A new report by Dell Technologies and Zinnov finds that Global Capability Centers poised to lead by 2030 won't be those running the most AI pilots, but those with strong data, governance and infrastructure foundations to scale AI responsibly.
Dell Technologies, in partnership with Zinnov, released a report at the Dell Technologies Forum 2026 outlining the next stage of evolution for India's Global Capability Centers (GCCs). Titled "India GCCs 2030: From Capability Centers to Agentic Transformation Engines," the report is based on surveys and interviews with more than 50 senior GCC leaders across the BFSI, retail, manufacturing and software sectors.
Its central finding: the GCCs set to lead by 2030 will not be defined by how many AI initiatives they run, but by whether they have built the groundwork needed to scale AI into measurable business outcomes.
Manish Gupta, President and Managing Director, Dell Technologies India, said, "The most influential GCCs of 2030 will not be measured by the number of AI initiatives they launch, but by their ability to industrialize AI responsibly and at scale. As they take on greater strategic ownership, robust foundations across data, infrastructure, and governance will become the bedrock of enterprise innovation. The GCCs that build these capabilities now will define how their organizations harness AI globally, and Dell Technologies is focused on enabling that journey from foundation to transformation."
Sidhant Rastogi, President, Zinnov, said, "The GCC model is reaching an inflection point. For the last two decades, the conversation was largely about scale, talent, and capability. The next decade will be about ownership. As AI and agentic systems become embedded into enterprise workflows, GCCs will increasingly be expected to own products, platforms, markets, and measurable business outcomes. Those that build the right data, technology, governance, and talent foundations now will move from being capability centers to becoming true transformation engines for the enterprise."
Sector scale and the pilot problem
India is home to over 2,100 GCCs employing 2.36 million people and generating $98.4 billion in revenue in FY26. The report found that 64 percent of GCC leaders hold dual global mandates, overseeing their India centre while also owning a global function, and 70 percent operate with a defined AI roadmap. Indian GCCs now account for roughly 28 percent of global GCC AI talent, with more than 1,200 centres having developed AI and machine learning capabilities. Notably, 66 percent of leaders already rank top-line business impact as a high priority in their AI strategy, suggesting the conversation has shifted well past cost and delivery.
The maturity curve is compressing too—27 percent of new GCCs now reach "Portfolio Hub" maturity within five years, down from nearly a decade historically, with AI mandates arriving earlier in a centre's lifecycle.
Yet scale alone isn't translating into results. Close to 70 percent of GCCs remain stuck at the pilot stage, unable to move proofs of concept into sustained adoption—hampered by fragmented data, legacy systems, unclear governance, weak security controls and talent models not built for an AI-driven environment.
Four levers for the next phase of growth
The report attributes stalled pilots to structural issues: production data is messier than test environments, governance tends to be bolted on after deployment rather than built in from the start, and use cases built outside standard enterprise platforms are hard to scale. Costs also shift sharply at production volume—agentic workflows can consume anywhere from 10,000 to 500,000 tokens per workflow, compared to just 1,000–2,000 for a typical chat interaction, meaning leaders who delay infrastructure planning often end up managing budget overruns instead of business outcomes.
The report identifies four forces set to reshape the GCC operating model: building repeatable, AI-enabled workflows rather than isolated experiments; treating data readiness, compute, security, governance and economics as a single integrated decision; taking genuine ownership of products, markets and outcomes rather than supporting them remotely; and redesigning workforces so talent shifts from routine execution toward engineering, product and business problem-solving.
With 55 percent of routine GCC work already exposed to AI-driven automation and 60 percent of the workforce expected to need reskilling by 2030, the report argues the need is not incremental training but a fundamental rethink of how work itself is structured.
The report also offers a framework for deciding which AI workloads to own versus lease—recommending greater control for sensitive, regulated or business-critical workloads, while exploratory ones can run on flexible, leased infrastructure. It introduces a "Sovereign Sandbox" model allowing GCCs to test regulated or proprietary data in a contained environment before scaling to production.
