Subsea well intervention — Gulf of Mexico supply-constraint brief

Takeaway

Supply-constrained. The deepest serviceable band (5,000-10,000 ft) holds the most subsea wells on record (270 wells, 55.7% subsea share) yet sits over the thinnest dedicated fleet: only ~3 GoM-resident dedicated intervention units [soft]. The wells are concentrated exactly where the intervention-asset supply is scarcest.

Existing subsea wells by water-depth band

Counts are from the BSEE subsea-borehole registry [on record]; the subsea-share column depends on full-population coverage and is [soft].

Subsea total on record: 593 wells [on record] (depth min 1,055 ft / median 4,609 ft / max 9,627 ft; n=593). Well population compared against: 29,286.

Serviceability — who can reach each well

The rule: Depth does not pick the asset; riser-need picks the asset. Water depth alone does not choose the vessel; whether the scope needs an intervention/workover riser does. These are [engineering judgement] from standard subsea-intervention practice, not a measured BSEE dataset.

Modifiers: A horizontal subsea tree (HXT) tubing pull forces a riser-based asset even for otherwise-light scope. Light riserless work in 5,000-10,000 ft is availability_constrained (thin supply of rated units), not capability-limited.

Asset classes: rlwi_monohull = Riserless light well intervention monohull (vessel_fleet #592); mpsv = Multipurpose support vessel (vessel_fleet #592); heavy_intervention_semi = Heavy-intervention semi, e.g. Helix Q-class (vessel_fleet #592); modu = Mobile offshore drilling unit (drillship / drilling semi).

Intervention-asset fleet

GoM-resident dedicated intervention units (the binding supply)

~3 units [soft] are GoM-resident and dedicated to well intervention. This — not any global roster — is the figure that constrains GoM well access:

Triangulated against independent research:

• gom resident dedicated intervention vessels: 2-4 [soft] — GoM-resident dedicated intervention units; small population
• global rlwi fleet: 15-25 [soft] — global riserless light well-intervention fleet, +/-30-50%
• helix purpose built units: 7-8 [soft] — Helix purpose-built well-intervention units worldwide
• rigs rated 20k psi worldwide: 2 [on record] — only two rigs worldwide rated to 20,000 psi

Global available roster (CONTEXT ONLY — NOT GoM supply)

These are WORLDWIDE fleet counts. They include shelf and non-GoM assets and must NOT be read as Gulf-of-Mexico intervention supply. They show the size of the global pool a GoM operator competes for, nothing more. Counts are DEDUPED distinct hulls (#599): 2,442 source listings collapse to 2,336 vessels (106 duplicates removed, e.g. the Helix Q-class name-spelling variants), so no hull is double-counted.

Indicative day-rates are an attached PUBLIC snapshot (#596), as of 2026-02-19 — issuer FSRs / filings / dated trade press, NOT a live subscription feed. Helix Q-class intervention semis and RLWI monohulls carry no public per-day figure (shown as dayrate not public).

Planned / projected new subsea wells

The on-record FID register itemises 18 announced US GoM developments totalling 72 new subsea wells [on record] — a FLOOR, since later infill phases are deliberately not itemised. Trion is excluded (Mexican waters).

New wells by first-oil year (on-record floor):

Analyst envelope: 12-20 new GoM subsea trees / wells installed per year (central 16, +/-15%) [projected], through 2030. This brackets the on-record floor (P10/P90); it is not summed with it. Sources: Westwood Subsea Tree Tracker (GoM ~16-19 trees/yr); Rystad Energy subsea market outlook.

Access gap — forward-looking demand vs supply

FORWARD-LOOKING ACCESS RISK, not a measured crossover. Per band: intervention demand rig-days/yr (subsea wells x parameterized frequency x band-effective heavy-intervention duration) vs available fleet rig-days/yr (eligible fleet x utilization x 365). A utilization_ratio > 1 means forward demand would outstrip that supply pool; it is an access-risk indicator, not an observed shortfall.

5,000-10,000 ft holds 270 subsea wells but only 2 GoM-resident heavy-intervention unit(s): forward demand would need ~4.4x the rig-days that resident supply can provide (access-RISK indicator, NOT a measured crossover).

Per band: intervention DEMAND rig-days/yr (subsea wells x intervention frequency ~0.1-0.2/well/yr [engineering judgement] x band-effective heavy-intervention duration, #629) vs available SUPPLY rig-days/yr (eligible fleet x utilization ~0.6-0.8 [engineering judgement] x 365). The GoM-RESIDENT fleet is the binding constraint; the global roster is CONTEXT ONLY. utilization ratio = demand / supply (>1x = forward demand outstrips that supply pool). All figures are [forward-looking risk] — an access-RISK indicator, NOT a measured crossover.

Global-roster context (NOT GoM supply): against the worldwide heavy-deepwater pool (~471 eligible units) the 5,000-10,000 ft demand ratio is only ~0.0x — the global pool is ample; the constraint is that almost none of it is GoM-RESIDENT. This is the access wall.

Forward-looking caveats carried with the synthesis:

• FORWARD-LOOKING: this is access RISK, not a measured demand/supply crossover; both parameters (frequency, utilization) are cited planning assumptions, overridable.
• FLOOR: the #630 empirical interventions_per_well cross-check is a lifetime upper proxy on the ~6% depth-stamped WAR subset (WATER_DEPTH populated on only ~6% of WAR rows, #627/#630), not an annual lambda.
• GEOGRAPHY (#628): GoM has ~0 RESIDENT dedicated LIGHT-intervention vessels -- this is geography (the global RLWI fleet is elsewhere) + the 6% WAR coverage, NOT evidence that deepwater light intervention is avoided. ~3 GoM-resident dedicated units total (#598/#628).
• UNKNOWN 44.5% (#627): a large share of WAR service-type is unclassified, so the empirical work mix is incomplete.
• Demand uses the HEAVY heavy_dead_well duration for all interventions (an upper bound on rig-days); real campaigns mix lighter scopes.
• Global-roster supply is a worldwide pool a GoM operator competes for; the GoM-resident view is the binding constraint.
• Subsea well counts are [on record] and likely undercount older completions (#583/#589).

Honest-framing notes

• Subsea band counts are [on record] from the BSEE borehole registry, which likely undercounts older completions; subsea-share is [soft].
• Serviceability is [engineering judgement], not a measured dataset; band-level availability should be calibrated against the fleet.
• The global by-asset-class fleet counts are CONTEXT only. The binding GoM intervention supply is ~3 resident dedicated units (~2-4 [soft]).
• The planned-wells total is an on-record FLOOR; the analyst rate is [projected] and brackets — never adds to — that floor.
• The access gap is FORWARD-LOOKING access RISK, not a measured crossover; intervention frequency and fleet utilization are cited, overridable planning parameters, and intervention cost now varies by depth band (#629 mobilization fix).

Sources

All figures render from these committed source YAMLs:

• well_inventory_by_band.yml — installed subsea wells by band (worldenergydata #583).
• serviceability_matrix.yml — band x scope serviceability + riser rule (worldenergydata #586).
• planned_subsea_wells.yml — on-record FID register + analyst rate (worldenergydata #587).
• intervention_fleet_catalog.yml — unified intervention fleet (worldenergydata #598).
• access_gap.yml — forward-looking demand vs supply rig-days + $ exposure by band (worldenergydata #638, capstone of #626); economics are now band-aware (#629) and demand cross-checks #630.