Subsea hyperspectral intelligence

See the plume.
In real time.

EV Deep is developing a deep-rated hyperspectral imaging system with on-board AI that detects, measures and reports deep-sea mining sediment plumes as they happen, not months later from a laboratory.

6,000 mDepth rating design target
25+ yrsBuilding cameras for extreme subsea and downhole environments
LivePlume data in minutes, not the weeks to months of lab sampling
24/7Built for continuous production operations
The challenge

Deep-sea mining will run around the clock. Plume monitoring can't wait for the lab.

Every seabed collector raises a sediment plume. Regulators, scientists and the public want to know where it goes and how dense it is. Under the US deep seabed mining rule in force since 21 January 2026, applicants must describe their environmental monitoring equipment and file a monitoring plan.

Today, plume data is gathered by ROV sampling, recovered to the ship, sent ashore and analysed in a laboratory. Results arrive weeks or months later. The first commercial-scale operations are expected within the next two years, running continuously. They will need continuous plume data.

Today

Weeks to months
  1. ROV deployment
  2. Sample capture
  3. Recovery
  4. Shipping
  5. Lab analysis
  6. Report

With EV Deep

Real time
  1. ROV deployment
  2. Live capture
  3. On-board AI
  4. Live report

Continuous, repeatable measurement replaces periodic sampling, giving operators and regulators a live picture of the plume.

The platform

One integrated system: sensor, AI and reporting.

EV Deep combines a deep-rated hyperspectral imager, edge AI that runs on the vehicle, and automated reporting in a single purpose-built system. It is designed around how mining operators actually work: offshore, continuously and with limited bandwidth to shore.

01

Detect

Identifies the collector plume by its spectral signature and separates it from clear water, seabed and natural particles.

02

Measure

Estimates plume density and spread continuously, building a live picture around the operation.

03

Report

Sends compact, reporting-ready results from ship to shore over limited satellite links.

04

Protect

Illumination designed around the visual sensitivity of deep-sea life, to reduce disturbance while imaging.

System architecture, methods and performance data are shared with qualified partners under NDA.
Illustration of a hyperspectral data cube, with many stacked colour bands and one layer showing a sediment plume
Illustrative hyperspectral data cube
Why hyperspectral

A normal camera sees three colours. Hyperspectral sees a fingerprint.

A standard camera records red, green and blue. A hyperspectral imager splits light into many narrow bands, so every pixel carries a full spectrum. That spectrum works like a fingerprint, telling materials apart where the eye and a normal camera cannot.

Published research has already shown hyperspectral imaging can classify materials on the seafloor at more than 4,000 m. EV Deep takes the technique to the plume itself and adds AI and real-time delivery.

3bands in a standard colour camera
Manynarrow bands per pixel with hyperspectral
Continuous mining needs continuous evidence.
EV Deep
Track record

Built by people who have already put cameras where others couldn't.

EV Deep is led by Jonathan Thursby, who spent 25 years as founder and CTO of EV Offshore (EVCam), developing camera systems for deep-sea and downhole environments of up to 10,000 m, 175°C and 15,000 psi.

  • 1999Founded EV Offshore in Norwich
  • 2011First colour downhole camera
  • 2014Queen's Award; £69m management buyout led by Dunedin
  • 2024EV sold to ADNOC-backed Enersol, with operations in 36 countries
  • TodayHyperspectral imaging and AI in daily use across the EV group, with university research partners
Extreme hardware

Pressure and temperature

Decades of designing and deploying optical systems that survive crushing pressure, heat and darkness.

Machine learning

Feature detection

Neural networks trained to find and measure features in subsea and downhole imagery automatically.

Hyperspectral

Spectral analysis

Hyperspectral and multi-sensor imaging with AI, in active development across the group today.

3D and motion

Measurement

Time-of-flight 3D imaging, camera motion tracking and image stitching from a single camera.

Data delivery

Secure cloud viewer

Browser-based platform handling terabytes of video and sensor data, secured and always available.

Offshore

Operational insight

Working with offshore specialists who have first-hand experience of deep-sea collection trials.

Roadmap

From feasibility to the first at-sea trial.

The programme is timed to the start of commercial operations, so a validated system is ready when operators need it.

Feasibility

Define what must be measured, how often and how accurately, and benchmark against current sampling methods.

Lab prototype

Build a bench system and a test tank with a recreated seabed and controlled sediment plumes.

Independent validation

A marine science partner verifies the results and the protection of deep-sea life.

At-sea trial

Deployment with an operator at sea, leading to production systems and data services.

Invest & partner

We are raising development funding now.

EV Deep is seeking investors and operator partners to take the system from feasibility to at-sea trial. If you are an investor, a mineral operator or a marine science organisation, we would like to talk.

Investor materials are confidential and shared under NDA. Nothing on this website is an offer or invitation to buy or subscribe for securities.

  • Regulation-driven demand. Operators must show how they will monitor the plume, and production runs around the clock.
  • Clear timing. The first commercial operations are expected within two years.
  • Few specialists. Plume monitoring is not yet a commoditised market, leaving room for one system that combines sensor, AI and reporting.
  • Proven builders. A team that has designed, sold and scaled extreme-environment imaging worldwide.
  • Recurring revenue. Systems plus ongoing monitoring data and reporting services.