AI in 2026: Why Sovereignty, Governance and Hybrid Cloud Will Define the Next Enterprise Shift

With rise in technology in 2026, enterprises no longer experiment with artificial intelligence rather it has become the backbone of business transformation. With automation at scale, several organisations are rethinking how data is governed, secured and activated at large. More than two-thirds of the surveyed executives worry that their AI efforts will fail due to poor integration with business operations and data, pointing to gaps in oversight, model accountability and necessary risk controls.

In this conversation, FE Futech spoke with Rishi Aurora, Managing Partner, IBM Consulting India & South Asia, about how AI is moving from experimentation to enterprise-scale execution, reshaping large business models across diverse industries. (Edited Excerpts)

Q1. How can enterprises rethink their AI strategies in 2026?

NITI Aayog expects that AI could contribute nearly USD 1.7 trillion to the Indian economy by 2035 – our nation is not only accelerating consumption but positioning itself as a global AI innovation leader. We are also seeing a global shift towards digital sovereignty - India’s own DPDP Act being a great example of this trend toward digital autonomy. 

In 2026, enterprises need to rethink their AI strategies through the lens of digital sovereignty, prioritising control over their data, operations, and technology. Sovereignty is no longer just about meeting regulations—it’s becoming a core part of business strategy. With hybrid cloud, AI, and workflow-centric transformation, companies now have the tools to turn this shift into a genuine competitive advantage.

Hybrid cloud blends public, private, and sovereign cloud environments in a way that gives enterprises the control they need, without losing access to global innovation. As AI works with sensitive, high-value data, sovereignty becomes an architectural requirement - data lineage, model governance, and accountability need to be built in right from the start. Unlocking enterprise data is equally important, given that 99% of enterprise data remains untapped and most of it is unstructured. AI must run where the data lives—at the edge, in sovereign clouds, or on-prem—to stay compliant and scalable. Finally, we’re seeing a huge shift toward workflow-centric transformation - enterprises are embedding AI agents directly into core workflows for procurement, HR, IT security, customer service, and so on, which enables action driven automation. 

We’ve done this within IBM itself – what we call the IBM client zero story - reinventing over 70 workflows leveraging AI and automation to drive $4.5 billion in productivity, with the savings reinvested in the R&D of new frontier technologies, solutions, and sales. 

Q2. What are the key challenges that enterprises continue to grapple with in their AI journeys?

One of the biggest challenges enterprises face today is a vision gap. IBM’s Enterprise in 2030 report revealed that 79% of executives expect AI to significantly contribute to their revenue by 2030, but only 24% can clearly see what the main sources of revenue will be - revealing unclear pathways as enterprises move from experimentation to impact. This also reflects critical talent shortages, with enterprises lacking the skills to integrate, govern and scale AI effectively.

Second, governance remains a huge challenge - many enterprises are also struggling with integrating AI into their core workflows, often leading to poor integration with business operations and data, pointing to gaps in oversight, model accountability and necessary risk controls.

However, Indian enterprises don’t need to solve these challenges in isolation.For example, banking’s rigor in governance, validation and human-in-the-loop systems can guide safety-critical sectors like healthcare or citizen services. By embracing cross-sector learning, investing in talent transformation and strengthening governance, enterprises can turn their AI ambitions into measurable impact for sustained growth.

Q3. SLMs vs LLMs – what is the way forward for businesses?

The future will not be about choosing between SLMs and LLMs – rather, building multi-model AI architectures. As per IBM’s Enterprise in 2030 report, 82% of respondents expect their AI capabilities to be multi-model by 2030, and 72% expect SLMs to surpass LLMs in usage. This shift is driven by how AI can create value inside the enterprise. 

SLMs offer efficiency, lower cost, and easier governance—making them ideal for domain-specific tasks, regulated environments, and workflows requiring tight control. LLMs remain powerful for open-ended reasoning and broader knowledge tasks. For example, a bank may leverage an LLM for a general chatbot, but for fraud detection or KYC processing, an SLM is the winner. It’s faster, cheaper, and keeps sensitive data within the bank’s sovereign private cloud. 

In this context, we have engineered the IBM Granite 3.0 models —specifically the 8B and 2B variants—to be enterprise "workhorses." They are instruct-tuned to excel at dense, domain-specific tasks while delivering 3x to 23x lower costs compared to larger frontier models on similar tasks. By running Granite 3.0 on a hybrid cloud stack, enterprises can get the performance of a giant with the privacy and agility of a specialist.

Q4. How can India transition from rapid AI adoption to true AI autonomy, with control over data, models, and governance?

To move from AI adoption to true AI autonomy, India needs to get three dimensions right – data sovereignty, operational sovereignty and technology sovereignty. Data sovereignty under the DPDP Act means that citizen data must stay under strict consent and purpose controls. For example, a state health system that keeps all medical records within India and logs every access. 

Operational sovereignty entails ensuring that systems that run essential public services remain fully under local authority. Imagine a city’s water-supply management system running on a sovereign-by-design platform – so all operational logs, sensors, alerts and AI predictions stay within local jurisdiction. 

And technology sovereignty means building India’s own indigenous, trusted models. Like how India’s Bhashini project builds indigenous models that understand regional dialects, ensuring critical communication stays independent of foreign proprietary software.

Q5. What is the role that consulting partners like IBM play in this journey?

IBM Consulting differentiates itself by blending industry expertise with hybrid cloud and AI, using the IBM Garage methodology to co-create and scale innovations rapidly. We don't just advise; we co-implement. Across BFSI, telecom, and government, we drive digital sovereignty through the IBM Consulting Advantage platform. This AI-powered library of agents and methods ensures faster, more secure, and transparent delivery. For instance, with Vodafone Idea, we are modernizing stacks for AI-driven resilience. 

Similarly, by supporting the design of BharatGen’s architecture, embedding responsible AI guardrails, and ensuring the ecosystem scales securely and compliantly, we are helping India build trusted, indigenous AI capabilities that keep sensitive data within the country, while enabling truly transformative innovation for impact.

Whether modernizing core banking or digital public infrastructure, we act as an end-to-end partner, ensuring India’s institutions transition confidently into a sovereign, self-reliant digital era.