Yotta NM1 Navi Mumbai
- Uptime Institute Tier IV certified
- Dual utility feeds from separate substations
Hopper · SXM
The H100 is the workhorse of AI training, and many Indian teams already own them. NVIDIA has stopped taking new orders, so we host existing H100s and quote H200 or B200 for new capacity.

Orders closed Lead time, October 2026: Not orderable · All GPU lead times
Good fit for

The hardware
8-GPU HGX server, 640 GB of GPU memory. 80 GB HBM3 per GPU, up to 700 W per GPU, 5u–8u, by chassis of rack space.
Hosting
H100 HGX servers are air-cooled but dense. We confirm the rack’s power budget, airflow and network uplinks for your exact chassis before it ships.
How managed hosting worksNVIDIA rates the DGX H100 at up to 10.2 kW. OEM HGX servers are in the same range.
No. NVIDIA is no longer taking new H100 orders. For Hopper, the H200 is the current option, with 141 GB per GPU. For more throughput, B200.
Rarely well. At around 10 kW it needs dedicated power, serious cooling and noise isolation that most offices don’t have.

Rack, power, cooling, network and remote hands for servers you own.
Learn more →
Lender-ready quotes and asset records for your bank or NBFC.
Learn more →
kW, amps and tons of cooling for every GPU class.
Learn more →
Tested, valued and sold, or traded up for newer GPUs.
Learn more →