The Astroworld Festival tragedy in November 2021 killed 10 people and injured hundreds more in a crowd crush during Travis Scott’s headlining set. It was the worst crowd disaster at a music event in the United States in decades, and its effects have rippled through the global live events industry in ways that are still playing out nearly five years later.

For everyone in event production, Astroworld was a wake-up call. Not because crowd safety was a new concern — it wasn’t — but because it forced an industry-wide reckoning with how well existing crowd management practices actually worked. The answer, in too many cases, was: not well enough.

What’s followed is a significant investment in crowd management technology. Some of it is genuinely useful. Some of it is security theatre. Here’s where things stand.

Crowd Density Monitoring

The most impactful technological development is real-time crowd density monitoring using camera-based AI systems.

The principle is straightforward. Cameras positioned around a venue or festival site feed video to an AI system that estimates the number of people per square metre in different zones. When density in any zone exceeds safe thresholds — generally considered to be around 4-5 people per square metre for standing crowds — the system alerts operators.

This matters because crowd crushes don’t happen suddenly. They build gradually as density increases in specific areas. By the time the crowd itself feels dangerously compressed, it’s often too late to decompress safely. Early detection gives event operators a window to intervene — opening additional exits, redirecting flow, pausing artist performances, or deploying crowd management staff to specific areas.

Several companies now offer this technology. Crowd Connected uses a combination of camera analysis and anonymised mobile phone location data. IDEMIA and Hikvision offer camera-only solutions with varying degrees of accuracy.

The accuracy of these systems has improved substantially since the early implementations. Modern systems can estimate density within about 10-15% accuracy in open-air environments, and better than that in structured environments with fixed camera positions.

Team400’s AI automation work includes similar computer vision applications, and the underlying technology — counting and tracking objects in video feeds — is well-established. The challenge in live events isn’t the AI itself but the integration with operational response procedures. A density alert that nobody acts on is worthless.

Wearable and Mobile Monitoring

Some events have experimented with wearable devices (RFID wristbands with accelerometers) that detect unusual crowd movement patterns — surging, crushing, or excessive compression. The data from thousands of wristbands creates a heat map of crowd pressure across the site.

The concept is sound, but the practical implementation has been mixed. Wristbands need to be worn consistently (some attendees remove them), the data processing requires significant computational infrastructure, and the false positive rate for “unusual movement” is high — a mosh pit looks a lot like a crowd crush to an accelerometer.

Mobile phone-based monitoring is more promising because it doesn’t require special hardware. By analysing the density of mobile phone signals (with appropriate privacy safeguards), event operators can track approximate crowd distribution across a site in real-time. This approach is less precise than camera systems but covers larger areas and works in environments where camera coverage is impractical.

Improved Communication Systems

One of the critical failures at Astroworld was the breakdown of communication between security staff on the ground, the production team, the artist, and emergency services. Ground-level security recognised the danger but couldn’t effectively communicate up the chain quickly enough to stop the show.

Post-Astroworld, many events have invested in redundant communication systems: dedicated radio channels for safety staff separate from production channels, push-to-talk smartphone apps as backup, and pre-agreed protocols for escalating crowd safety concerns directly to the stage manager with authority to stop a show.

The Event Safety Alliance has published updated communication protocols that specify who can trigger a show stop and under what circumstances. The key principle is that any qualified safety officer should be able to initiate a stop without requiring approval from production or artist management. This was contentious initially — stopping a headliner’s set has massive financial implications — but the industry has broadly accepted it.

Venue Design and Flow Engineering

Technology alone doesn’t prevent crowd disasters. Physical space design is equally important.

Post-Astroworld, event designers are paying more attention to crowd flow modelling — using simulation software to predict how crowds will move through a site under different scenarios (normal operation, weather event, artist changeover, emergency evacuation).

These simulations can identify pinch points where density is likely to build: narrow pathways between stages, bottlenecks at entrances and exits, dead-end zones where crowds can become trapped. Identifying these risks during the design phase is far more effective than trying to manage them in real-time.

Some festivals have reconfigured their site layouts specifically to improve crowd flow. Wider pathways, multiple exit routes from standing areas, barriers that channel flow without creating traps, and designated decompression zones near stages are all design features that reduce crowd crush risk.

What’s Still Missing

Despite the progress, there are gaps.

Standardisation. There’s no universally adopted standard for crowd density monitoring at live events. Different systems use different metrics, different thresholds, and different alert protocols. What counts as a dangerous density in one system might not trigger an alert in another.

Small event coverage. The crowd monitoring technology described above is expensive enough that it’s mainly deployed at major festivals and large-capacity events. The 2,000-person regional festival or the 500-capacity venue show — where crowd incidents also occur — generally can’t afford these systems.

Training. Technology is only as useful as the people operating it. Many event staff receive minimal training in crowd dynamics and don’t fully understand what the monitoring data means or how to respond to alerts. Investment in technology needs to be matched by investment in training.

Post-event accountability. There’s still limited systematic collection and sharing of crowd safety data across the industry. Events that experience near-misses rarely publicise them, which means the industry loses opportunities to learn from incidents that didn’t quite become disasters.

The Cultural Shift

Beyond technology, the most important change post-Astroworld is cultural. There’s a growing acceptance that crowd safety isn’t just a compliance checkbox — it’s a core production responsibility that ranks alongside sound quality, artist management, and ticketing.

I’ve been in this industry long enough to remember when crowd management was an afterthought — a few blokes in hi-vis standing near the stage. That’s not good enough, and it was never good enough. What Astroworld did was make that fact undeniable.

The technology helps. Better monitoring, better communication, better site design — all of it reduces risk. But the foundation is a culture that takes crowd safety seriously at every level, from the festival director to the security guard on the barrier. Technology supports that culture. It doesn’t create it.