Closer is faster
Local Bay Area POPs cut round-trip time dramatically
Run training and inference on NVIDIA H100 GPUs over our low-latency Bay Area ring. Local, fast, and reliable – designed for teams that need real-time
Regional RTT
Interconnected by Fiber
Arista CVP + diversity

NG Compute
$1,950-$2,700 / 30-day commit
NG XP Compute
$4,500-$4,800 / 30-day commit
Choose a commitment term that works for you and unlock increasing discounts and exclusive features.
30-day rolling
No Commitment
Quarterly commitment
Save on your deployment
Semi-annual commitment
Great value
Annual commitment
Maximum savings
Pricing targets the $3-$8/hr sweet spot to outperform national GPU clouds on latency and total cost of results.
Private Connectivity? We can extend a low-latency wave from our ring to your office or DC for line-rate access to GPUs. Ask about dedicated links and security zones.
*Specs and pricing subject to capacity & availability. NVIDIA® and H100® are trademarks of NVIDIA Corporation.