a developer actually prices:
The Three Questions to Ask Your Utility (or Your Broker) Before You Commit
- Does the utility have a network-level digital twin, and does it cover my substation? Not every feeder in a Triton-covered territory benefits equally — ask specifically about the substation or zone your site sits on, not the utility's headline capability.
- What's the delta between "study requested" and "study delivered" today, not in the utility's press release? A digital twin only helps if it's actually shortening the queue you're standing in, not just the demo queue.
- Is my load large enough to trigger a full interconnection study regardless of the tooling? Loads above the threshold that forces a full system-impact study (in the US, typically tied to state large-load rules; in Taiwan, the 5,000 kW / 5MW contract-capacity line — see our companion piece on that threshold) don't get to skip the queue just because the utility owns better software.
For more on how sensing and grid-facing infrastructure decisions cascade into building operations, browse our full CRE intelligence library or the digital twin archive.
Taiwan's Version of the Same Problem
Taipower has stated it plans to "integrate cutting-edge technologies such as digital twins, 5G, and artificial intelligence to continuously strengthen the flexibility and security of the power grid" as part of its net-zero and smart-grid roadmap (Taipower Sustainability). That's a stated direction, not a shipped Triton-equivalent — Taipower's public materials describe smart-grid buildout and DER integration tooling, but no published "time to power" reduction figure comparable to National Grid's 70%. In the meantime, the more immediate Taiwan constraint for large buildings is regulatory, not modeling: Taipower has suspended new applications above 5MW north of Taoyuan since September 2023 to protect grid stability, and any facility at or above the 5,000 kW contract-capacity threshold now carries a separate renewable-obligation clock on top of that connection constraint. If you're scoping a build in that corridor, the connection risk isn't "will the utility's software be fast" — it's "is the application window even open," which is the harder, blunter version of the same underlying story: grid capacity, not the building itself, is increasingly the limiting factor on your delivery schedule.
What I'd Do If This Were My Building
I'd stop treating "utility interconnection" as a line item my MEP consultant handles quietly in the background and start treating it as a Go/No-Go gate that gets evaluated at the same stage as land cost — because in a growing number of markets, it now has more schedule risk than the building's own construction timeline. Concretely: before signing a letter of intent on a site with meaningful new electrical load (large-format retrofit, EV fleet charging, any AI-HVAC or electrification pilot pulling material new capacity), ask the utility directly whether connection assessments run on a digital-twin/scenario tool or a manual study queue, and get the current study-to-delivery timeline in writing, not from a press release. In a Triton-covered UK feeder, that answer might shave months off your schedule. In a PJM or Taipower-constrained corridor, it might be the reason you walk from an otherwise-good site.
Have a question about this topic? Ask our CRE AI Agent →