I worked with an engineer once who walked the roof of a Caribbean property I was involved with and said something I never forgot. "You see thatch, the guests see paradise, and I see accelerant." He wasn't being dramatic. He was being accurate.

Friday morning, a fire ripped through the Viva Dominicus Beach by Wyndham in Bayahibe, Dominican Republic. 84% occupancy. Roughly 1,690 guests on property, including more than 200 kids. One woman... an Italian tourist, 46 years old... didn't make it out. Nine others were injured. Reports indicate the fire spread with terrifying speed, fueled by wind and sections of thatched or cane roofing material that are standard construction in tropical resort markets. The adjacent Viva Dominicus Palace wasn't touched. Same ownership group, same beachfront, completely different outcome. That tells you this wasn't some act of God that no one could have predicted. It was a building-specific failure with building-specific causes.

Here's the detail from the New York Post headline that should sit with every operator reading this: guests reportedly charged back into the burning building to retrieve their belongings. Passports. Phones. Wallets. Everything they needed to prove who they were and get home. I want you to think about that for a second. Your guests' most critical possessions are in a room they can't reach, and the only people standing between them and a fire are your staff... whoever happens to be on shift at 11 AM on a Friday. That's the moment where your emergency plan either works or it's just a binder collecting dust behind the front desk. And I'll tell you something... I've been in this business 40 years, and I've seen plenty of binders. I've seen far fewer teams that have actually drilled on what to do when the building is on fire and a panicked guest is trying to run past them back inside.

The broader context matters here. Wyndham has been expanding aggressively in Latin America and the Caribbean... roughly 75 new hotels planned over the next five years in the region, with the Dominican Republic as a primary growth market. All-inclusive already represents nearly 30% of their rooms there. This fire doesn't just affect one property. It puts a spotlight on construction standards, fire safety protocols, and emergency response capability across an entire development pipeline. When you're growing that fast in markets where building codes and enforcement may not match what operators are used to in the US, the question isn't whether something like this will happen. It's whether you're ready when it does.

Let me be direct about something the post-incident PR will never say. Evacuation of 1,690 people with one fatality and nine injuries is, statistically, a result that could have been catastrophically worse. But "could have been worse" is not a standard. A woman is dead. Her family flew to the Dominican Republic for a vacation and one of them isn't coming home. That's the number that matters... not the occupancy percentage, not the room count, not the brand's expansion plans. One. And the question every operator should be asking right now isn't "could that happen here?" It's "what am I doing this week to make sure it doesn't?"

I worked with a GM years ago who ran fire drills quarterly. Not the kind where you pull the alarm, everyone rolls their eyes, and the front desk supervisor marks it complete on a checklist. Real drills. Middle of the night. Full evacuation. Guests and all. His ownership group told him he was crazy... said he was going to generate complaints. He told them he'd rather have a one-star review from a guest who got woken up at 3 AM than a wrongful death lawsuit from a family who never got woken up at all. He kept his job. And when a kitchen fire broke out on a Saturday night two years later, his team moved 340 guests to the parking lot in under nine minutes. Nobody got hurt. Nobody panicked. Because they'd done it before.

Friday morning at the Viva Wyndham Dominicus Beach in the Dominican Republic, a fire ripped through the resort around 11 AM. The property's thatched palm roofing... the kind that looks gorgeous in the brochure photos... caught and spread in high winds. An Italian tourist named Francesca Valentino, 46 years old, died. Nine others were injured. Roughly 1,700 guests and staff were evacuated. Reports say the resort was "almost completely destroyed." The cause is still under investigation, but preliminary findings point to exactly what you'd expect... highly flammable traditional roofing materials and wind conditions that turned a fire into an inferno before anyone could contain it.

Let me be direct. This is not a story about one resort in the Caribbean. This is a story about every property in your portfolio that has a life safety plan sitting in a three-ring binder behind the front desk that nobody has opened since the last brand QA visit. Every property where the fire panel is in the engineering office and the overnight guy doesn't know how to read it. Every property where the emergency contact list has phone numbers for people who left two years ago. Every property where the thatched palapa by the pool, or the decorative wood paneling in the lobby, or the aging electrical in the east wing hasn't been assessed for fire risk since the last renovation... if it was assessed then. I've seen this movie before. The fire itself is the headline. The real story is what happens in the 90 seconds between when the alarm sounds and when your team either executes or freezes.

Here's what bothers me about this one specifically. This property had renovations in 2020. They updated rooms, added pools, refreshed the lobby. And the thatched roofing... the most obvious fire risk on the entire campus... stayed. Because it's part of the aesthetic. It's what sells the resort experience. It's in the marketing photography. And somebody, somewhere in that chain of decisions, chose the look over the risk. I'm not saying that's what killed this woman. The investigation will determine causation. But I am saying that in 40 years of operations, I have watched ownership groups and management companies defer life safety investments because the ROI doesn't show up on a spreadsheet. Fire suppression doesn't generate revenue. Sprinkler upgrades don't improve your TripAdvisor score. New roof materials don't get mentioned in the brand newsletter. They just keep people alive. And that math only becomes obvious after someone dies.

The Dominican Republic welcomed 5.6 million visitors in the first five months of this year alone. Local authorities were quick to say tourism in the surrounding area continues normally. And it will. Markets recover from incidents. Properties get rebuilt. But Francesca Valentino doesn't come back. And every operator reading this right now should be asking themselves one question... not "could this happen at my property?" Because you already know the answer. The question is: "If it happened tonight, at 2 AM, with my skeleton crew... what happens in the first 90 seconds?"

I worked with a chief engineer once... quiet guy, 20-something years in the business... who kept a laminated card taped to the inside of the generator room door. It listed exactly what ran on backup power and exactly what didn't. He made every new GM read it on day one. "Because when the lights go out," he told me, "nobody remembers what the generator covers until they're standing in a 95-degree lobby full of guests asking why the air conditioning just died." Most of them were shocked at how short that list was.

That card is what I keep thinking about this week. Duke Energy just got a federal emergency order allowing them to run power plants past their emissions limits to keep up with demand. ERCOT in Texas is forecasting peak demand above 92 gigawatts... a record. Heat index values across a corridor from Texas to the Carolinas are pushing 106 degrees. This is not a forecast problem. This is a right-now problem. And the thing that should keep every GM in the Southern half of the country up tonight is this: your generator almost certainly cannot run your HVAC. It runs your emergency lighting, your fire suppression, your elevators, maybe your PMS server. It does not run the system that accounts for 50-60% of your total electrical load. So when that rolling blackout hits at 3 PM during your check-in rush... you have a building full of people, a lobby that's climbing past 90 degrees, electronic door locks that may or may not have recovered cleanly, and a front desk team that has never drilled for this scenario. That's not an inconvenience. That's a safety liability and a reputation event rolled into one afternoon.

Here's what makes this different from the generic "be prepared for summer" advice you've heard a hundred times. The grid is structurally more fragile than it was five years ago. Population growth, data centers, crypto mining facilities... they've all piled onto infrastructure that was already aging. The 2003 Northeast blackout data is instructive and terrifying: 98% of affected hotels lost air conditioning, 89% lost computers and cooking systems, 88% lost elevators. Eighty-five percent of hotels with generators maintained power to their critical systems. But "critical systems" and "guest comfort systems" are two very different lists, and nobody in that lobby cares about the distinction. They care that it's 100 degrees in their room and the ice machine is dead.

The financial angle here is real and it's hiding in plain sight. Hotels spend an average of $2,196 per available room per year on energy. HVAC is more than half of that. Demand response programs... where your utility pays you to reduce load during peak events... can cut 10-15% off your summer energy costs. Marriott ran a program across 264 properties with projected savings of $9.9 million over five years. Some utilities offer $50 per kilowatt of demand reduction. That's not pocket change for a 200-key property. But enrollment requires advance setup, and if you haven't signed up yet, you're leaving real money on the table while also carrying the full risk of a grid event you have no buffer against.

And then there's the labor piece that nobody wants to talk about until somebody gets hurt. Your valet staff, your pool attendants, your maintenance crew working the loading dock or the roof... OSHA heat illness standards require water, rest, and shade, and enforcement is tightening. One heat-related incident involving a hotel employee this summer in Texas or Florida or Georgia, and you're looking at a workers' comp claim, an OSHA investigation, and a local news story that will do more damage to your employer brand than anything your competitors could cook up. The cost of shade structures and hydration stations and adjusted break schedules is trivial compared to the cost of getting this wrong. This is one of those things where the Invisible P&L is working against you... the expense of prevention never shows up as a line item, but the cost of failure will eat you alive in ways that don't hit the P&L until it's too late to do anything about them.

A 57-year-old woman is dead because the best escape plan available to her was tying bedsheets together and climbing out a fourth-floor window. Her husband watched it happen. The hotel was a Hyatt Regency. The city was Kathmandu. The date was September 9, 2025, during Nepal's anti-corruption protests that killed over 50 people and eventually toppled a prime minister. A mob of 100 to 150 people breached the property, set fires, looted guest belongings, and burned what they didn't take. The hotel told guests to move to higher floors. That advice trapped them.

Let that sit for a second. "Shelter in place, move to higher floors." That's the standard fire response in most hotel SOPs. It makes sense when the fire is accidental and the fire department is coming. It makes zero sense when the fire is intentional and the people setting it are still in the building. The husband just had his $12 million compensation claim dismissed by a Delhi court on procedural grounds... he sued Hyatt's Indian consulting arm trying to establish jurisdiction for something that happened in Nepal. The legal theory was shaky. The court kicked it. He can still file a civil suit. But here's what matters to you and me: the legal outcome is almost irrelevant compared to the operational question this case puts on every GM's desk. What is your plan when the threat isn't a kitchen fire or a gas leak... but people?

I've been through hurricanes, bomb threats, power failures that lasted days, and one situation I'd rather not describe in detail involving an armed individual in a lobby at 3 AM. Every one of those had a playbook. Every one of those playbooks assumed a functioning civil infrastructure... police respond, fire department arrives, the cavalry comes. Kathmandu in September 2025 had none of that. The cavalry wasn't coming. The police were overwhelmed. And the hotel's SOP, designed for orderly emergencies, became a death trap in a disorderly one.

If you're operating internationally... especially in regions with political instability, protest movements, or weak rule of law... you need a separate protocol for civil unrest. Not a paragraph in your emergency manual. A separate protocol. It needs to address evacuation routes when ground-floor exits are compromised. It needs to address communication when cell networks go down (they did in Kathmandu). It needs to address the possibility that "shelter in place" is the wrong call. And it needs to be something your night auditor, working alone at 2 AM, can execute without calling a regional VP who's asleep in a different time zone. The Hyatt Regency Kathmandu is still closed for reconstruction. Nepal's luxury hotel sector reported significant financial losses through the autumn tourist season. One family lost a wife and mother. All because the playbook assumed the world would behave the way it's supposed to.

This isn't just an international problem, by the way. Domestic hotels have faced protest-related incidents, civil disturbances during political events, and situations where local law enforcement was unavailable or delayed. If your emergency plan assumes help is always 10 minutes away... you're making the same bet that hotel in Kathmandu made. And sometimes the bet doesn't pay.