How deep a commercial diver can actually go

Four different answers to one question

Ask how deep a commercial diver can go and you will get four numbers back, usually with no label attached. They describe four different things, and conflating them is the single biggest reason the internet is wrong about this topic.

The operational depth is where paid work actually happens. For the great majority of saturation jobs that is somewhere between about 60 and 180 metres. It is set by where the hardware is, not by what a human can survive.

The certification and regulatory limit is what your ticket and the law in that jurisdiction permit. These are hard numbers and they are mostly shallow: 50 m on air in the UK, 30 m on scuba under the Australian scheme, 190 fsw on surface-supplied air in US federal rules.

The rated limit of the equipment is what the saturation system on the vessel is certified for. This is the 300 metre number, and it is remarkably uniform across the world fleet.

The experimental record is what has been done once, under laboratory conditions, with gases and compression schedules nobody uses commercially. That is where 534 m and 701 m live.

A diver who tells you they have been to 300 m is probably describing their system's certificate. A diver who tells you 150 m is probably describing their deepest actual bell run. Both can be telling the truth.

Where the law stops you: the regulatory limit

There is no global depth limit for commercial diving. What exists is a patchwork of national rules, and the striking thing about them is how little they say about deep diving and how much they say about shallow diving.

In Great Britain, the governing document is the Health and Safety Executive's Approved Code of Practice L103, second edition, published in 2014 and in force since 8 December 2014. Paragraph 43 sets the maximum depth for compressed air or oxygen-nitrogen mixtures at 50 metres of water, and recommends a maximum oxygen partial pressure of 1.4 bar. Paragraph 46 is one sentence long and says closed bell techniques should be used deeper than 50 metres. And that is the last depth figure in the document. The UK sets no upper limit at all. Below 50 m you move to a closed bell and the regulation switches from numbers to process: competent people, assessed risk, a system certified for the job.

The United States is the opposite. US federal rules are explicit about shallow depths and silent about the bottom. Under OSHA's commercial diving standard, scuba is prohibited deeper than 130 fsw (29 CFR 1910.424(b)(1)), and a decompression chamber must be ready for use for any dive deeper than 100 fsw or outside the no-decompression limits. Surface-supplied air diving is capped at 190 fsw, with a concession to 220 fsw for bottom times of 30 minutes or less (1910.425(b)(1)). For mixed gas, a bell is required deeper than 220 fsw and a closed bell deeper than 300 fsw (1910.426(b)(1)(i) and (ii)). The US Coast Guard's parallel rules for diving from vessels, 46 CFR Part 197 Subpart B, repeat the same figures: 130 fsw for scuba, 190/220 fsw for surface-supplied air, closed bell beyond 300 fsw.

Note what is missing. Neither OSHA nor the Coast Guard names a maximum depth for commercial diving. The deepest-specific provision in the OSHA standard is 1910.423(c)(2), which requires that for dives deeper than 300 fsw a chamber capable of recompressing the diver to the maximum depth of the dive be available on site. That is a capability requirement, not a ceiling.

Norway is the most interesting case, because the Norwegian ceiling is real but it is not written as a law. The framework regulations bring manned underwater operations in petroleum activity under the Working Environment Act, but the detailed requirements are delegated to the NORSOK U-100 standard. A 2024 review commissioned by the Norwegian regulator and carried out by NUI AS in Bergen looked at exactly this and reported that it could not identify a statutory Norwegian depth limit. What it found instead was a standard built out to a specific depth and no further: NORSOK's compression rate of 1 metre per minute is specified for the range 0 to 180 msw, and its excursion window runs from 14 to 193 msw. There is no NORSOK schedule for 250 m because NORSOK was never written for 250 m. The review itself was titled for the range 0 to 250 msw, which tells you where the regulator thinks the open question sits.

Sources: HSE L103, 2nd edition 2014; eCFR 29 CFR 1910 Subpart T; eCFR 46 CFR 197 Subpart B; ADCI Consensus Standards 6.5, PDF dated 27 July 2025; Risberg & Segadal, NUI report 2023-69 Rev 1, 22 January 2024. All consulted 8 October 2026.

Where your ticket stops you: the certification limit

A separate limit travels with the diver rather than the job. Your qualification is written for a technique and often for a depth, and no amount of experience overrides the certificate. The HSE's List of Approved Diving Qualifications, dated 1 January 2026, is the most useful public document on this because it maps qualifications from more than a dozen countries onto British categories, with the depths attached.

Read that list and the structure of the trade becomes obvious. Nearly every national scheme has a shallow scuba tier at 20 to 40 m, a surface-supplied tier that tops out at 50 m, a mixed-gas surface-supplied tier reaching 70 to 75 m, and then a closed bell tier with no depth written against it at all. The closed bell ticket is where the depth limit stops being a number on your certificate and becomes a property of the ship you are standing on. That also explains a hard fact about training: of the 34 schools in satdiver's directory, only five teach saturation at all.

Source: HSE, List of Approved Diving Qualifications, dated 1 January 2026, with depths for the Australian scheme cross-checked against ADAS certification and the Canadian levels against the Diver Certification Board of Canada (CSA Z275.4). Consulted 8 October 2026.

Air has a ceiling, and it is 50 metres

The reason air diving stops around 50 m has nothing to do with the hardware. It is the gas. Breathe compressed air deep enough and two things happen at once: the nitrogen becomes narcotic, and the oxygen becomes toxic. L103 names both in paragraph 43 and sets the ceiling accordingly, with a recommended maximum oxygen partial pressure of 1.4 bar. We go through the first of those problems in more detail in our piece on why air has a ceiling.

This is also the honest answer to the question people search as "what diving depth requires decompression". Decompression is not a switch that flips at some depth; it is a function of depth and time together. HSE's Table 1 in L103 gives the British limits, and it is worth reading as a shape rather than a set of numbers: at 12 m you have four hours, at 30 m you have 50 minutes on in-water or surface decompression, and at 51 m you have 20. The deeper you go, the faster the clock runs.

Source: HSE L103, Table 1, maximum bottom time limitations, 2nd edition 2014, consulted 8 October 2026. Bottom time runs from first exposure to pressure above atmospheric until decompression begins.

Above 50 m the whole method changes. You stop going home at the end of the shift. Instead you live under pressure in a chamber on the vessel and ride a closed bell down to work. The gas changes too: nitrogen comes out and helium goes in, for the reasons we set out in why saturation divers breathe heliox.

Where the fleet actually stops: 300 metres, 17 times over

This is the part no other page on the topic can do, because it needs the hardware data rather than the regulations. satdiver's dive support vessel directory holds 46 vessels, 45 of them saturation-capable. Twenty-two of those publish a rated depth for their saturation system. Of those 22, seventeen are rated at exactly 300 metres.

That is not a coincidence and it is not a physical law. It is an industrial standard that hardened some time in the 1980s and never moved. Two vessels are rated 350 m — the Skandi Arctic (2009) and the Deep Explorer (2016), both twin-bell, both 24-diver systems. Two Gulf of Mexico vessels are rated 305 m, which is 1,000 fsw written in metres. One, the Seven Pelican, is rated 284 m. The median rated depth across the fleet is 300 m and the spread is 66 m from top to bottom. For a global fleet spanning build years from 1977 to 2022 and operators on five continents, that is extraordinarily tight. We look at why in the 300 metre wall.

Source: satdiver's own dive support vessel directory, 46 vessels, compiled from operator and class records, September 2026. Median rated depth 300 m; 21 of the 22 vessels that publish a figure are rated 300 m or deeper. Twenty-eight vessels publish a diver capacity, totalling 493 berths, with a median system size of 18 divers.

Two caveats matter. A rated depth is a certificate, not a logbook: a vessel rated 300 m may never have run a bell below 120 m in its working life. And more than half the fleet does not publish a figure at all, which is a transparency problem rather than a secret — operators simply do not treat system depth as marketing.

Why 300 metres is a wall and not a cliff

Nothing dramatic happens at 301 metres. What happens is that between roughly 180 m and 300 m several costs rise at once, and past 300 m they rise faster than the value of the work.

High-pressure nervous syndrome. Compress a human on a helium-based mixture and at some point the central nervous system becomes hyperexcitable: tremor, myoclonic jerks, disturbed sleep, measurable losses in dexterity and cognition. The NUI review for the Norwegian regulator put the practical threshold plainly — clinically relevant HPNS generally requires about 180 msw or more — while noting that tremor has been reported after rapid compression to as little as 131 to 152 msw, and that one study observed mild to moderate HPNS once depth passed 200 msw. In a Norwegian simulated dive in which 18 divers were compressed on heliox towards 360 msw, symptoms appeared at 240 msw. A 2019 paper in Scientific Reports frames the onset in pressure terms, above roughly 11 bar, and stresses that susceptibility depends on both compression rate and the absolute pressure held. We go further into the mechanism in high-pressure nervous syndrome.

HPNS can be managed rather than cured. Compressing slowly, holding stops on the way down and adding a narcotic gas — nitrogen to make trimix, or hydrogen to make hydreliox — all push the threshold deeper. None of it is free: slow compression means more days of chamber time before anyone does any work.

Gas density. Helium is light, but at 30 bar even helium is thick. The deeper you go the more effort each breath costs, and the harder it becomes to clear carbon dioxide. That is the quiet limit: a diver who cannot move enough gas through their lungs cannot shed CO₂, and a sudden burst of exertion at depth can overwhelm them. It is why deep mixtures hold nitrogen to a minimum, and why hydrogen — lighter still, but flammable — keeps being revisited.

Heat. Helium conducts heat roughly six times better than air, and a diver at depth loses heat through the lungs as well as the skin. L103 paragraph 81 requires active heating of the inspired gas for dives deeper than 150 metres. DMAC's guidance on bell lockouts puts the practical threshold at around 150 msw too. Below that depth you are not just heating a suit, you are heating the breathing gas, and if the heating fails the dive ends immediately.

Decompression time. Coming up from saturation takes roughly a day per 30 metres. A 300 m sat means around ten days of decompression before anyone steps out of the chamber, on top of the days spent going down and the days spent working. That cost is why the deeper end of saturation has never been cheap at any oil price.

How long you can stay, which is a depth question too

Depth and duration are the same problem seen from two angles, and the duration rules are stricter and better documented than the depth rules. The Diving Medical Advisory Committee's guidance note DMAC 21, revision 2 of June 2006, applies where no national regulation does, and says a saturation should not normally run beyond 28 days. The surface interval afterwards should normally match the length of the saturation, or at minimum be the lesser of half that length and ten days. Cumulative exposure should stay under 182 days in any twelve months.

Then comes the depth-specific clause, and it is the clearest regulatory acknowledgement anywhere that deep saturation is different in kind. For deep dives — DMAC's own example is beyond 200 metres — the surface interval should be at least as long as the saturation itself, with 28 days preferred. Go deep and you buy yourself a month ashore whether you wanted one or not.

Inside the saturation, the working day is governed by DMAC 20, revision 1 of November 1997. North Sea practice limits a diver to four hours locked out of the bell on any one bell dive; with three-man bell runs, six-hour lockouts are permitted, with each diver taking the bellman seat every third run. DMAC is blunt that no physiological parameter sets these limits — fatigue does, and the diver and supervisor decide. It also notes that lockouts elsewhere in the world are often considerably longer.

Sources: DMAC 21 Rev. 2, June 2006; DMAC 20 Rev. 1, November 1997; HSE L103, 2014; Risberg & Segadal, NUI report 2023-69 Rev 1, 2024. Consulted 8 October 2026.

The experimental records, with their conditions attached

These are the numbers that get quoted without context, so here is the context.

701 metres, 20 November 1992, dry chamber. Théo Mavrostomos, at Comex in Marseille, during the Hydra 10 programme, breathing hydreliox — hydrogen, helium and oxygen. He spent about two hours at the simulated depth. Guinness World Records lists it as the deepest dive in a hyperbaric chamber. It was a gas test. There was no seabed, no tooling and no work.

686 metres, 1981–82, dry chamber. The Duke University work on trimix and compression rate, published by Bennett, Coggin and McLeod in 1982 and still cited in the Norwegian regulator's 2024 review. This is the dive usually called Atlantis III. We are citing it through that reference list rather than from the original paper, which we could not obtain.

534 metres, 1988, open sea. Comex divers in the Mediterranean, during the Hydra 8 programme. The 2019 Scientific Reports paper describes it as the deepest known working dive, at about 53.7 bar. This is the one record on the list that involved actual work.

501 metres, 1977, open sea. Comex's Janus IV. We have this only from secondary accounts rather than a primary source, so treat the figure as reported rather than verified.

The gap between 534 m in 1988 and the 300 m the fleet is built for in 2026 is the whole story of this topic. The record is 38 years old. Nothing has gone deeper commercially since, because nothing needed to.

Sources: Guinness World Records, deepest dive in a hyperbaric chamber; Bliznyuk, Grossman & Moskovitz, Scientific Reports 9:10814 (2019); Bennett, Coggin & McLeod (1982) as cited in NUI report 2023-69; satdiver vessel directory. Consulted 8 October 2026.

The one-offs: deep dives that happened and never became normal

Between the 300 m norm and the 534 m record sits a small set of real commercial dives that went deeper than the industry's comfortable range. They are worth knowing about precisely because they did not change anything.

The best documented recent one is a commercial saturation off Australia in 2019. The NUI review describes 15 divers over 27 days of exposure, with storage depths of 234 and 250 msw and a maximum excursion depth of 273 msw. The review's own sources for it are telling: media coverage, a conference presentation, personal communication and a 2023 court decision. When a commercial dive goes that deep, the paper trail tends to run through litigation rather than through technical journals.

The review also records that one of its authors had previously identified 248 saturation dives deeper than 300 msw across several countries. That is a historical total, accumulated over decades, not an annual rate. Set against a fleet that today runs thousands of bell runs a year at 60 to 180 m, 248 is a rounding error — and the review states plainly that diving deeper than 200 msw is uncommon.

One operator's procedures reviewed in the report carried saturation storage tables out to 350 msw with maximum excursions in the 300 to 350 msw band. So the paperwork for deeper diving exists. The demand does not.

What depth is actually for: platforms, not records

Here is the commercial logic that sets the real limit, and it has nothing to do with physiology. Divers are used where structures are. satdiver's platform directory holds 87 offshore platforms, 85 of which publish a water depth. Only 35 of them — 41 per cent — sit at 300 m or shallower. Every one of the other 50 is beyond the reach of a conventionally rated saturation system, and always has been.

Split that by basin and the geography of diving work falls out of it.

Source: satdiver's own offshore platform directory, 87 platforms of which 85 publish a water depth, compiled September 2026. Deepest in the set: Stones, Gulf of Mexico, 2,900 m.

That table explains why a North Sea diver and a Gulf of Mexico diver with the same certificate do different jobs, which is the subject of same job title, different job. It also explains why the fleet never grew past 300 m: everything shallower than 300 m is diver work and everything deeper went to remote systems decades ago. There was never a band of structures sitting at 400 or 500 m large enough to justify building a fleet for it. We set that out in full in how much of the offshore estate is beyond diver depth.

Three coincidences in the data are worth noticing. Troll A stands in 303 m of water — three metres past the standard rating. Heidrun and Snorre A both sit in 350 m, the exact rating of the two deepest vessels in the fleet. And the shallowest Gulf of Mexico platform in the set, Cognac at 312 m, is already out of diver range and has been decommissioned.

Scuba, welding and the other depth questions

How deep can a commercial diver go on scuba? Not far, and in offshore work usually not at all. US federal rules stop scuba at 130 fsw; ADCI stops it at 100 fsw, or the no-decompression limits, whichever comes first; the Australian scheme writes 30 m on the certificate. IMCA maintains a specific recommended practice on the limitations of using scuba offshore, D033, last revised May 2026. In offshore saturation and surface-supplied work, scuba is a tool for inshore and inspection jobs, not a method for depth.

How deep does a diving bell go? As deep as the system it hangs from is rated, which in practice means 300 m. The bell is not the limiting component; the chamber complex, the gas reclaim system and the decompression schedule are. Regulation is clear about when you need one: deeper than 50 m in the UK, deeper than 300 fsw under US federal rules.

How deep can you weld underwater? The welder's depth limit is the diver's depth limit — there is no separate number. What changes with depth is quality. TWI's own technical note explains why wet welds are weaker: bubbles obscure the arc and disturb the weld pool, and the surrounding water quenches the joint fast enough to risk cracking. Dry hyperbaric welding inside a habitat removes those problems and is done at saturation depths, but it is a far larger and more expensive operation. We found no authoritative published depth cap for wet welding and are not going to invent one.

What about deep-sea mining and other new deep work? People assume new deepwater industries mean deeper diving. They do not. The depths involved are in the thousands of metres, which has never been a human range, and the work is done by machines. Deep-sea mining is not a diving job and never was.

What the internet gets wrong about this

"Saturation divers work at 300 metres." They are certified to systems rated for it. The Norwegian regulator's commissioned review says in plain terms that diving deeper than 200 msw is uncommon. Most bell runs on most vessels happen between 60 and 180 m, because that is where the platforms and pipelines are.

"The record is 701 metres, so that is the human limit." That dive lasted about two hours, in a dry chamber, on an experimental hydrogen mixture, in 1992. No seabed, no tooling, no work. It is a gas-physiology result, not an operating depth.

"1,000 feet means 1,000 metres." A 1,000 fsw rating is about 305 m. Two vessels in the directory carry exactly that number, which is why 305 m keeps appearing alongside 300 m in fleet data.

"Depth limits exist because of the bends." Decompression illness is a time-and-ascent problem, and saturation diving largely solves it by keeping the diver at pressure. The limits at the deep end are HPNS, breathing-gas density and heat loss. Different mechanisms, different fixes.

"Technology will push the limit deeper." It has had 38 years to and has not. The constraint is commercial: there is no band of offshore structures between 300 m and 500 m big enough to pay for a deeper fleet, and remote systems already cover everything below. Our fleet audit shows a median build year of 2009 and a rated depth that has not shifted since the 1980s. If demand were there, the paperwork already exists — one contractor's reviewed procedures run to 350 msw.

"Deeper means better paid, proportionally." Depth pay is real and it is contractual, which means there is no public rate card for most markets and anyone quoting you a precise per-metre figure as a global rate is guessing. We set out how the mechanism actually works in what depth pay is and how it is calculated. What the data does suggest is that the premium is thinly earned: the deep dives that attract the largest premiums are the ones that almost never come up.

So what should you take away

If you are weighing up this trade, the useful number is not 701 or 534. It is the band between 60 and 180 metres, because that is where you will spend your career, and the 50 metre line, because that is the one your first certificate will stop at.

The 300 metre figure matters as the fleet's design ceiling and as the reason the industry has the shape it does: 45 saturation-capable vessels for the whole world, 493 berths between them, and a rated depth that has not moved in 40 years. Depth is not the scarce thing in commercial diving. Work inside diver depth is.

If you want the history of how the 300 m system came to exist in the first place, start with the history of saturation diving.

Sources

1. HSE — L103, Commercial diving projects offshore: Diving at Work Regulations 1997, Approved Code of Practice and guidance, 2nd edition 2014, ISBN 978 0 7176 6592 1, consulted 8 October 2026.
2. HSE — Diving Approved Codes of Practice (L103–L107, in force 8 December 2014), consulted 8 October 2026.
3. HSE — List of Approved Diving Qualifications, dated 1 January 2026, consulted 8 October 2026.
4. eCFR — 29 CFR Part 1910 Subpart T, Commercial Diving Operations (OSHA), consulted 8 October 2026.
5. eCFR — 46 CFR Part 197 Subpart B, Commercial Diving Operations (US Coast Guard), consulted 8 October 2026.
6. ADCI — International Consensus Standards for Commercial Diving and Underwater Operations, Edition 6.5, PDF dated 27 July 2025, consulted 8 October 2026.
7. Jan Risberg and Kåre Segadal — Compression and excursion parameters for saturation dives 0–250 msw, NUI AS, Bergen, report 2023-69 Rev. 1, 22 January 2024, prepared for the Petroleum Safety Authority Norway and published by Havtil, consulted 8 October 2026.
8. Havtil (Norwegian Ocean Industry Authority) — Regulations, including the framework regulations on manned underwater operations, consulted 8 October 2026.
9. DMAC 21 — Guidance on the duration of saturation exposures and surface intervals between saturations, Rev. 2, June 2006, consulted 8 October 2026.
10. DMAC 20 — Duration of bell lockouts, Rev. 1, November 1997, consulted 8 October 2026.
11. Diving Medical Advisory Committee — Guidance notes index, consulted 8 October 2026.
12. IMCA — Guidance Document Register, issued 5 October 2026, consulted 8 October 2026.
13. IMCA — D 014, International Code of Practice for Offshore Diving, Rev. 3.3, March 2025, consulted 8 October 2026.
14. A. Bliznyuk, Y. Grossman and Y. Moskovitz — The effect of high pressure on the NMDA receptor: molecular dynamics simulations, Scientific Reports 9:10814, 2019, doi:10.1038/s41598-019-47102-x, consulted 8 October 2026.
15. Guinness World Records — Deepest dive in a hyperbaric chamber (Théo Mavrostomos, 701 m, Marseille, November 1992), consulted 8 October 2026.
16. ADAS — Certification parts and scope (Australia), consulted 8 October 2026.
17. Diver Certification Board of Canada — Certification (CSA Standard Z275.4), consulted 8 October 2026.
18. TWI — What is underwater welding?, consulted 8 October 2026.
19. Wikipedia — Deep diving, used only as a flagged secondary source for the 1977 Comex Janus IV figure, consulted 8 October 2026.
20. satdiver — dive support vessel directory (46 vessels) and offshore platform directory (87 platforms), own data, compiled September 2026.

Frequently asked questions