Edge Intelligence at Scale: GCCs powering AI on the edge for global clients

The explosion of data in the digital age is fundamentally reshaping how enterprises operate, compete and innovate. To keep up the pace, centralised cloud processing workloads must support growing demands for real-time decision-making, enhanced data privacy and operational efficiency.

Edge intelligence, also called Edge AI, offers a solution to help enterprises make faster, autonomous and context-aware decisions, while simultaneously reducing costs and ensuring compliance with evolving regulations.

Deloitte research shows realising returns from edge investments rose by 13 percentage points between 2023 and 2024. Further reinforcing this trend, 25 percent of enterprises already using generative AI plan to launch agentic AI pilots in 2025. This figure is expected to double to 50 percent by 2027. These developments reflect a growing demand for real-time, autonomous AI workloads with capabilities that edge intelligence is uniquely positioned to deliver.

Indian Global Capability Centers (GCCs) are emerging as strategic execution engines that power AI at the edge for global enterprises. GCCs are using their unique blend of engineering scale, ecosystem partnerships and cost advantages to achieve this.

Edge intelligence effectively addresses three critical challenges, as follows:

Latency and autonomy: AI, generative AI and agentic AI demand sub-second decision making. By processing data locally on edge devices, delays caused by cloud roundtrips are eliminated, enabling real-time, autonomous responses.

Bandwidth and cost: The surge in video and sensor data strains networks. Edge AI mitigates this by processing most data locally, transmitting only essential summaries to the cloud, drastically reducing bandwidth costs.

Data sovereignty and privacy: Tightening regulations increasingly restrict cross-border data flows. Edge AI supports local data processing, ensuring compliance without stifling innovation.

Indian GCCs combine scale, expertise and operational efficiency to accelerate Edge AI innovation. They bring substantial engineering firepower, with multitudes of AI experts optimising models for deployment across diverse edge devices, including Central Processing Units (CPUs), Graphics Processing Units (GPUs) and Neural Processing Units (NPUs). Their deep ecosystem partnerships with hyper-scalers, chip vendors and academia speed up the journey from proof-of-concept to production. This is further supported by proactive government initiatives aimed at expanding market access and promoting local production, providing ideal conditions to test and refine products. Additionally, wage and real estate arbitrage provide a significant cost advantage, enabling rapid prototyping and de-risking global rollouts.

Across sectors, Indian GCCs have spearheaded transformative Edge AI initiatives. For instance, a Pune-based manufacturing GCC compressed vibration models from 45 MB to under 5 MB, enabling real-time inference on 4,000 Computer Numerical Control (CNC) machines with latency below 20 milliseconds, significantly reducing unplanned downtime.

In retail, a Bengaluru-based GCC deployed vision AI across 9,000 stores, personalising digital signage locally to reduce bandwidth usage and ensure data residency compliance. In financial services, a Hyderabad-based GCC developed edge analytics for 3,500 bank branches, enabling offline Anti-Money Laundering (AML) scoring that maintains compliance even during intermittent connectivity. In automotive, a Chennai based GCC created lightweight defect detection models for embedded inspections, cutting false rejects by 28 percent and protecting sensitive intellectual property on-premises.

To fully harness the potential of Indian GCCs and Edge AI, global enterprises must strategically locate pilots where the talent resides and India’s GCC workforce offers both scale and sophistication to accelerate delivery at an affordable cost. Designing for device heterogeneity is essential, as edge environments vary widely; GCC engineers bring deep expertise in container orchestration, lightweight models and Over-the-Air updates. Embedding security from the outset is critical since decentralised environments inherently broaden attack surfaces. Finally, integrating edge and agentic AI roadmaps is paramount, as mastering edge orchestration ahead of the agentic AI surge in the coming years unlocks a disproportionate competitive advantage.

Success with Edge AI depends on overcoming several challenges. Model optimisation is crucial because edge devices have limited compute capacity, necessitating advanced compression, quantisation and federated learning techniques to maintain accuracy. Security frameworks must evolve to be robust and decentralised, safeguarding increasingly distributed data assets. System integration remains complex but is essential for reliable edge AI operations. Most importantly, enterprises must grant Indian GCCs full end-to-end ownership of the Edge AI journey.

Fragmented mandates or disjointed onshore development risk siloed innovation, delays, security vulnerabilities and integration failures. Indian GCCs, with their deep expertise and tooling, are uniquely positioned to lead holistic innovation, from model design through deployment, unlocking scalable, sustainable edge AI solutions. With global data generation projected to reach 175 zettabytes annually by 2025, Edge AI adoption is accelerating rapidly.

Indian GCCs’ unique scale, domain expertise and ecosystem integration position them as key players in this transformation. Moving beyond traditional operational hubs, they are evolving into strategic innovation centres that will power intelligent, decentralised AI systems, defining the future of global enterprise operations. Parent companies embracing this integrated GCC model stand to unravel Edge AI’s full potential, turning it into a sustainable competitive advantage in a rapidly digitalising world.

(Rohan Lobo is a Partner & GCC Head at Deloitte)

(The examples and case studies referenced in this article are based on Deloitte’s analysis and publicly available information. They are provided for contextual understanding and do not represent endorsements or specific claims about individual organisations. Deloitte does not assume responsibility for any outcomes related to the use or interpretation of these examples.)