Microsoft Opens Fourth India Cloud Region in Hyderabad, Betting Big on AI Infrastructure
Hyderabad, India — September 22, 2026 — Microsoft has formally launched its India South Central cloud region in Hyderabad, Telangana, marking the company’s fourth hyperscale datacenter footprint in the country and its most significant bet yet on India’s artificial-intelligence economy.
Unveiled at the Microsoft AI Infrastructure Summit India on Monday, the new region is designed from the ground up for AI workloads, featuring three availability zones and sovereign-ready controls aimed at regulated industries such as banking, financial services and digital commerce.
“Demand is not the issue. We need to get the supply,” said Puneet Chandok, President of Microsoft India and South Asia, addressing the summit. He noted that India generates roughly 20 percent of the world’s data but accounts for only 3 percent of global datacenter capacity — a gap Microsoft is racing to close.
Built for the AI Era
The Hyderabad region joins Microsoft’s existing Indian cloud footprints in Pune (Central India), Chennai (South India) and Mumbai (West India), giving the company what it describes as the largest hyperscale cloud presence in the country.
Microsoft said the region became generally available on August 6, but the formal launch at this week’s summit underscored its strategic importance. The facility is engineered with resilient design aligned to India’s regulatory and seismic-zone requirements, and is already serving early-access customers including Adani Group, Bajaj Finserv, HDFC Bank and PB Pay.
For HDFC Bank, the new region doubles as a dedicated disaster-recovery site. “Having a dedicated DR region perfectly complements our Central India footprint, strengthening our business continuity and cross-cloud resilience,” said Atish Bhanushali, Senior Vice President and Head of Cloud Infrastructure at HDFC Bank. “With its ample capacity and access to the latest compute capabilities, it will support our efforts to accelerate our AI initiatives.”
Sovereignty, Security and Sustainability
With enterprise AI adoption shifting from pilots to production, Chandok stressed that Indian organisations want assurances that their data and AI processing remain within the country’s borders. Microsoft is offering sovereign public and private cloud options under a “your data, your control, your keys” principle, and confirmed it does not train foundation models on commercial customer data.
The Hyderabad campus also advances Microsoft’s environmental commitments. The facility uses next-generation air-cooled chillers that consume effectively zero water for cooling, and the company plans to create 5 million litres of groundwater recharge capacity locally.
Globally, Microsoft says it matched 100 percent of its annual electricity consumption with renewable energy in February 2026, and in India it has contracted more than 1,000 megawatts of new solar, wind and hybrid projects.
The $20.5 Billion Bet
The Hyderabad region is part of a cumulative $20.5 billion investment commitment Microsoft has made toward India’s cloud and AI future, including a $3 billion pledge announced in January 2025 and a further $17.5 billion in December 2025.
Industry analysts say the timing is critical. IDC forecasts India’s public-cloud services spending will reach $45.7 billion by 2030, growing at 22.2 percent annually, with AI spending expanding at roughly double that rate.
“With enterprise AI adoption moving from pilots to production at pace, proximity, sovereignty and resilience are no longer differentiators but baseline requirements,” said Dr. William Lee, Senior Research Director at IDC.
Microsoft also highlighted the human-capital dimension of its India push: the company has committed to equipping 20 million Indians with AI skills by 2030 and says it has already trained 5.6 million people since January 2025.
With the Hyderabad region now live, Microsoft is signaling that India is not merely a fast-growing market, but a strategic hub where the next wave of global AI infrastructure will be built and scaled.
