INTERVIEW: Darkocean charts AI-led course in offshore survey industry

Zawya

2 days ago

Founded in 2023, marine geo-intelligence firm Darkocean aims to disrupt the offshore survey industry by fusing traditional geoscience with real-time Artificial Intelligence (AI), robotics, and modular deployments.
CEO Anupam Thakur told Zawya Projects that the company was founded out of a desire to challenge the inertia in the offshore survey industry.
'We saw a clear opportunity to offer faster, smarter, and more modular geo-intelligence services - without being weighed down by legacy systems or rigid hierarchies,' he said.
'Darkocean reimagines offshore data acquisition as a streamlined digital service, blending AI-powered analysis, unmanned systems, and modular deployments. This gives our clients rapid insights, reduced environmental footprint, and significant cost savings.'
Thakur said the company employs a unique 'first-pass-right' approach, which combines pre-survey satellite analysis, AI-driven onboard processing, and lightweight modular equipment to cut costs and data turnaround times.
'Our unmanned platforms and intelligent pipelines reduce mobilisation overheads by up to 40 percent and data turnaround times by 60 percent, without compromising on accuracy,' he explained.
Darkocean's client base spans energy majors, NOCs (National Oil Companies), EPC (Engineering, Procurement, Construction) contractors, and sovereign bodies in Qatar, UAE, Saudi Arabia, India, and the UK, with projects spanning offshore wind, subsea infrastructure, and coastal resilience.
A standout project, according to Thakur, was the geotechnical study for a green ammonia project in Oman.
He elaborated: 'Our work on the green ammonia project in Duqm, Oman, involved an integrated geotechnical and oceanographic campaign. We overcame challenging seabed variability by deploying dynamic CPT [Cone Penetration Test] and pressure coring, with post-processing enhanced using AI-driven soil consistency models. The resulting data fed directly into the EPC's foundation design and environmental models.'
Strategic alliances with Malaysia-based Helms Geomarine and UK-based Rovin Subsea strengthen Darkocean's geotechnical and ROV [Remotely Operated Vehicles] capabilities, enabling it to offer bundled solutions worldwide while tapping into partners' decades of expertise.
Thakur said: 'Our alliance with Helms Geomarine gives us access to high-end offshore geotechnical rigs, while the Rovin Subsea partnership expands our reach into ROV operations and inspection services. These collaborations allow us to scale rapidly while maintaining quality control.'
He pointed out that while field surveys remain the revenue mainstay, AI offerings -- especially in bathymetry prediction, real-time QC [Quality Control] and automated log interpretation -- are seeing the fastest growth.
'Clients increasingly value predictive insights over raw data, and that's where our proprietary tech is proving most disruptive,' he said.
Having bootstrapped the business thus far through founder equity and early client revenue, the company is currently raising a $2 million seed round to accelerate AI productisation, team expansion, and robotic equipment upgrades.
'This will be followed by a planned $25 million institutional round in 12–18 months to scale operations and pursue acquisitions,' disclosed Thakur.
He said the company growth strategy includes acquisitions, regional expansion, and a platform-based SaaS [Software as a Service] model.
'We're actively exploring acquisition of distressed marine tech assets, and licensing our ML modules for third-party use. The long-term plan includes building a SaaS layer to turn raw survey data into strategic decision intelligence for asset developers and EPC companies.'
Excerpts from the interview:
How do your AI models, ROVs, geophysical sensors, and survey vessels work together to deliver insights?
Our system is built around interoperability. Survey vessels (manned or unmanned) stream raw data to our cloud AI engines via edge computing units. ROVs collect targeted samples or imagery based on ML [Machine Learning]-detected anomalies. Geophysical sensors trigger automated quality checks. Together, they enable a seamless, looped feedback process from seabed to boardroom.
Darkocean recently introduced its first AI-powered satellite-derived bathymetry (SDB). What sets it apart from existing global solutions?
Our SDB solution integrates proprietary AI correction layers tailored for Arabian Gulf turbidity and seabed reflectance. Unlike global off-the-shelf models, ours is trained on localised sonar and chart data, offering sub-meter vertical accuracy in optimal conditions and enabling rapid reconnaissance for nearshore engineering.
Could you elaborate on how your proprietary ML frameworks enable higher seabed resolution and how you validate its outputs?
Our ML framework uses ensemble learning to reconcile discrepancies across satellite imagery, sonar archives, and in-situ samples. Validation is conducted through cross-checks with multibeam echo sounder data and legacy charting, ensuring our models meet accuracy thresholds suitable for feasibility and reconnaissance-level planning.
What's your approach to data security, timelines, and cost optimisation?
Data is encrypted end-to-end across acquisition, processing, and transfer layers. Our AI pipelines significantly reduce idle vessel time and manual QC loops, allowing us to commit to tighter timelines and leaner budgets. We adopt a cost+ model in many contracts to maintain transparency while remaining competitive.
What's next for Darkocean? What technologies or platforms are you currently incubating?
One of our most strategic initiatives is the development of a proprietary unmanned surface vessel (USV) platform, tailored for energy clients in Northern Europe. This programme is not merely about autonomy—it represents a fundamental reimagining of how analogue geophysics and ultra-high-resolution (UHR) surveys are performed.
We are engineering a 10-metre twin-hull USV capable of 3D ultra-high resolution surveys along with traditional surveys. This eliminates the need for multiple mobilisations or crewed support vessels, drastically reducing environmental footprint and operational costs. The vessel integrates edge AI for onboard QC and real-time processing, and its modular design is built with a clear IP trajectory, covering both hardware and software innovations.
In parallel, we've developed in-house 3D UHR seismic processing capabilities, purpose-built for handling dense datasets acquired in shallow and transition zones. Our workflows combine pinger, chirp, and boomer data to create true 3D imaging—essential for resolving buried objects, shallow gas, and micro-faults with exceptional accuracy. This capability is a game changer for offshore wind farm planning, cable route optimisation, and site investigation in constrained environments.
The USV platform is set to undergo pilot deployments in the North Sea and Baltic region, forming the cornerstone of our expansion into fully automated, AI-powered marine geo-intelligence.
Lastly, how does Darkocean align with net-zero goals in its projects?
Darkocean's unmanned platforms and AI tools reduce vessel time by over 90 percent, directly cutting CO2 emissions. Our bathymetry models aid in siting renewable infrastructure and nature-based solutions. We are also working towards developing ESG dashboards to help clients quantify the carbon footprint of their survey campaigns.
(Reporting by Anoop Menon; Editing by SA Kader)
([e-mail])