NovaCore Becomes India's First GPU Cloud to Deploy NVIDIA Blackwell
Today we're announcing that NovaCore has become India's first GPU cloud platform to deploy NVIDIA Blackwell servers. Our Hyderabad cluster is live and available for remote deployment worldwide.
Why Hyderabad
Hyderabad's datacenter ecosystem offers something unique: an elite power grid, competitive electricity costs, and proximity to one of the largest technical talent pools in the world. These aren't abstract advantages — they translate directly into lower operating costs and faster support response times for our customers.
India's datacenter costs run well below the U.S. equivalent. As our co-founder Ranbir Badwal puts it: "What some call an almost datacenter bubble in India has kept hosting costs well below the U.S." For AI teams running large training jobs over weeks or months, that cost difference compounds significantly.
The hardware
Central to our platform is the NVIDIA B200 — Blackwell's successor to the H100 and H200 generation. The numbers speak for themselves:
- 2.3x higher peak performance over prior generations
- 2x real-world AI speed compared to H100/H200
- 192GB HBM3e memory with 8TB/s bandwidth
- Up to 3x faster training and 15x inference throughput
- 25x greater energy efficiency than previous hardware
What we're offering
We're building NovaCore for the teams that need serious compute but don't want to navigate the complexity and cost of U.S.-only hyperscalers. That means:
- Bare metal GPU access with no virtualization overhead
- Custom configurations tailored to your workload
- Pre-built templates for major ML frameworks
- Free trials for qualifying startups and research labs
- Dedicated engineer support
What's next
This deployment is our first step. We're actively expanding capacity and will have more to share in the coming weeks about partnerships and new GPU configurations.
If you're running AI workloads and want access to Blackwell hardware at competitive rates, get in touch.