On Thursday, 14 May 2026, six divers died in the Maldives. Five were Italian guests aboard the liveaboard Duke of York; the sixth was a Maldivian military rescue diver who entered the same cave system two days later to recover their bodies. As of this writing, only one body — that of the dive instructor leading the group — has been recovered. The Maldives Ministry of Tourism has suspended the operator's license, and a Finnish technical recovery team has been called in.
This is the deadliest single diving incident in the country's history, and one of the worst recreational/technical diving accidents recorded anywhere in the past decade. It is also, on the available evidence, an avoidable one. What follows is a sober, evidence-based breakdown of what happened, what likely killed them, and what every diver — recreational or technical — should take away from it.
What we know about the dive
The group descended at Devana Kandu, a cave system near Alimathaa Island in Vaavu Atoll, roughly 100 km south of Malé. Vaavu is well known to recreational divers for its dramatic channel diving, drift dives, and shark encounters. Less well known — and far less forgiving — are the cave systems that extend through the carbonate reef structures around several of its islands.
Devana Kandu is not a karst cenote of the kind found in the Yucatán. It is a series of coral-reef tunnels, chambers, and overhangs at depths reported between 50 and 60 meters (164 to 197 feet). The recreational diving depth limit in the Maldives is 30 meters. Diving beyond that requires explicit prior authorization from the local authorities — authorization the group did not seek and the booking operator says it would never have granted.
The five Italian guests entered the system together. According to media reconstructions, the only body recovered so far is that of dive guide Gianluca Benedetti, the operations manager for the local liveaboard arm of Albatros Top Boat. Italian press reporting, citing search-team accounts, says his body was found inside the second chamber of the system at approximately 60 meters, his cylinder pressure at zero. A spokesperson for the office of President Mohamed Muizzu, who visited the site, said Benedetti was found "near the mouth of the cave." The two accounts are not yet reconciled in public reporting. What both accounts agree on is that his breathing gas had been completely consumed.
The other four divers remain inside the system. They include two University of Genoa research staff — associate professor of ecology Monica Montefalcone and research fellow Muriel Oddenino — along with biomedical engineering student Giorgia Sommacal (Montefalcone's daughter) and recent marine biology graduate Federico Gualtieri. The University of Genoa has publicly stated that the dive was personal, not part of any sanctioned scientific mission.
Certification, training, and the line you do not cross
The first thing experienced cave divers noticed about this incident was a mismatch between the kind of training reportedly held by some of the group and the kind of dive they attempted. Italian press reporting indicates at least one of the divers held a "Level 1" cave certification from CMAS- and FIPSAS-aligned agencies — an entry-level introduction.
The training pipeline matters here. Every recognized cave diving curriculum — whether from TDI, NSS-CDS, NACD, GUE, or the CMAS family — distinguishes sharply between three progressively more committing levels:
- Cavern Diver / Level 1. An overhead-environment introduction. Penetration is limited (typically 60 meters / 200 feet of linear distance from the surface). Depth is limited to roughly 21 meters. Crucially, the diver must remain inside the "daylight zone" — the area where natural light from the entrance is still visible. The course teaches a single continuous guideline and basic overhead protocols.
- Intro Cave / Apprentice Cave. The first true cave certification. The diver moves beyond the daylight zone but follows a single continuous main line, with no jumps, gaps, or T-intersections. Gas planning, propulsion, and emergency response in zero visibility are formally drilled.
- Full Cave. The complete certification, covering jumps, gaps, circuit traverses, and stage decompression. Depth and complexity beyond this point typically require additional specialty training in mixed gas and decompression procedures.
A Level 1 / Cavern certification is, by every published standard, explicitly not a qualification to enter a fully enclosed, light-less, 60-meter coral cave with multiple chambers. Doing so is roughly equivalent to using an open-water license to attempt a 90-meter Trimix wreck penetration. The skills, equipment configuration, gas planning, and decompression doctrine are different categories of activity. We do not know what each member of the group had trained for. We do know that the environment they entered demanded Full Cave certification at minimum, layered with mixed-gas and decompression specialties — and that this is a non-negotiable industry standard, not a guideline.
The DIVEVOLK guide to cave diving safety, line techniques, and emergency protocols covers the training progression in more depth. The short version: cave certifications are not interchangeable, and skipping levels is the single most consistent risk factor in cave diving fatalities worldwide.
The gas physics: why Nitrox 28 cannot survive 60 meters
Italian media have reported, citing search-team observations and expert commentators, that the group was diving on Nitrox 28 — a breathing gas containing 28% oxygen and 72% nitrogen. Whether this was the actual fill in every cylinder has not been officially confirmed; final analysis will depend on the cylinders being recovered. But the hypothesis is widely repeated in the Italian press and has been advanced by, among others, pulmonologist Claudio Micheletto and Alfonso Bolognini, president of the Italian Society of Underwater and Hyperbaric Medicine (SIMSI).
If correct, this single fact is a more or less complete explanation for what happened.
Nitrox is breathed safely when the partial pressure of oxygen (ppO₂) stays inside accepted physiological limits. Divers Alert Network and most agency standards cap working ppO₂ at 1.4 ATA, with a contingency limit of 1.6 ATA reserved for short emergency excursions only. Above 1.6 ATA, the risk of acute central nervous system oxygen toxicity — sudden seizures, loss of consciousness, drowning — rises sharply.
The math is unforgiving. ppO₂ equals the fraction of oxygen in the mix multiplied by the ambient pressure in atmospheres absolute (ATA). Ambient pressure is 1 ATA at the surface and increases by 1 ATA for every 10 meters of seawater.
- For Nitrox 28 (28% O₂), the working maximum operating depth at 1.4 ATA is roughly 40 meters. The contingency limit at 1.6 ATA is roughly 47 meters. These are the standard MOD values trained into every Nitrox-certified diver.
- At 50 meters (6 ATA), Nitrox 28 yields a ppO₂ of 1.68 ATA — already above the contingency limit.
- At 60 meters (7 ATA), the ppO₂ becomes 1.96 ATA — well into the territory where acute oxygen toxicity becomes statistically likely with continued exposure.
This is not a marginal exceedance. It is a depth-pressure regime where the gas itself becomes the threat, independent of any error in navigation or buoyancy. CNS oxygen toxicity hits without reliable warning. A convulsion underwater, with a regulator in the mouth, will usually result in the regulator being expelled and the diver drowning before help can reach them. In a confined cave at 60 meters, "help" is in any case minutes of carefully managed ascent away.
The technically correct gas for a 60-meter cave dive is not Nitrox at all. It is a helium-based Trimix — typically something like 18/45 or 21/35 (oxygen/helium fractions, with the remainder nitrogen) — chosen specifically to keep ppO₂ below 1.4 ATA at maximum depth while reducing nitrogen narcosis. Anyone trained in technical cave diving knows this. The reported use of a Nitrox blend that exceeds its MOD by 13 to 20 meters is the central physical fact of this incident.
The empty cylinder: silt-out and gas exhaustion
There is a competing — or more accurately, complementary — explanation that the publicly reported physical evidence supports. Benedetti's cylinder was depleted to zero. That is not consistent with a sudden, instantly fatal oxygen toxicity convulsion, which typically leaves a diver immobilized before significant additional gas is consumed. It is consistent with a prolonged struggle inside the cave system: a navigation error, a silt-out, separation from the guideline, and frantic, accelerated breathing as gas supplies dwindled.
Coral-reef caves are particularly prone to "silt-out" — a sudden collapse of visibility caused by fine carbonate sediment being kicked up off the floor. A single careless fin stroke can reduce visibility from clear blue water to zero in seconds. In trained cave divers, this triggers a rehearsed response: stop, establish contact with the guideline, exit by touch following directional markers. In divers without that training, or without a continuous guideline back to open water, it triggers panic — and panic in a sealed overhead environment is a death spiral. Breathing rate spikes. Gas is consumed at three to five times the normal rate. The exit, even if technically reachable, becomes unreachable on the remaining gas budget.
It is plausible that both factors were in play: the group descended on a gas inappropriate for their depth, one or more divers suffered an oxygen-toxicity event in the deeper chambers, and the remaining divers — including Benedetti — were trapped trying to manage the emergency until their own cylinders ran dry. The final investigation will turn on cylinder analysis, dive-computer data downloads, and the recovery of the remaining four bodies. None of that data is yet public.
The secondary tragedy: a rescue diver lost to decompression illness
Two days later, on Saturday 16 May, MNDF rescue diver Staff Sergeant Major Mohamed Mahudhee entered the cave system as part of the body-recovery operation. He cleared the first two of three chambers without finding additional remains. On surfacing, he was evacuated to Malé with suspected decompression illness. He did not survive.
Cave recovery at this kind of depth is, by orders of magnitude, more dangerous than the original dive. The recovery team is not simply visiting a known site at a known depth on a known gas — they are entering an environment that has just killed a team of five, with limited intelligence about exactly where each body is, in conditions that may already be heavily silted from the original incident. Physical exertion is unavoidable: navigating restrictions in poor visibility, working with body-bag equipment, often dragging significant negative or neutral loads. Every additional minute deep extends the decompression obligation; every spike in workload increases inert-gas uptake.
If the recovery dive runs longer than planned, or workload pushes gas consumption past the deco budget, the diver faces a brutal choice: complete the prescribed stops and risk running out of gas, or shorten the stops and surface knowing that inert gas may form bubbles in joints, tissue, and — most dangerously — the spinal cord and brain. Decompression illness can be treated if recompression is reached in time, but the window is narrow and the outcome is never certain. Our coverage of immersion-related diving emergencies and dive injury prevention covers the broader risk landscape, but high-overhead recovery work is at the extreme end of it. Sergeant Major Mahudhee was, by all accounts, an experienced military diver. Cave systems at 60 meters take experienced divers too.
After his death, the Maldives suspended further recovery operations. Three Finnish technical cave divers, all affiliated with the global Divers Alert Network, arrived in the Maldives on Sunday 17 May to advise on a new recovery plan. As of writing, recovery operations remain paused, with a yellow weather warning compounding the operational difficulty.
Regulatory and operational failure
Several actors are now in scope of the official investigation, and the broad shape of the failure chain is already visible.
The Maldives 30-meter limit was bypassed. Any dive deeper than 30 meters in Maldivian waters — recreational, technical, or scientific — requires advance authorization. The Ministry of Tourism has stated no such authorization was issued for this group. The operating license of the Duke of York has been suspended indefinitely pending the inquiry.
The booking agent has disclaimed authorization. Albatros Top Boat, the Italian tour operator based in Verbania that sold the cruise package, has stated through its legal representative that the cruise was sold as a coral-sampling itinerary at standard recreational depths, that any deep cave dive at Devana Kandu would have required additional permits, and that the agency neither owned the vessel nor employed the crew operating it locally. The accuracy and completeness of this statement will form part of the inquiry.
The University of Genoa has stated the dive was personal. Although four of the five Italian guests had academic affiliations with UniGe and one was on a research trip to the Maldives, the university has formally distanced itself, stating the deep cave dive was not part of any official scientific program. This matters institutionally and legally; whether it changes the operational picture on the day is another question.
The failure here was not at any single point. It was systemic: a recreational liveaboard environment in which a small group was able to organize, gear up for, and execute a deep technical cave dive that no regulator, no booking agent, and no published certification standard authorized. The forensic question is which links in that chain failed actively, and which failed by simply not being there.
What every diver should take from this
It is tempting to read incidents of this magnitude as exceptional — extreme divers attempting an extreme dive — and conclude that nothing here applies to the rest of us. That reading is wrong. The decisions that led to the Devana Kandu accident are scaled-up versions of decisions recreational divers make on every trip. The same shape of failure kills divers every year at 20 meters. A short, practical list:
1. Treat depth limits as engineering limits, not social conventions. The 30-meter recreational limit, the MOD of your nitrox fill, your certification's maximum depth — these are not bureaucratic boundaries. They are derived from physiology and equipment performance, and they have safety margins built in for a reason. The margin is what keeps you alive when something else goes wrong.
2. Know the MOD of every gas you breathe. Every diver, recreational included, should be able to calculate the maximum operating depth of their fill in under ten seconds. If you don't do that calculation, you can't catch the error when a tank gets mislabeled, a fill goes wrong, or a dive plan drifts past your gas's safe envelope.
3. Overhead environments are a category, not a feature. Wrecks, ice, caverns, coral reef tunnels, and full caves are all overhead environments. They share one fatal property: no direct ascent. If you have not been formally trained for overhead diving, the answer to "should I follow them in for a quick look?" is always no. Light from the entrance is not a substitute for guideline, redundancy, and rehearsed exit procedures.
4. Buddy-team capability is set by the least-capable member, not the most. A team in a cave is only as safe as the diver with the weakest gas planning, weakest emergency response, and weakest overhead training. If even one diver in a group is below the dive's required level, the dive itself is below standard.
5. Communication is a safety system. A surprising fraction of dive incidents trace back to ambiguous communication: misunderstood plans, missed signals, divers who could not raise the surface or signal the boat when something went wrong. Practice your dive hand signals, run pre-dive briefings the same way every time, and consider modern underwater communication tools where they fit the dive profile.
6. Stories like this one only work if we learn from them. Cave diving in particular has a long tradition of public, technical accident analysis — examining what went wrong in each fatality so the next generation can avoid the same failure modes. That tradition has made certified, trained cave divers among the safest groups in the sport. It only works if the rest of us treat each report as a study, not a spectacle.
Six divers — five Italian visitors and one Maldivian rescuer — entered the water at Devana Kandu in mid-May 2026 and did not return. The full forensic story will take months and depends on equipment yet to be recovered. But the broad arc is already legible: a complex overhead environment, a gas mix beyond its operating depth, a team mismatched to the dive, and a regulatory framework that was bypassed. None of those failures was novel. All of them are preventable in any dive any of us is going to plan this year. That is the only useful response to a story like this one.
