Haein GPUaaS Enables Tanger Factory Outlet Centers to Scale Retail AI Across Malls
- Tanger can quickly provision segregated GPU environments for shopper flow, energy optimization, and predictive maintenance.
- Multi-tenant scheduling and unified visibility let Tanger tailor privacy-preserving models per property and update them frequently.
- Tanger's GPUaaS adoption lowers AI costs, centralizes functions, speeds partnerships, and improves tenant services and resilience.
Haein milestone reframes AI tools for mall operators
SK Telecom marks six months of operation for Haein, its GPU cluster for GPU-as-a-Service (GPUaaS) built to run large-scale AI workloads. The cluster’s full-stack platform and rapid provisioning model are sharpening attention among commercial real estate operators that rely on data-driven tenant services and facility management, including outlet mall owner Tanger Factory Outlet Centers. Industry executives say such platforms make advanced computer vision, demand forecasting and personalized marketing models more practical to deploy at scale across multiple properties.
Main Topic — Tanger can scale mall intelligence with GPUaaS
Tanger Factory Outlet Centers is positioned to benefit from the capabilities Haein demonstrates. The Petasus AI Cloud virtualization and the GPUaaS Service Orchestrator enable quick, secure partitioning of GPU resources and multi-tenant model development, letting retail landlords run experiments on shopper flow analysis, energy optimization and predictive maintenance without long procurement cycles. For a company operating dozens of outlet centers, the ability to spin up segregated GPU environments within hours supports faster proof-of-concept work and shorter paths from model training to live deployment.
The platform’s multi-tenant scheduling and unified visibility help chain operators coordinate concurrent model training across property teams and third-party partners while preserving performance and data isolation. Tanger can use those features to tailor models for specific properties or tenant mixes, run privacy-preserving experiments on aggregated anonymized video and sensor feeds, and update models on more frequent cycles. That capability is especially relevant for seasonal outlets where short lead times for demand forecasting and promotional targeting are commercially important.
Strategically, adoption of GPUaaS reduces barriers for outlet mall operators to centralize AI functions rather than duplicate costly hardware at each site. Tanger can lower total cost of ownership for AI initiatives, move quickly on partnerships with analytics vendors and improve tenant services such as targeted promotions and footfall-driven leasing decisions. The capability also supports operational resilience by enabling remote diagnostics and model retraining to respond to changing shopper behavior.
Platform specifics
Haein’s stack—Petasus AI Cloud, AI Cloud Manager and GPUaaS Service Orchestrator—virtualizes heterogeneous xPU environments and transport protocols like NVLink and InfiniBand, enabling secure, high-performance isolation of GPU and network fabric at the tenancy level in under an hour rather than days.
Operational gains
Real-time visibility, intelligent job scheduling and unified orchestration help reduce deployment lead times and operational overhead, giving mall operators predictable performance and the flexibility to scale AI-driven services across properties.