World Surf League
Live broadcast analytics
Offshore conditions blocked live data. Now, wearables stream telemetry to AI, giving commentators instant insights in seconds.
- Insight retrieval in ~2 seconds for commentators
A one-second delay was too slow for live sports. Cloud inference now delivers stats 10x faster, powering 60+ AI initiatives.
The global leader in sports data manages 7.2 petabytes of proprietary information and 200+ AI modules to power broadcasters, betting markets, and leagues worldwide.
Live sports broadcasts require hyper-accurate statistics delivered in milliseconds, making even a one-second delay unacceptable. Previous inference...
“We don’t just collect data. We’ve defined the language of sport. From KPIs to the way leagues and broadcasters describe performance, a lot of it came from us.”
Sports data and AI for betting, media, and team performance analysis.
LPU hardware and cloud platform for high-speed AI inference.
Stats Perform's Real-time sports insights is part of this use case:
Related implementations across industries and use cases
Offshore conditions blocked live data. Now, wearables stream telemetry to AI, giving commentators instant insights in seconds.
Most of the field played in silence — raw data, no story. Now AI narrates every shot with leaderboard position and player history.
14s lags limited AI play to <5%. Cutting latency 4x unlocked real-time tactics, driving AI modes to >60% of total gameplay share.
Managing 200TB of player data required heavy manual oversight. Now, serverless pipelines let scientists deploy models themselves.
Feedback lagged for 15 seconds. Switching to specialized hardware cut response times to two seconds while halving costs.
1M+ players flooded a Super Bowl puzzle chatbot. A real-time AI security layer absorbed the surge and blocked malicious prompt injections.
Translating updates one language at a time bottlenecked distribution. Now, AI drafts multilingual commentary in minutes for editor review.
Querying Wikidata required specialized syntax, locking out most AI systems. Vector search now lets LLMs navigate 100M+ entities in plain language.
On-premise systems, dispersed and brittle, bottlenecked every release. AI agents now run routine dev steps — hours cut to minutes.
A one-second delay was too slow for live sports. Cloud inference now delivers stats 10x faster, powering 60+ AI initiatives.
The global leader in sports data manages 7.2 petabytes of proprietary information and 200+ AI modules to power broadcasters, betting markets, and leagues worldwide.
Live sports broadcasts require hyper-accurate statistics delivered in milliseconds, making even a one-second delay unacceptable. Previous inference...
“We don’t just collect data. We’ve defined the language of sport. From KPIs to the way leagues and broadcasters describe performance, a lot of it came from us.”
Sports data and AI for betting, media, and team performance analysis.
LPU hardware and cloud platform for high-speed AI inference.
Stats Perform's Real-time sports insights is part of this use case:
Related implementations across industries and use cases
Offshore conditions blocked live data. Now, wearables stream telemetry to AI, giving commentators instant insights in seconds.
Most of the field played in silence — raw data, no story. Now AI narrates every shot with leaderboard position and player history.
14s lags limited AI play to <5%. Cutting latency 4x unlocked real-time tactics, driving AI modes to >60% of total gameplay share.
Managing 200TB of player data required heavy manual oversight. Now, serverless pipelines let scientists deploy models themselves.
Feedback lagged for 15 seconds. Switching to specialized hardware cut response times to two seconds while halving costs.
1M+ players flooded a Super Bowl puzzle chatbot. A real-time AI security layer absorbed the surge and blocked malicious prompt injections.
Translating updates one language at a time bottlenecked distribution. Now, AI drafts multilingual commentary in minutes for editor review.
Querying Wikidata required specialized syntax, locking out most AI systems. Vector search now lets LLMs navigate 100M+ entities in plain language.
On-premise systems, dispersed and brittle, bottlenecked every release. AI agents now run routine dev steps — hours cut to minutes.