With the retail industry’s increased challenges for skilled practitioners and monitoring inventory globally, AI adoption can address some of their current and future concerns. However, retailers must first address the requirements of their aging digital infrastructure to handle advanced computing needs and facilitate integrated AI adoption.
The retail sector is embracing artificial intelligence (AI) to boost productivity, automate operations, increase profits, and optimize business processes. According to global research firm Gartner, “Global AI software spending in the retail market is forecast to increase 15.8% in 2024 to $7.8 billion and reach $12.5 billion by 2027, with a five-year CAGR of 16.5%. Technology and service providers can use this presentation to support retail industry planning activities for 2024 and beyond.”* Moreover, “In 2025, retailers continue to pursue adoption of AI, with >30% planning to increase budgets for AI and GenAI,”** which we believe is a strong indication of long-term high adoption and interest.
Current applications of AI in retail
Retail giants and innovative startups are deploying AI to:
- Personalize shopping
- Improve chatbot experiences
- Enhance upselling precision
- Streamline checkout processes
- Manage inventory
- Forecast demand
- Assist customers
While some are already far along in implementing their AI strategies, others are still planning their roadmaps for incorporating AI into their operations. They face a significant challenge: an aging IT infrastructure with poorly integrated systems. Large chains often struggle with a patchwork of disparate systems resulting from years of mergers and acquisitions. This typically requires a major refresh across the entire data center, point of sale, and distribution landscape.
Growth goals mean infrastructure challenges
Retailers want to use AI to predict trends and make real-time adjustments based on market demand. If one area is experiencing a surge in particular orders, AI can help them identify a growing need and recommend a shift in supply chain configuration to better accommodate it.
AI computing demands high-density servers far more powerful than previous generations’ compute capability. Hundreds or thousands of graphics processing units (GPUs) require significantly higher power, cooling, and connectivity. The heat generated by these processors and servers necessitates advanced cooling, and in most cases, only direct-to-chip liquid cooling technology can do the job.
Hence, the retail sector finds itself with a dilemma:
- Strategy calls for AI to be implemented broadly and at once.
- The underlying infrastructure – power, cooling, networks, and servers – is aging and cannot cope with AI workloads.
Take the case of one large US retail chain. Its existing equipment is more than adequate for traditional POS, online sales, and management. While this infrastructure is nowhere near its end of life, it has suddenly become obsolete as their current infrastructure is not even close to what this retailer needs to realize its AI vision.
Preparing retail for AI
Preparations for transitioning systems to accommodate AI in the retail sector must encompass three areas, each one contributing to the alignment of the whole:
- Stores: As the heartbeat of retail operations, stores require uninterrupted service. Cash registers and power supplies can’t ever go down. One retailer estimated losses in the millions per hour of outage. Stores need the latest in PDUs, UPS, and backup power to keep them up and running.
- Distribution centers: Distribution centers may not need the same level of equipment sophistication as primary data centers, but their requirements for automation are steadily growing and expanding. Robots are being introduced for inventory management, and AI-integrated sensors and cameras can further improve efficiency. It is also becoming increasingly common for distribution centers to need several racks of modern IT equipment that protect against outages.
- Data centers: Retail data centers are increasingly incorporating high-density racks (up to 100 kW) to serve AI operations. Next-generation servers, cabinets, power, and cooling systems are crucial for processing and analyzing information from all stores, detecting trends, and responding to market conditions.
Strategizing for success in retail
The retail sector faces a clear dilemma: while strategy demands broad and immediate AI implementation, the underlying infrastructure is inadequate for AI workloads. Retailers should assess their current infrastructure against their AI strategies and roadmaps. Operation teams are advised to review the state of their stores, distribution centers, and data centers to see how they align. The ideal approach is to find a partner capable of providing end-to-end solutions with a large, skilled service team to assist in rapidly AI-enabling disparate systems, creating standardized blueprints for systematic implementation.
The changing retail landscape requires a fresh approach that facilitates AI and bakes it into every facet of operations. Done correctly, AI can boost response to consumer demand, elevate in-store and e-commerce experiences, and improve the energy efficiency of digital systems and responsible business goals.
*Compare AI software spending in the retail industry, 2023-2027,Sandeep Unni, Robert Hetu, James Ingham, Inna Agamirzian, 23 April 2024
**Emerging tech impact radar: Artificial intelligence in retail, Robert Hetu, Sandeep Unni, 11 April 2025
GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.
