Brijesh Desai 
India is going big on artificial intelligence with a new 100-kilometer data city planned for Visakhapatnam—10 GW AI hyperscale hub with ₹1,000 crore Phase 1 investment.
India is going big on artificial intelligence with a new 100-kilometer data city planned for Visakhapatnam, Andhra Pradesh IT Minister Nara Lokesh confirmed this week. Spanning a staggering 100 sq.km radius ecosystem, the project promises 10 GW compute capacity powered by dedicated nuclear energy, positioning Vizag as Asia’s AI hyperscale capital and Greater Bay Area rival. Phase 1 kicks off with ₹1,000 crore investment, scaling to ₹5,000 crore total by 2030.
I’ve followed India’s AI push since the 2023 national mission, but this feels different—continental ambition matching China’s Shenzhen or US Abilene clusters. Lokesh calls it “whole nine yards” infrastructure: data halls, subsea cables, green hydrogen, BARC R&D. Google’s largest non-US AI hub validation seals the deal.
Massive Scale & Specs
Core Infrastructure:
- 10 GW nuclear-powered capacity (surpasses Mumbai’s 600 MW)
- 5M rack units hyperscale data centers
- 1M+ NVIDIA H200/B200 GPUs
- Direct US-Europe subsea cable landings
- 100% renewable baseline
Anchor Investments:
• Google: Largest AI hub outside US ($15B+)
• Meta/Sify: 500 MW Waterworth cable landing
• Reliance/Brookfield/Digital Realty: $11B JV
• AdaniConneX: Hyperscale colocation
Strategic Location Advantages
Visakhapatnam Trifecta:
Geography: Deep-sea port, subsea cable hub
Power: 10 GW nuclear allocation (Kudankulam proximity)
Land: 500 sq.km available (industrial zones)
Talent: IIITVizag, IIT nearby English-speaking engineers
Cooling: Natural seawater immersion
Economic Impact Projections
Job Creation:
- Direct: 50,000 high-skill AI/data roles
- Indirect: 200,000 construction/ops
- State GDP boost: +15% Andhra Pradesh
National Targets:
- $50B AI GDP contribution by 2030
- 10,000 AI startups
- 1M AI professionals trained
- Top-5 global AI nation status
Phase 1 Roadmap (2026-2028)
Q3 2026: Land/power substation acquisition
Q1 2027: 500MW first data hall operational
Q4 2027: Subsea cables live (Google/Meta)
Q2 2028: 2 GW Phase 1 complete
Global Competition Context
| Hub | Capacity | Investment | Status |
|---|---|---|---|
| Visakhapatnam | 10 GW | $12B | Phase 1 2028 |
| Guangdong China | 8 GW | $20B | Operational |
| Abilene Texas | 6 GW | $10B | Live |
| Singapore Tuas | 2 GW | $5B | Live |
- 100% FDI automatic approval
- 15-year tax holiday for AI infra
- Carbon-neutral mandate
- Sovereign data localization
- English-first talent pipeline
Enterprise Anchor Commitments
Reliance Jio: 2 GW AI cloud services
Tata Digital: Enterprise AI platforms
AdaniConneX: Hyperscale facilities
Microsoft Azure: Regional sovereignty cloud
NVIDIA: DGX SuperPOD clusters
Technical Differentiators
Networking: 400Gbps intra-campus fibre
Storage: 100 EB NVMe capacity
Cooling: Seawater immersion + liquid
Connectivity: Direct SG-US-EU cables
Security: Biometrics, 24/7 monitoring
RAND AI Power Report 2025 Context
- Vizag projected top-5 global AI capacity
- Current India: 21 GW trending 100 GW by 2030
- Vizag surpasses Mumbai’s 600 MW overnight
- 5-6% global data center contribution
India’s Visakhapatnam data city catapults nation into AI hyperscale league. 10 GW sovereign compute rivals nation-states while attracting $26B private investment. Phase 1 2028 delivery makes India unavoidable for global AI workloads.