Indian enterprises rethink the cloud after a costly 2025

For much of the past decade, cloud adoption in Indian enterprises followed a simple rule: move fast, move everything. In 2025, that rule quietly began to break down. Across banking, telecom, retail, manufacturing, and the public sector, senior technology leaders began reassessing whether public cloud still made sense for every workload. The result was not a cloud exit, but a selective pullback that is now shaping enterprise technology strategies for 2026.

This recalibration was driven less by ideology and more by economics. As AI and analytics workloads scaled, cloud costs became harder to predict. CIOs and CFOs found that constant GPU usage, data movement charges, and consumption-based pricing were eroding the business case that originally justified aggressive migration.

“Hybrid cloud has moved from being a transitional architecture to the default enterprise operating model,” said Mayank Chaturvedi, Hybrid Cloud Solutions Sales Leader at HPE India. According to him, AI-driven use cases place new demands on data proximity, performance, security, and governance—requirements that traditional cloud-first strategies were not designed to address fully.

AI workloads changed the economics

Artificial intelligence proved to be the turning point. While public cloud platforms remain effective for experimentation and model training, inference workloads tell a different story. Always-on AI systems used in fraud detection, document processing, customer analytics, and operational forecasting created steady compute demand, making long-term cloud usage expensive.

Several industry leaders said Indian enterprises underestimated these operating economics. “Inference costs caught many organisations off guard,” said Srividya Kannan, founder and CEO of Avaali Solutions. “When AI runs across thousands of daily transactions, the compute and data movement costs quickly add up.”

“Many organisations did not anticipate how quickly AI-related cloud costs could scale without strong governance frameworks,” said Saloni Jain, co-founder of Plus91Labs. She pointed to Gartner projections showing global public cloud spending crossing $723 billion, even as nearly half of enterprises cite unpredictable cloud costs as a growing concern. In India, she said, this reality has accelerated the shift toward hybrid and multi-cloud strategies to balance performance, compliance, and financial sustainability.

Anant Adya, EVP and Service Offering Head at Infosys, said the discussion is also being shaped by trust and sovereignty. As AI adoption accelerates, enterprises are reassessing where data and inference should reside, leading to greater interest in localized cloud zones and sovereign-by-design architectures.

Regulation, predictability, and performance take centre stage

Regulatory pressure added another dimension in 2025. As enforcement of India’s data protection framework became more tangible, enterprises handling financial, healthcare, and citizen data began favouring environments where auditability, access control, and data residency could be demonstrated more clearly.

“The dialogue in India has definitely moved past the speed-to-migrate phase,” said Peter Lees, Head of Solutions for Asia Pacific at SUSE. He noted that enterprises are now focused on correcting “accidental architectures” created during rapid lift-and-shift migrations, particularly where steady-state workloads ended up attracting perpetual subscription costs without delivering proportional value.

Cost predictability has emerged as a common theme. “Cloud strategy is no longer owned by IT alone,” said Ritwik Batabyal, CTO and Innovation Officer at Mastek Group. CFOs, regulators, and business leaders are increasingly shaping workload placement decisions, especially for core systems such as ERP, billing platforms, and internal analytics.

A broader industry reset

For some leaders, the reassessment is also about efficiency. Ankush Sabharwal, founder and CEO of CoRover, said many heavy workloads were moved to the cloud without architectural optimisation. With advances in lightweight AI models, edge computing, and “AI-in-a-box” architectures, enterprises can now deploy applications closer to data sources, reducing latency, cost, and energy use. This does not signal reduced cloud adoption, he said, but more efficient consumption.

Across the industry, cloud re-evaluation has become a multi-stakeholder effort. CIOs and CTOs lead architecture discussions, CFOs scrutinise cloud spend, regulators influence data placement, and CEOs demand predictable economics tied to business outcomes. As Sabharwal noted, cloud strategy now spans business, risk, and innovation functions.

2026 outlook: cloud-first gives way to cloud-smart

Looking ahead, Indian enterprises are unlikely to abandon the cloud. Instead, cloud is becoming a choice rather than a default. Public cloud will continue to power experimentation, scaling, and customer-facing innovation. Predictable, regulated, and data-heavy workloads are expected to remain in private or hybrid environments.

“What replaces cloud-first is a more deliberate, cloud-smart approach,” said Rahul Jain, co-founder and CEO of Pixeldust Technologies, where workload placement is driven by economics, compliance, and control.

Cloud is no longer just infrastructure but an intelligent operating fabric that must align with AI, sustainability, and trust. After the lessons of 2025, Indian enterprises enter 2026 with a clearer view: cloud remains central to digital strategy, but control over cost, data, and outcomes has become equally strategic.