Mapping the Hidden City with Underground Digital Twins

By Eric Kamande

Cities know where every traffic light is, yet many still guess at what lies three metres below. Gas mains installed in the 1950s, fibre bundles from the 1990s, and stray power conduits often coexist in the same trench, poorly documented and occasionally forgotten. Miss one, and a routine excavation turns into a ruptured line, a blown schedule, and a multimillion‑dollar claim.

The cost of that uncertainty is mounting. According to the Common Ground Alliance, utility strikes in North America alone caused US$30 billion in direct and indirect losses last year. With urban densification and huge broadband build‑outs underway, the margin for error is shrinking fast. That is why investors are suddenly pouring money into a new breed of “underground digital‑twin” companies that promise centimetre‑accurate, multi‑utility maps delivered as frequently as traffic data.

From One‑Off Surveys to Living Subsurface Twins

Traditional ground‑penetrating radar or vacuum potholing captures a single point in time; asset owners still end up stitching PDFs. A digital‑twin approach treats the subsurface like critical infrastructure that needs continuous version control. Multiple sensing modalities; ground radar, magnetometry, acoustic tomography, even passive electromagnetic sniffers, are fused in real time, then aligned against historical as‑builts and LIDAR surface scans. The output is not a static CAD drawing but a cloud model that can be queried by depth, utility type, or confidence score and updated after each dig.

That living model underpins a new workflow: planners run clash‑detection before issuing permits; contractors navigate with augmented‑reality helmets instead of hand‑drawn mark‑outs; utilities reconcile legacy schematics with what is actually in the ground.

Exodigo’s Multi‑Sensor Stack

Tel Aviv‑based Exodigo, led by Co-Founder & CEO Jeremy Suard, exemplifies the shift. Fresh off a US$96 million Series B, the company deploys sled‑mounted arrays that sweep a site in minutes, capturing a terabyte‑scale data set of reflected radio, magnetic, and acoustic signals. Proprietary machine‑learning models, trained on thousands of kilometres of labelled subsurface imagery, classify pipes, cables, and voids with claimed 90 % precision.
What differentiates Exodigo is not one sensor but the fusion layer. Each modality compensates for a rival’s blind spots: radar struggles in saturated clay, magnetometry ignores plastics, acoustics falters near rebar. By weighing all three against historical utility plans, the system produces a probability‑weighted twin that utilities can interrogate via a browser. California’s Department of Transportation says pilot projects cut exploratory potholing by 70 %, translating into double‑digit savings per lane‑mile.

Why the Market Is Ripe

• Gigabit rollouts need trench‑certainty: Fibre operators are racing to lay thousands of route‑miles before rivals secure pole and duct rights. Faster, more reliable clearance data is now a balance‑sheet issue, not an engineering nicety.
• Public owners are rewriting the rule book: The UK’s National Underground Asset Register will require all utilities to supply accurate GIS layers by 2027; similar mandates are brewing in Australia and several US states. Digital‑twin providers slot neatly into that compliance gap.
• Autonomous machinery is coming: Excavators and horizontal‑drill rigs equipped with vision and GNSS can already follow programmed paths. Feeding them a live subsurface map unlocks true “robot‑ready” digging—zero surprises, less rework, safer crews.
• Climate adaptation funds are flowing: Waterfront cities planning seawall upgrades and stormwater overflows need to model how existing pipes interact with new drainage. The same data lake that prevents strikes also supports resilience modelling, opening another budget line.

Market analysts estimate that subsurface mapping and analytics will reach US$4 billion in annual spend by 2030, triple today’s level, driven chiefly by telecom and municipal infrastructure projects. Early adopters report payback periods measured in a single construction season.

Challenges Still to Solve

• Standards: No global schema yet exists for depth accuracy or confidence scoring, complicating tender specifications.
• Data ownership: Contractors ask who holds liability if an algorithm mislabels a gas main. Insurers are only now drafting new coverage language.
• Update cadence: A twin ages the moment a utility crew digs. Automating “as‑built” uploads from job‑site tablets remains a weak link.

Looking forward

As cities race to bury cables, harden grids, and install micro‑tunnels for district heating, the ground beneath our feet is becoming as congested as the sky above. Companies like Exodigo are betting that continuous, multi‑sensor digital twins will become mandatory infrastructure, no different from traffic signal timing plans or above‑ground GIS. The winner will be the platform that not only locates assets but keeps that map evergreen, integrates with permit systems, and feeds autonomous digging equipment without human redraws.

Get innovation insights from The FutureList weekly. Subscribe to our newsletter here