HomeCompaniesSubVysion
SubVysion

Autonomous rovers to make the ‘Google Maps’ of underground utilities

We build centimeter-accurate 3D maps of underground utilities to help construction contractors dig safer and faster. With underground intelligence, contractors can avoid dangerous utility strikes and costly project delays.
Active Founders
Maxwell Stone
Maxwell Stone
Founder/CEO
Founder at SubVysion. Previously Machine Learning Engineer at Roboflow, where he built on-device inference SDKs deployed in 10,000+ pharmacies for a Fortune 100 healthcare company and developed ML solutions for enterprise customers including railyard mapping and defect detection. MSE and BSE candidate in Computer Science from UPenn.
Wenson Tang
Wenson Tang
Founder/COO
Co-founder and COO at SubVysion, Inc. Computer Science and Consumer Psychology from UPenn. Previously Software Developer at CentralSquare Technologies, working cross-jurisdiction emergency dispatch allowing first responders to reach emergencies three minutes faster.
Company Launches
SubVysion: Preventing $150B is wasted construction spending
See original launch post

https://youtu.be/5IY3vHEJ7J4

Construction has a mapping problem. Of the $2.2T in annual US construction spending, $150B is wasted due to a lack of subsurface knowledge. All construction starts from the ground up, and more often than not, unexpected assets underneath the surface will hold a construction project back. Whether it's an unknown utility line, unexpected rock layer, or old foundation that wasn’t on any design map, the costs of delays from underground conflicts quickly add up. 

Take utility strikes, when an excavator hits an active utility line. For every dollar spent on the explicit repair cost of that utility line itself, the contractor loses an additional $30 per repair-dollar in project delays, insurance claims, idle work crews, and increased insurance premiums. For a hyperscale project such as data center construction, a single day of delay can cost the cloud-provider up to $14 million dollars for that single day.

In metropolitan areas, maps are incomplete at best and none create a unified picture of critical infrastructure, making excavation especially dangerous when natural gas or electric lines are broken.

Enter SubVysion. We’ve built a fleet of autonomous rovers carrying ground-penetrating radar and other sensors that create a centimeter-accurate map of subsurface infrastructure. Our hardware is the first and the best to map the entirety of US infrastructure. Using our rovers, we comprehensively map construction sites, city streets, and even larger indoor buildings, at 100x better granularity and 10x the speed (with a single rover) compared to our competitors. We process vast amounts of data using machine learning and agents, and output that data into an easy-to-understand 3D scene. 

Who are we? Maxwell Stone was an early ML engineer at Roboflow (YC S20), and Wenson Tang built the world’s smallest and cheapest LiDAR-capable drone for adverse environments. We both studied computer science at the University of Pennsylvania. Together, we’re building SubVysion to map the entirety of America’s subsurface infrastructure.

We’re currently on job sites in Philadelphia and Texas, helping our first market of utility contractors save hundreds of thousands of dollars every year. We’re soon to expand our mapping services to municipalities and energy providers as we scale up our rover fleet and capacity!

If you’re in the energy, infrastructure, or hyperscale construction world, we’d love to hear from you at founders@subvysion.com!

SubVysion
Founded:2026
Batch:Summer 2026
Team Size:2
Status:
Active
Location:San Francisco
Primary Partner:Gustaf Alstromer