Executive Summary
The subsea supply base has consolidated into a handful of integrators bundling SPS, SURF, survey, ROV/IMR and diving under one accountability – the Saipem–Subsea7 merger (combined revenue ~€21bn, backlog ~€43bn) being the clearest signal. For operators, the procurement question is no longer which specialist to hire per scope but how deeply to integrate, how to qualify an integrator against mature machinery (IMCA competence, eCMID, IOGP Report 423, Achilles, DNV class) and where to keep capability in-house. We set out the single-integrator versus multi-vendor trade-off and the field-concentration test that decides build-vs-buy for resident ROV/IMR.
What Changed in the Market
For two decades, subsea procurement was tightly compartmentalised. One specialist ran the site survey, another the ROV/IMR work, another the diving spread, and a third the seabed mapping. Each brought its own contract, its own interface and its own contingency, and the operator carried the integration risk between them. That market has contracted. After the 2014–2016 price collapse, the supplier base consolidated into a small number of contractors that combine SPS, SURF, survey, ROV/IMR and diving under a single management line. The clearest signal is the Saipem–Subsea7 merger. The two companies signed a binding merger agreement in July 2025 to form Saipem7, a 50/50 combination headquartered in Milan and listed on the Milan and Oslo exchanges, with combined revenue of about €21 billion and an order backlog of roughly €43 billion. Completion is expected in the second half of 2026, subject to regulatory clearance. When two contractors of this scale combine, the operator’s bargaining position narrows whether it likes it or not. The management problem has therefore shifted from choice to architecture: how far should integration go, how should an integrator be qualified so that its name actually carries weight, and where should the operator keep internal capability so that it is not negotiating without full information? We take each of these in turn, anchored to the standards and contract structures that already exist.
The Case for Integration
The operator-relevant point comes first: an alliance that cannot stand behind a single liability claim is not an integrated contractor – it is two contractors and a handshake. Everything that follows is the evidence for why genuine integration, with single-line accountability, is worth buying. The integrated EPCI (Engineering, Procurement, Construction and Installation) model took shape during the downturn. The logic is simple to state and hard to execute: combining SPS and SURF removes the interface buffers, overlapping contingencies and contractual gaps that sit between separate contractors. The cost evidence is real. TechnipFMC has reported that bringing its own scope into an integrated delivery cut cost by roughly 20% on its own iEPCI scope alone, and the saving deepens further when front-end engineering (FEED) is bundled into the integrated EPCI (iEPCI) scope – the deeper the integrated scope, the larger the saving. On Equinor’s Trestakk field, the collaborative integrated solution reduced development cost by roughly 50% against earlier market bids, which is what made the project viable. The mechanism that delivered this was structural, and it is the reason the liability point matters: Technip and FMC Technologies formed the Forsys Subsea joint venture in March 2015, won their first integrated EPCI award on Statoil’s Trestakk field in November 2016, and began operating as the merged company TechnipFMC on 17 January 2017 – the alliance becoming a single company is precisely what let it stand behind one liability claim. Field architecture reinforced the trend: the downturn pushed development toward subsea tiebacks rather than new platforms, growing the SURF scope and tilting the economics toward integration.
The Market You Are Now Buying From
The integrators are not interchangeable, and the structural differences matter in a tender. TechnipFMC was the original combined SPS+SURF company. Subsea7 delivers integrated SPS and SURF through the Subsea Integration Alliance with OneSubsea, the subsea business of SLB. The Subsea Integration Alliance is a non-incorporated strategic alliance – not a separate company – yet it bids and wins genuine integrated EPCI (SURF+SPS) work, including bp’s Murlach development in the UK North Sea. It clears the single-liability bar not by incorporation but by contract: the alliance presents one integrated EPCI contract to the operator with a defined lead-contractor wrap, so the operator faces a single liability interface and the SPS/SURF split is handled inside the alliance rather than crossing the operator’s desk. That is the mechanism to verify in any alliance bid – which entity holds the prime contract and warrants the integrated scope – because an alliance that cannot point to one is back to two contractors and a handshake. Oceaneering is the ROV/IMR leader rather than a SURF house and runs one of the largest working ROV fleets in the world. Fugro remains an independent geo-data specialist focused on site characterisation and asset integrity. DOF has absorbed Maersk Supply Service into its offshore/subsea fleet, while Helix is positioned in well intervention. Boskalis sits at the opposite end of the do-everything-in-house spectrum: it entered a long-term global partnership with ROVOP to place ROV systems across several of its vessels rather than build and operate that capability itself. A heavy-lift, cable-lay and pipelay contractor making a deliberate build-versus-buy choice on ROV capability is exactly the decision operators should also be making consciously.
Use the Qualification Machinery as a Gate, Not a Checkbox
The machinery for qualifying subsea contractors is mature. The failure mode is treating it as a formality rather than a selection filter, and the centre of gravity is competence. IMCA does not certify individuals; its competence assurance framework requires contractors to operate their own documented systems against IMCA’s minimum standards for safety-critical roles, with separate competence tables for offshore operations, diving, marine survey, and remote systems and ROVs. Your tender should therefore require a documented competence scheme assessed against the relevant IMCA tables – and you should ask for the assessment evidence, not a logo.
Vessel assurance runs in parallel. eCMID (IMCA M 149) is the standard inspection format for vessels of 500 GT and above, including dive support vessels; eMISW (IMCA M 189) covers vessels below 500 GT. Inspections are carried out by IIMS-accredited vessel inspectors who are independent of the vessel’s crew, and reports are loaded into a searchable database. eCMID reflects the contractor’s own safety management system and assesses the vessel’s condition; OCIMF’s OVID is the oil majors’ tool for verifying operational readiness, and most offshore vessels need both. Confirm a valid eCMID is in place before work starts on a contractor vessel, not after mobilisation.
For diving and ROV scopes, specify the codes explicitly. IMCA D 014 is the International Code of Practice for Offshore Diving and requires personnel qualified and competent against IMCA competence guidance with site-specific familiarisation. IMCA R 002 covers entry-level requirements and the introductory course for new ROV personnel; IMCA R 004 is the guidance for safe and efficient ROV operations. On the client side, IOGP Report 423 (HSE management guidelines for working together in a contract environment) provides guidance across selecting contractors, setting expectations, awarding the contract and managing execution, with an early risk assessment determining whose management system governs the work. The accompanying “Contractor HSE capability assessment and scoring system – Supplement to Report 423” provides the actual assessment basis: a capability questionnaire, a scoring system and an audit checklist, rather than a description of a process. Used as a gate, it looks like the table below – a banded score with a defined reject cutoff, not a single pass mark applied after the fact:
| Capability band | Score | Tender decision |
|---|---|---|
| High | ≥ 80% | Proceed to commercial evaluation |
| Medium | 60–79% | Proceed only with a closed-out improvement plan and audit |
| Low | < 60% | Reject – not eligible for award on this scope |
Set the cutoff before bids open and apply it to the assessed score, so a low-banded bidder is rejected on the questionnaire and audit evidence rather than rescued by a competitive price.
Prequalification registration runs through Achilles JQS in Norway and Denmark and through FPAL in the UK and Western Europe (now merged), aligned to NORSOK S-006 and IOGP 423; registration is admission to bid, not a pass. On equipment and class, DNV-OS-E402 is the standard for certifying saturation and surface diving systems, DNV 2.7-3 certifies portable offshore units (including ROV systems) for transport, and DNV 2.22 covers launch and recovery systems. ISO 9001 and ISO 45001, often alongside ISO 14001, are the baseline management-system standards. None of this is new. The discipline is using it to reject deficient bids, not to wave registered bidders through.
Contract Structures and Where the Risk Actually Sits
The headline integrated-contractor instrument is the lump-sum, turnkey EPCI contract, which transfers schedule and budget risk to the contractor while fixing the price. Diving, survey and subsea intervention work is more often billed on a day-rate or reimbursable basis, a model inherited from the drilling sector, in which the contractor carries little exposure once work begins. Most life-of-field work runs under a Master Service Agreement (MSA) with call-off work orders: the MSA holds the terms, warranties, insurance and performance standards, and the work order is one or two pages rather than twenty or thirty. The prevailing liability model is knock-for-knock, with carve-outs for gross negligence and wilful misconduct. The IMR frame agreement is the primary instrument for managing a field across its life, and current awards show the typical shape. DeepOcean’s long-running IMR relationship with Equinor, which began in the mid-2000s, has been extended on a multi-year basis and is supported by a remote operations centre at Killingøy, a shared subsea equipment pool, and the next-generation newbuild IMR vessel Rem Ocean. Oceaneering won a multi-year IMR and ROV contract from bp Mauritania on the Greater Tortue Ahmeyim development, and Vår Energi awarded DeepOcean a five-year IMR frame agreement with options to extend by up to four further years. Long frame agreements buy continuity and a learning curve, but they also deepen supplier dependency – which is the central management tension. The way to keep a frame agreement honest over its life is to tie payment to measured availability of the subsea production system rather than to vessel days or task counts. That means agreeing up front how subsea uptime is defined and, critically, how downtime is attributed: split production-system availability by node (wells and trees, manifolds and jumpers, the SURF/riser system, and topside-tieback interfaces) so that a loss of availability is booked against the node that caused it, and separate planned shutdowns from unplanned failures. Without that attribution rule, every downtime event becomes a dispute over whether the cause sat in the operator’s scope or the contractor’s, and the incentive structure collapses. Settle the attribution method in the frame agreement, not in the first outage.
Single-Integrator Versus Multi-Vendor: Staff for the Model You Choose
The single-integrator EPCI approach dominated large subsea developments in West Africa and the Gulf of Mexico through roughly 2000–2015, and its advantages are real. The operator avoids carrying the EPC project-management load – useful for smaller companies – and can concentrate on field development and commercial work. The disadvantage is concentrated risk: the contractor may take on more technical, operational and commercial scope than it can deliver, the operator may not see it early enough, and the resulting overruns surface as change orders that breach the supposedly fixed budget. The multi-vendor route offers the opposite trade-offs: phased capital release, genuine price competition, flexibility and reduced dependence on any single contractor. The cost is interface risk, because every additional contract is another integration point that becomes a problem if it is not actively managed. That model only suits operators with the in-house experience and resources to own the work-package interfaces. The two control mechanisms that justify its cost are an interface matrix, maintained as a controlled technical document with named responsibilities on both sides of each interface, and a project coordination forum with real escalation authority rather than monthly status meetings. The matrix is not elaborate – each row is one physical or scope interface with a named owner on each side, a deliverable and a date – but it has to be live and controlled:
| Interface | Contractor A (responsible) | Contractor B (accepting) | Deliverable | Due |
|---|---|---|---|---|
| Tree-to-jumper hub | SPS vendor | SURF installer | Hub geometry and make-up envelope | FEED close |
| Spool metrology data | Survey contractor | SURF installer | IMCA S 019 report to agreed TPU basis | Pre-fabrication |
| Tie-in to riser system | SURF installer | Topside/EPC | As-built tolerances and load case | Pre-commissioning |
Every blank cell or unnamed owner in that table is an unmanaged interface and a future change order. The fragmentation is not only operator-driven; specialist contractors are increasingly reluctant to absorb full project risk, so scopes are splitting across the market in any case. The honest conclusion is that the single-integrator model gives a single point of contact and removes interface buffers but carries vendor lock-in and concentrated risk, while the multi-vendor model gives competition and flexibility but demands interface-management overhead the operator must resource. The decisive question is not which is theoretically better, but whether you have the people to manage the interfaces. Choosing multi-vendor to save money and then under-resourcing the interface is a false economy that ends in overruns.
Build-Versus-Buy: Resident ROV and IMR
The sharpest build-versus-buy contest is in resident and autonomous subsea vehicles, where the technology is no longer the constraint. The spectrum runs from tethered resident systems to autonomous docking vehicles: IKM Subsea’s Merlin UCV operates resident at Equinor’s Snorre B on a roughly three-month cycle with onshore piloting over a fibre-optic link; Oceaneering’s E-ROV is a battery-powered, semi-resident vehicle piloted from shore over a 4G link from a surface buoy; Saab Seaeye’s Sabertooth offers autonomous docking; and Saipem’s Hydrone-R, deployed on Equinor’s Njord field under the first worldwide subsea-drone service contract, completed a continuous residency of 167 days under control from Stavanger. Precisely because the capability now exists, operators are tempted into the wrong deployments. Residency only pays on large, concentrated fields. On remote or low-activity fields the vehicle sits idle and the economics collapse on task coverage, not on availability: if a resident vehicle cannot complete even a meaningful share of a field’s tasks and a crewed vessel still has to sail on an unpredictable schedule, you pay for both the dock and the vessel. Many fields need inspection only every couple of years, which is too thin a duty cycle to justify dedicated infrastructure. The test needs a number, but the number is yours to derive rather than borrow. The crossover is where the annualised cost of the resident system – dock capex amortised over its service life, plus residency opex and the through-life data link – falls below the avoided cost of the crewed alternative, which is the vessel and ROV spread day-rate multiplied by the on-station and transit days it would otherwise consume, plus each mobilisation and demobilisation. Build the case from those four inputs (dock capex/opex, the link, crewed vessel day-rate, and mobilisation cost) against your field’s real task volume, and the threshold falls out of your own numbers. As a sanity-check band rather than a target, a dedicated resident system rarely repays itself below the order of several hundred annual operating hours – roughly equivalent to keeping the vehicle productively occupied for a meaningful fraction of the year rather than for a single annual inspection campaign – and the share of the field’s task list that the resident vehicle cannot perform without a crewed vessel should sit well into single-digit percentages, not a quarter or more, because every such task re-introduces the vessel and mobilisation cost the residency was meant to avoid. A single resident dock that genuinely earns its keep is usually one positioned to serve a cluster of fields, not one well-centre; if the business case rests on a single low-activity field, the answer is almost always buy-as-a-service. Treat those figures as the screening band, then confirm with the operator’s own task register rather than the vendor’s utilisation model. Infrastructure economics decides location. Equinor’s standardised subsea docking-station design (developed with Blue Logic) is intended to serve many fields across the Norwegian shelf from a common, multi-station infrastructure rather than from one dock per field. Shared, contractor-managed dock infrastructure beats operator-owned stations almost everywhere, because the utilisation that justifies a dock comes from pooled demand across several fields, not from one operator’s inspection schedule. Build your own docks only where a concentration of assets and a variety of tasks guarantees high utilisation. Otherwise buy it as a service and leave the utilisation risk with the contractor.
Where Operators Get It Wrong
1. Treating Prequalification as a Badge Check
An Achilles-registered contractor with ISO 9001 certification and a list of IMCA references is not automatically qualified for your scope – it is merely eligible to be assessed. A common error is accepting an IMCA competence claim without checking the contractor’s actual competence scheme against the relevant tables for diving, marine survey and remote systems/ROVs. Equally common is accepting a vessel without a valid eCMID and the corresponding OVID record.
2. Buying Single-Integrator Without Funding the Interface It Eliminated
Centralised accountability is workable, but it does not relieve the operator of its technical-authority role. Operators who shed their own subsea and survey engineers under an EPCI contract lose the ability to challenge change orders or to recognise when the contractor has taken on more than it can deliver. You need enough internal capability to audit the integrator, not just to receive its reports.
3. Specifying Embedded Survey and ROV Competence Loosely
When survey/metrology and ROV/IMR work is folded into an EPCI award rather than let to specialists, there is a temptation to let the integrator set its own standards. That raises, not lowers, the operator’s assurance burden. The contract must specify the acceptance criteria for metrology data, positioning standards, the relevant IMCA R-series and survey competence tables, and as-built tolerances; otherwise you accept whatever the integrator deems adequate. Name the references and the numbers. For spoolpiece metrology, require the methodology and reporting of IMCA S 019 (Guidance on subsea spoolpiece metrology) and state the hub-to-hub acceptance budget explicitly – relative-distance tolerances are commonly in the 50–150 mm band and relative-angle tolerances in the 0.5–1.0 degree band, derived from the connector make-up envelope and the spool stress analysis rather than from the survey contractor’s convenience. For the measurement itself, set the expected positioning performance: a modern INS-aided (sparse-)LBL acoustic metrology spread should deliver relative-distance accuracy of the order of 30–50 mm, with INS or photogrammetry where geometry or turbidity rules out clean acoustics. Tie every quoted accuracy to a stated total propagated uncertainty (TPU) basis and confidence level so that an as-built result is accepted against an uncertainty budget, not a bare number.
4. Chasing Residency on the Wrong Field
The case for resident vehicles is too often argued on technological capability and field-development upside rather than on duty cycle. The deciding criteria are field concentration and task coverage, not vehicle autonomy. A routine inspection requirement does not justify a permanently installed vehicle, and any tasks beyond the vehicle’s reach erode the savings.
5. Signing Long Frame Agreements Without Managing Lock-In
Long IMR and service agreements buy continuity but deepen dependency. Without contractual benchmarks, performance-linked incentives and the ability to exit without losing access to assets, a multi-year agreement becomes a liability for the operator and renewal leverage shifts entirely to the contractor.
What We Would Do
- Run prequalification as a selection filter using the IOGP Report 423 supplement: apply its capability questionnaire, scoring system and audit checklist, and reject bids that fall below a defined threshold rather than maintaining a preferred-bidder list until a score is reached.
- Verify vessel assurance and competence evidence before award, not after. Require a valid eCMID (M 149) or eMISW (M 189), the corresponding OVID record, and a documented contractor competence scheme assessed against the relevant IMCA tables. For diving, require DNV-OS-E402 certification of the saturation/surface system and confirmed compliance with IMCA D 014.
- Match the model to the staffing plan. Choose a single integrator only if you are prepared for vendor lock-in and concentrated risk. Choose multi-vendor only if you fund the interface matrix as a controlled document with named responsibilities and a coordination forum with real escalation authority. Do not pick multi-vendor for competition and then underfund the interface.
- Tie the IMR frame to measured subsea availability with reward/penalty bands. Define subsea uptime and its attribution rule before award, not after the first outage: book each loss of availability against the node that caused it (wells and trees, manifolds and jumpers, the SURF/riser system, topside-tieback interfaces), separate planned from unplanned downtime, and link a meaningful share of contract value to the availability the contractor actually controls.
- Run a field-concentration test before any resident-vehicle deployment, against explicit thresholds rather than as a principle. Quantify the annual task volume (a dedicated resident system rarely repays below the order of several hundred operating hours a year), the share of tasks the vehicle cannot perform without a crewed vessel (keep it well into single digits), and whether the dock serves a cluster of fields rather than one. Default to shared, contractor-managed dock infrastructure rather than operator-owned stations.
- Keep technical authority in-house. Retain enough subsea-engineering, geodesy and ROV capability to audit the integrator, scrutinise change orders and hold the line on survey and metrology standards. Integration internalises this expertise into the contractor; it does not remove your obligation to maintain your own.
