Everybody has stood in a full stadium with a phone showing full signal and nothing loading. It's one of the most reliably frustrating experiences in modern spectating, and it happens in venues that have spent enormous sums on connectivity.
The reasons are genuinely interesting from an engineering perspective, and most of them are not the reason people assume, which is that there isn't enough bandwidth.
The density problem
A stadium is possibly the most hostile radio environment that consumer wireless has to deal with. Sixty thousand devices in a bowl a few hundred metres across, most of them within a few metres of dozens of others.
Wi-Fi and cellular both work by sharing a limited amount of radio spectrum among devices. The more devices in range of the same access point, the more time each spends waiting for a turn to transmit. Beyond a certain density, the protocol overhead — devices announcing themselves, negotiating, retrying after collisions — starts consuming a substantial share of the available capacity.
So the constraint isn't the fibre coming into the building, which is usually enormous. It's the last few metres of air between a phone and an access point.
Why the standard answer doesn't scale
The usual approach to density is more access points, each covering a smaller area with fewer users. This works, up to a point.
The problem is that access points close together interfere with each other. Wi-Fi has a limited number of non-overlapping channels, particularly in the 2.4 GHz band, and once you've reused them all, additional access points start competing rather than adding capacity.
Stadium deployments deal with this by using highly directional antennas — often mounted under seats or in handrails, pointing at a small group of seats and shielded from everything else. That's a genuinely clever solution and it's why modern stadium Wi-Fi installations involve thousands of access points rather than dozens.
It's also why retrofitting an older stadium is so expensive. You're running cable to thousands of physical locations inside a concrete structure that wasn't designed for it.
The uplink asymmetry
Here's the thing that broke a lot of early stadium deployments. Networks are traditionally engineered assuming download dominates upload — people consume more than they produce.
In a stadium, at the moment of a goal, that assumption inverts violently. Tens of thousands of people simultaneously upload video. Uplink demand spikes to many times what the network was provisioned for, for maybe ninety seconds, and then collapses back.
Provisioning for that peak is enormously expensive and the capacity sits idle 99% of the time. Most venues deliberately under-provision and accept that connectivity degrades at exactly the moments people most want it, which is a rational economic decision and an infuriating user experience.
The concrete
Physical structure matters more than people expect. Stadium bowls are reinforced concrete, which attenuates radio signals substantially and reflects them unpredictably.
Concourses under the stands are effectively radio caves — surrounded by concrete on all sides, with signal only leaking in from the ends. That's why connectivity often gets worse when you go to buy a drink, at exactly the moment you might want to use a payment app.
And the bowl shape creates reflections. A signal bounces off the far side of the stadium and arrives at a receiver delayed relative to the direct path, which the receiver has to disentangle. Modern protocols handle multipath reasonably well but it costs efficiency.
The human factor
An underrated one. Sixty thousand human bodies are, from a radio perspective, sixty thousand bags of water, and water absorbs at the frequencies Wi-Fi uses.
A stadium full of people has measurably worse propagation than an empty one, which means the network that tested beautifully during commissioning performs differently on matchday. Modern deployment practice accounts for this but it's a genuine source of surprise in older installations.
What the good deployments do
A few things distinguish venues where it works.
Under-seat or handrail access points rather than overhead. Shorter distance, more directional, fewer users per radio.
Aggressive band steering pushing devices onto 5 GHz and 6 GHz where more spectrum and more channels are available, and away from the congested 2.4 GHz band.
Separate networks for operations. Point-of-sale terminals, ticketing scanners and broadcast equipment on isolated infrastructure, so that a crowd hammering the public network doesn't stop people getting through the turnstiles.
Cellular offload coordination. Distributed antenna systems for mobile networks inside the stadium, so that phones have somewhere to go that isn't the Wi-Fi.
Whether it's worth it
Honest question, since these installations cost millions.
The business case is usually built on in-seat ordering, app engagement, sponsor activation and data collection. Those returns are real but modest, and several venues have found the numbers less impressive than projected.
The stronger argument is that connectivity is now an expected part of the product. A supporter who can't post a video from a match feels the venue has failed, in a way that would have been incomprehensible twenty years ago. It's become infrastructure rather than an amenity, like working toilets — nobody praises you for it and everybody notices its absence.
Which is a slightly grim business case for a multi-million pound investment, but it's the accurate one.