India is on course to add roughly 240 million square feet of warehousing in five years. Whether that space gets smarter along with bigger is the question a new report from PwC India and FICCI, Transforming warehousing with AI. The report estimates that total warehousing stock will grow from about 610 million sq. ft. in 2025 to 850 million sq. ft. by 2030. Its central finding is that most Grade A and Grade B facilities still operate at the “digital” or “automated” stage of maturity. AI-led decision-making, it says, is at a very early stage of adoption.
The constraint has moved from execution to decisions
Over the past decade, GST-led network consolidation, e-commerce, organised retail, quick commerce and 3PL growth have turned the warehouse from a passive storage buffer into an operational hub. Regulation has helped. The report notes that 7,512 WDRA-accredited warehouses were operating as of 31 March 2026, and that the government is making warehousing “more bankable” through e-NWRs.
Warehouse management systems, barcode and RFID scanning, conveyors and robotics now give operators real-time visibility and faster execution. But the report argues the bottleneck has shifted. Land near consumption centres is scarce and expensive, labour is harder to staff, and SKU ranges and delivery windows keep widening. When space and headcount can’t be expanded freely, the main lever left is the quality and timeliness of decisions about how existing space, inventory, labour and equipment are used.
A warehouse management system can show what has happened. The report’s argument is that AI can anticipate bottlenecks, forecast labour and replenishment needs, flag inventory or product-condition risks, and recommend or initiate a response. It builds on WCS and WES layers rather than replacing them.
Where Indian warehouses stand today
Asked how close Indian warehouses are to this shift, Jaffrey Thomas, Partner, Transport and Logistics at PwC India, is direct. “While the foundational automation in Indian warehouses is now mainstream, the adoption of advanced intelligence layers is still a rarity,” he says.
He points to three obstacles holding back AI at scale:
Commercial models. Arrangements are still based on real estate (square feet) rather than productivity, so operators have little financial incentive to invest in smarter operations.
Fragmented data. There is no integrated data layer, and supply chains “still grapple with fragmented data and isolated operational silos,” Thomas says.
Skills. There is a shortage of engineers and operations managers trained to orchestrate, interpret and maintain AI-driven systems on the warehouse floor.
The third point matters because the report is clear that AI is only as good as its inputs. “Foundational readiness determines feasibility,” it says. Organisations that skip data quality, system integration and human oversight will be limited “not by the capability of the model but by the quality and reliability of the data it receives.”
A closed loop, not a bolt-on
The report proposes an “intelligent warehouse framework” built as a closed loop of six stages: sense, understand, predict, decide, execute and learn. Conventional automation ends once a pick is completed or a carton is sorted. In the intelligent model, outcomes such as pick times, inventory exceptions, equipment performance and dispatch results feed back into the earlier stages, so each decision improves the next.
The report adds that the loop only works whole. Sensing without learning gives continuous monitoring but no improvement. Deciding without executing produces recommendations that never become operational outcomes. Execution can be advisory, with an operator confirming, or autonomous, depending on how material the decision is and how strong the oversight is.
The report is also wary of AI hype. It positions AI as “a strategic business capability, not a default technology layer to be applied everywhere.” Where the governing logic can be clearly written down, existing rules-based systems remain “the more rational choice.” AI earns its place where decisions are dynamic, uncertain and interdependent, such as demand forecasting, inventory placement, order fulfilment, labour planning and predictive maintenance. The report suggests judging it on five dimensions: operational efficiency, service excellence, risk reduction, sustainability and responsible adoption.
Will new capacity be smarter capacity?
The report’s stock projections raise an obvious worry: that India builds far more warehouses without improving how they operate. Thomas sees a split.
Grade A, institutional-quality assets, which he says will make up more than half of the 2030 stock, are likely to be shaped by tenants’ demands. “Better racking systems and higher loading and throughput per unit land area” will push developers and operators to plan around operational efficiency, he says.
The picture is less encouraging elsewhere. “Regional and unorganised warehousing might still depend on real-estate expansion rather than intelligent workflow design,” Thomas cautions. He adds that the end-use industries absorbing this supply “will be key stakeholders in determining the adoption of efficiency-driven planning.”
The report suggests newer facilities may have an edge, since recently developed warehouses face fewer legacy-system integration challenges than those in more established markets.
The policy backdrop
Government initiatives are laying groundwork. The report cites PM GatiShakti, the National Logistics Policy and multimodal logistics parks. India’s rise to 38th among 139 countries in the World Bank’s 2023 Logistics Performance Index reflects progress, and the policy’s goal of a top-25 rank by 2030 “would require more efficient and connected warehouses.”
The Unified Logistics Interface Platform (ULIP) had, as of August 2025, enabled API-based integration across more than 30 digital systems and over 160 crore digital transactions. In 2026, the Department of Food and Public Distribution launched an AI-powered smart warehousing system across FCI and CWC facilities, with AI-based monitoring, IoT-enabled environmental sensing and automated gate management.
Three changes for the next five years
The report envisions an intelligent, connected and sustainable warehousing ecosystem by 2047, in line with Viksit Bharat. Asked for the three most important changes businesses and policymakers should make in the next five years, Thomas names these:
1. Build a skilled workforce. The talent gap he identified earlier has to close first.
2. Grow a local deep-tech ecosystem. He calls for “an ecosystem of local deep-tech startups” building solutions for Indian warehousing conditions.
3. Set national standards. India needs national-level data-sharing and cybersecurity standards for IoT devices in logistics, which also answers the fragmented-data problem he raised.
For technology leaders, the message is to resist treating AI as a technology programme. The report calls the intelligent warehouse “an operating philosophy” in which accumulated operational experience is systematically converted into better decisions. In practice, that means identifying the specific decisions where today’s rules fall short, getting data and integration in order first, keeping human oversight proportionate to the stakes, and closing the loop so predictions lead to actions and actions lead to learning.
Warehouses have historically been designed to execute decisions made elsewhere. As Indian capacity scales toward 850 million sq. ft., the report suggests the winners will be those whose facilities make better decisions, not just those with more space.