The Minnesota Department of Transportation has adopted an AI-driven utility intelligence platform to support its statewide road construction programme, embedding remotely collected subsurface data into project workflows across a highway network stretching more than 12,000 miles. The move, announced on 11 May 2026, marks a significant step in how public transportation agencies handle the persistent challenge of unknown or poorly documented underground utilities.
A $1.5 Billion Capital Programme Demands Better Data
MnDOT is managing more than 200 road construction projects in 2026, with a combined budget of approximately $1.5 billion. At that scale, the cost of incomplete utility records is not abstract: on a single transportation project in Minnesota, delays, damages, and relocations linked to unidentified utilities can exceed $1 million. Gathering utility data across a network of this size has traditionally required intensive manual effort, from archival records research to on-site locating, a process that is both slow and prone to gaps.
The agency has integrated the utility intelligence platform developed by Austin, Texas-based 4M Analytics into its existing planning and engineering workflows. The platform aggregates and georeferenced millions of utility records drawn from GIS systems, blueprints, as-built drawings, and permit data, entirely through remote collection methods.
How the Technology Works
Rather than relying solely on physical on-site investigation, 4M’s system applies computer vision to satellite and street-level imagery to detect above-ground utility objects, including manholes, inlets, poles, and hydrants, and then generates subsurface line estimates based on network topology, even where no formal records exist. The result is a continuously updated utility map that project teams can access before any crew reaches the site.
In a pilot phase conducted ahead of the full deployment, 4M’s independently assembled dataset showed more than 60% overlap with MnDOT’s existing records, and the system verified more than 90% of identified utility features through object detection. Over the next twelve months, the platform is set to be implemented across 500 project sites throughout the state.
“Utility intelligence means reducing risk, solving conflicts, and preventing damages based on the most advanced data available for every project site, while significantly lowering the time, costs, and safety hazards associated with collecting that information manually,” said Itzik Malka, CEO and co-founder of 4M Analytics, in the company’s May 2026 press release. “By augmenting standard methods with AI-powered utility data, MnDOT is pioneering a new protocol for public agencies to build and operate roads, safely and cost-effectively.”
A Growing State DOT Client Base
MnDOT is joining a roster of state transportation departments that have adopted the 4M platform, including TxDOT, GDOT, Michigan DOT, and CDOT. Engineering firms operating at national scale, among them WSP, Stantec, AtkinsRéalis, HNTB, and MasTec, are also listed among 4M’s users, giving the platform a footprint across both public agencies and the private contractors that execute their projects.
The platform’s remote-only data collection model is relevant in contexts where mobilising physical locating teams across a large, distributed network is impractical, particularly for early-phase project planning when site access may be limited or premature.
Infrastructure Intelligence as a Sector
The MnDOT deployment reflects a broader trend in US transportation infrastructure management, where agencies are seeking to move from reactive, on-demand data collection toward persistent, continuously available digital records. The Buffalo Department of Public Works recently formalised a comparable step in a different discipline, as Kurrant previously reported, partnering with CYVL to deploy AI-powered LiDAR scanning for pavement and sidewalk condition assessment across its road network. Both cases point to a similar logic: the cost of acting without comprehensive data routinely exceeds the cost of acquiring it upfront.
4M Analytics was founded in 2019 and is backed by Insight Partners, Viola Ventures, and angel investors including Noam Bardin, former CEO of Waze. The company’s most recently disclosed funding round, a Series A extension led by Insight Partners in September 2022, brought its total raised to $45 million at that time.
