THE BIG WHY — February 2026
11,500 spectators generate over a megawatt of heat. The refrigeration system was never sized for that.
Why Olympic Ice Turned to Slush in Milan
By Herbert Roberts, PE
Ilia Malinin was supposed to be untouchable. The 21-year-old American, widely regarded as the most technically gifted figure skater of his generation, entered the men’s free skate at the Milano Cortina Olympics as the overwhelming favorite. He fell twice. He finished eighth. The skating world was stunned.
But Malinin wasn’t alone. Falls cascaded through the competition like a chain reaction — skater after skater hitting the ice in ways that defied their training, their talent, and their competitive records. Over on the short track, races restarted as many as five times because of crashes at the starting line. Canada’s Steven Dubois, a silver medalist in the mixed team relay, called it “probably the worst ice of the year.” American speed skater Corinne Stoddard was more specific: “This ice isn’t the best for short track right now. I think it’s ... like figure skating ice.” Soft. Slushy. Wrong.
The media blamed nerves. The organizers said the ice was “constantly monitored.” Neither explanation survives contact with the physics.
So — why?
How Arena Ice Actually Works
An Olympic ice rink is not a frozen pond. It is an engineered thermal system — a concrete slab threaded with refrigerant piping, maintained at precise temperatures that vary by discipline. Figure skating ice runs warmer, typically around -3°C to -4°C at the surface, which produces a slightly softer layer that allows blade edges to grip during spins and landings. Short track speed skating demands colder, harder ice — closer to -7°C — because skaters cornering at 30+ mph need a surface that won’t deform under the lateral forces generated by their blades.
The Milano Ice Skating Arena — the Unipol Forum in Assago — hosted both disciplines, which means the facility had to alternate between these two thermal specifications throughout the Games. Each transition requires hours of controlled temperature change: cooling the slab for speed skating, then allowing it to warm slightly for figure skating. Done properly, this is routine. Done under Olympic scheduling pressure, with back-to-back sessions and minimal recovery windows, it becomes a thermal management problem that compounds with every passing day.
The refrigeration system beneath the ice does not simply “make cold.” It removes heat. Every watt of thermal energy that enters the ice surface — from the arena air above, from the lighting rigs, from the Zamboni’s hot water resurfacing, from the building envelope itself — must be extracted by the refrigeration plant and rejected outside the building. The system has a finite capacity, measured in tons of refrigeration, and that capacity was sized for a specific set of operating assumptions.
Those assumptions did not account for the Olympics.
No subscription needed to continue on this issue.
The Megawatt Nobody Modeled
Here is the calculation that should have driven every thermal management decision for this venue, and apparently did not.
A seated human body at rest radiates approximately 100 watts of thermal energy — a combination of convective heat from skin and clothing, radiant heat from body temperature, and latent heat from respiration. The Milano Ice Skating Arena holds 11,500 spectators. At full capacity, that is 1.15 megawatts of continuous thermal energy being released into the arena air column directly above the ice surface.
For a single evening hockey game or ice show, this thermal load is manageable. The event runs two to three hours, the building clears, and the refrigeration system has 12 to 16 hours of overnight recovery to pull the slab temperature back down well below competition specification — building what engineers call thermal reserve. The next evening, the ice starts cold, and the refrigeration system spends the event fighting to hold position rather than recover lost ground.
The Olympic schedule destroyed that recovery cycle. Competition sessions ran morning through evening, separated by hours rather than overnight rest. Practice sessions, warm-ups, and early spectator entry further compressed the gap. The building never fully cleared. The refrigeration system never got its recovery window.
The result is thermal debt — each competition day starting with an ice surface slightly warmer than the day before, the refrigeration plant running at or near maximum capacity just to hold the line, with no margin left for the inevitable spikes when 11,500 spectators file in and the broadcast lighting comes to full power. Add the latent heat from spectator breath — humid exhalation that condenses directly on the ice surface and releases energy as it transitions from vapor to liquid — and the slab temperature creeps upward through the session. By the final skating groups, competing in the warmest ice of the day, the surface has crossed from firm to compliant to soft.
That is not nerves. That is thermodynamics.
A Converted Venue Operating Beyond Its Design Envelope
The Unipol Forum is a premier European sports and entertainment venue. It hosts basketball, concerts, and — through its resident skating school — regular ice events. It was modernized for the Olympics, outfitted with a permanent rink, and the organizers rightly noted that it had been proven in competition before.
What had not been proven was sustained, multi-week, full-capacity ice operations under an Olympic-density schedule. A venue that performs flawlessly for a Tuesday night hockey game followed by a quiet Wednesday is operating within its design envelope. A venue asked to deliver competition-grade ice for 16 consecutive days of sold-out, multi-session, multi-discipline events is operating outside it.
This is a pattern I have seen across 32 years of engineering practice — not in ice rinks, but in every system that fails under sustained loading it was never designed to carry. The turbine blade that survives the certification test but cracks after 10,000 cycles. The composite panel that holds under static load but delaminates under fatigue. The common thread is always the same: the system was validated for a condition that does not match the condition it actually faces in service.
The ice in Milan was not defective. The refrigeration system was not broken. The venue was asked to perform beyond its thermal capacity, under a schedule that eliminated the recovery periods its design assumed, and the physics responded exactly as the physics always does.
What Should Have Been Done
The organizers stated that the ice was “constantly monitored.” I do not doubt that. But monitoring is detection. It tells you the ice is degrading. It does not give you the refrigeration capacity to prevent the degradation. Confusing detection with prevention is among the most common — and most consequential — failures in engineered systems.
The proactive safeguards that were missing fall into three categories:
(a) Oversized refrigeration capacity. The thermal plant should have been specified not for the venue’s normal operating profile, but for the Olympic operating profile — sustained full-capacity loading with compressed recovery windows. That likely required temporary supplemental refrigeration units, which is standard practice for major ice events in warm-climate venues.
(b) Mandatory deep-freeze recovery periods. The competition schedule should have included enforced building-clear windows — not just for ice maintenance, but for thermal recovery. Dropping the slab temperature 3 to 5 degrees below competition specification overnight, then allowing it to rise during the next session, would have provided the thermal reserve the system needed. This requires scheduling discipline that prioritizes ice quality over broadcast convenience.
(c) Aggressive dehumidification. The latent heat from 11,500 people breathing in an enclosed space is not trivial. Supplemental dehumidification systems — pulling moisture from the arena air before it can condense on the ice surface — would have reduced the single largest uncontrolled heat input to the slab.
None of these measures are exotic. None are cost-prohibitive at the scale of an Olympic budget. All of them require recognizing, during the planning phase, that an Olympic ice venue does not operate like a normal arena — and designing the thermal management system for the actual conditions rather than the assumed ones.
The Transferable Lesson
Every system has a design envelope. Every system will be asked, at some point, to operate outside it. The question is never whether the physics will enforce the boundary — it always does. The question is whether someone, during the design phase, had the foresight to ask: what are the actual operating conditions this system will face, and have we sized our capacity for those conditions rather than the ones we are comfortable assuming?
In Milan, nobody asked. And the best figure skater on the planet paid the price.
Herbert Roberts is a Licensed Professional Engineer based in Southwest Ohio with 32 years of aerospace R&D experience and 62 patents. His forensic engineering practice translates technical failures into language the rest of the world can act on. THE BIG WHY publishes monthly. Subscribe to the Engineering Mindset Blog for new posts twice weekly.