Summary
Speed, security, and capacity are just some of the things those investing in AI-powered video surveillance must consider. Read on to learn the rest.
Retail shrink, or shrinkage, describes inventory loss that stems from various internal and external factors. Top among them is theft, which can occur at all stages of the retail supply chain, from the point of manufacture to the point of sale. That includes shoplifting by customers, theft by employees, vendor fraud, and organized retail crime (ORC).
Retail shrink is rising — and eroding retailers’ bottom lines. According to research, it’s projected to reach $132 billion in 2024, and there was a 93% increase in average annual shoplifting incidents in 2023 compared to 2019.
To combat shrinkage and create a more secure environment for customers and employees, retailers of all sizes are stepping up investments in modern, edge-enabled video surveillance solutions, including those powered by artificial intelligence (AI).
On an episode of The Pure Report podcast, “Pixels and Petabytes: The Intersection of Data and Vigilance,” I discussed the importance of modern video surveillance solutions in various sectors, and how Pure Storage® can help organizations meet their infrastructure needs. We didn’t have enough time to dig into everything, so here’s a follow-on look at some ways that retailers can maximize their future-forward video investments with modern data storage.
Next-level Surveillance with AI-powered Systems
Modern AI systems can reduce false alarms by up to 90% through smart detection that distinguishes genuine threats from harmless movements. These eyes and ears are capable of:
- Individual person detection and tracking across multiple cameras and views
- Item recognition that cross-checks removed products with sales transactions
- Behavioral analytics and anomaly detection such as loitering or abandoned objects
- Deep learning algorithms that differentiate between humans, vehicles, and other objects with greater precision
Post-incident, they also help on the forensics front. AI enables instant video searches by filtering footage based on object attributes like clothing color, vehicle type, or license plates, dramatically reducing investigation time.
Edge Computing: Processing Intelligence at the Point of Capture
Edge computing has emerged as a transformative force in retail surveillance for 2025, fundamentally changing how video data is processed and stored. With Gartner predicting that 75% of enterprise data will be processed at the edge by 2025—up from just 10% in 2018—retailers are increasingly deploying intelligence directly at camera locations rather than relying solely on centralized data centers.
The benefits are perfectly aligned with the outcomes sought:
Real-Time Analytics and Instant Decision Making
Modern edge-enabled surveillance cameras process video data instantly, providing actionable insights in real-time without the latency associated with cloud-based processing. This capability reduces response times to under 5 milliseconds, compared to the 20-40 milliseconds typical of traditional cloud computing architectures. For retail loss prevention, this means suspicious behaviors like shoplifting or fraud attempts can be detected and flagged immediately, enabling security personnel to respond within seconds rather than discovering incidents hours later during footage review. These capabilities are particularly valuable in high-traffic retail environments where split-second detection can mean the difference between preventing and experiencing loss.
Bandwidth Optimization and Storage Efficiency
Rather than transmitting hours of raw footage to central servers, edge devices intelligently filter and process data locally, sending only relevant alerts, processed clips, or summary analytics. This approach dramatically reduces bandwidth consumption—a critical consideration as retailers deploy higher-resolution 4K cameras across multiple locations.
The result is significant cost savings on network infrastructure while ensuring that storage systems receive only the most pertinent data for long-term retention.
Enhanced Security and Privacy Compliance
Processing sensitive surveillance footage locally at the edge reduces exposure to potential breaches during data transmission. This localized approach is increasingly important for retailers operating under strict privacy regulations like GDPR and CCPA, as it minimizes the movement of personally identifiable information across networks. Edge computing also provides resilience against connectivity disruptions—cameras continue analyzing and storing critical events locally even when network connections to central systems are temporarily unavailable.
Scalability Challenges and Infrastructure Considerations
While edge computing offers compelling benefits, retailers must navigate significant implementation challenges in 2025. With the number of connected devices projected to reach 75 billion globally, managing distributed edge infrastructure becomes increasingly complex. Industry data shows that 10-15% of edge locations experience connectivity issues at any given time, making reliable device management crucial for maintaining surveillance effectiveness.
The storage implications are substantial. If each edge device requires local storage capacity for temporary buffering, AI model storage, and critical event retention, this requires careful planning around data lifecycle management, ensuring that locally stored footage is appropriately synchronized with central repositories while maintaining compliance with retention policies. As retailers adopt these systems, they need to ensure they have the right IT infrastructure in place to support to get the most value from the data they’re collecting day in and day out.
Managing the Data Deluge from AI-powered Surveillance
AI-powered video surveillance systems generate vast amounts of data, ranging from high-definition video feeds to metadata extracted through advanced analytics. As retailers deploy technologies like facial recognition, behavioral analysis, and object tracking, the volume, velocity, and variety of data escalate significantly. This surge in data necessitates storage solutions that can accommodate the unique demands of AI-enhanced surveillance.
Check out the solution brief to learn more
Make the Most of AI-powered Video Surveillance Investments with the Pure Storage Platform
Modernizing data storage goes hand in hand with implementing AI-powered video surveillance. Legacy storage cannot create the foundation that retailers need to win their war against shrink and threats like ORC and cyberattackers.
Pure Storage solutions for data and analytics provide the tools and capabilities your business needs to meet modern security and safety demands; reduce risk for customers, employees, and your operations; and increase efficiency and profitability. Each requirement for edge-enabled surveillance is met:
| Storage Capacity | Ensure storage capacity can keep pace with ever-growing volumes of unstructured data, including high-resolution video footage and associated metadata. | – FlashBlade//E is designed to manage massive unstructured data growth- Easily scales alongside growing volumes of high-resolution video footage- Capacity-optimized unified file and object storage platform for multi-petabyte surveillance deployments |
| High Performance and Throughput | AI applications in video surveillance require high-performance storage solutions with fast read and write speeds. Quick access to data is crucial for real-time analytics and prompt responses to security events. | – All-flash storage delivers blazing fast performance for real-time analytics- 85% faster queries for rapid video retrieval and analysis- High-bandwidth performance specifically engineered for AI workloads- Multi-dimensional performance across file and object protocols simultaneously |
| Retention Period and Compliance | Retailers must adhere to specific retention periods for video footage, which regulatory requirements often dictate. Storage solutions should facilitate long-term retention without compromising performance. Compliance with data protection regulations is essential to avoid legal and regulatory repercussions. | – Data services aligned with regulatory requirements and compliance frameworks- Ability to implement flexible retention policies through SafeMode snapshots- Seamless scalability for growing retention requirements without disruption |
| Redundancy and Reliability | Redundancy and reliability are crucial to prevent data loss in case of hardware failures. Implementing fault-tolerant storage solutions enhances data reliability and ensures continuous surveillance operation. Solutions should offer not only data backup but also fast recovery at scale. | – 99.9999% uptime guarantee, including during upgrades- Systems designed to continue operating when internal components fail- Fault-tolerant architecture that survives multiple overlapping device failures- Long-term reliability with products from 10+ years ago still in service and continually modernized |
| Data Transfer Speeds | AI-powered surveillance systems with real-time analytics capabilities rely on rapid access to stored data. Systems need to be backed by robust storage that can support high-speed data transfer and keep pace with the continuous recording and retrieval of video data. | – Support for high-speed data transfer protocols including NVMe, NFS, SMB, S3, iSCSI, GPUDirect- Optimized for rapid access to stored data for real-time analytics- High-bandwidth Ethernet focus for efficient data movement- Predictable high-bandwidth performance architecture |
| Compatibility with AI Workloads | Data storage solutions for AI-driven video surveillance should be optimized to handle the specific workloads associated with AI analytics. This includes the efficient retrieval of video segments for analysis, storing metadata, and parallel processing to accommodate the demands of multiple AI algorithms running simultaneously. | – NVIDIA AI Data Platform integration with FlashBlade for AI-ready storage- Purpose-built storage optimized for AI analytics, machine learning, and artificial intelligence- Support for parallel processing to accommodate multiple AI algorithms simultaneously- Removes bottlenecks in data access that directly impact GPU efficiency |
| Edge Storage for Decentralized Systems | – FlashBlade and FlashArray solutions with high-density, low power consumption ideal for edge locations- Portworx storage layer for containerized workloads at the edge- Easy remote management capabilities reducing operational overhead- 85% less energy consumption than competitive all-flash storage for sustainable edge deployments | |
| Cybersecurity Measures | Encryption, robust access controls, and regular security updates are essential to help retailers safeguard sensitive video data. Data storage should mitigate ransomware attacks by securing critical data from being altered or destroyed—and enabling a quick and complete recovery. | – SafeMode Snapshots for ransomware protection with immutable backups- Strong administrator authentication controls for secure access- Pure Storage cyber resilience solutions enabling rapid recovery from incidents |
For more insights on how video surveillance is transforming everyday operations for organizations in sectors from law enforcement to transportation, check out this episode of The Pure Report.
Written By:
