NVIDIA Hopper, 80 GB HBM3, SXM5
The NVIDIA H100 SXM is the Hopper-generation flagship with 80 GB HBM3 memory and 3.35 TB/s of bandwidth. 3× faster than A100 on transformer workloads. FP8 Tensor Cores and NVLink 4.0 make it the standard for frontier model training and fast inference. Coming soon to packet.ai.
Pricing to be confirmed at launch
The H100 SXM combines 80 GB HBM3 with NVLink 4.0 for the fastest multi-GPU training fabric available.
FP8 Tensor Cores deliver 3× the throughput of A100 on transformer workloads. The current frontier training standard.
HBM3 delivers 68% more bandwidth than A100. Critical for large batch sizes and long context lengths.
900 GB/s NVLink for tight multi-GPU coupling. Essential for tensor-parallel training across 8+ GPUs.
Hardware-accelerated FP8 mixed precision with dynamic scaling. Native support in PyTorch and JAX.
Dedicated or monthly — plus multi-node clusters.
Full H100 SXM reserved exclusively for you. Zero noisy-neighbour risk, 99.99% SLA.
Join waitlist →Reserved H100 at a flat monthly rate. Full single-tenant isolation, predictable cost exclusively for you. 99.99% SLA, zero noisy-neighbour risk.
Join waitlist →Scale frontier training across multiple H100 nodes with NVLink 4.0 and InfiniBand interconnect.
FP8 Tensor Cores deliver 3× A100 throughput on transformer workloads.
3.35 TB/s HBM3 bandwidth enables high-throughput token generation.
NVLink 4.0 tightly couples H100s for efficient tensor-parallel training.
Hopper flagship: 80 GB HBM3, 3.35 TB/s, 3× A100 on transformer workloads. The current frontier training standard.
Coming soon to packet.ai. Join the waitlist to be notified at launch.
H100 SXM is ~3× faster on transformer workloads, has HBM3 vs HBM2e, supports FP8, and has NVLink 4.0 at 900 GB/s vs 600 GB/s.
30B at FP16 natively, 70B at 4-bit. For full FP16 70B, use H200 or B200.
Yes. Up to 7 isolated MIG instances for multi-tenant inference serving.
Join the waitlist for early access to NVIDIA H100 SXM on packet.ai.
On-demand · hourly billing · US & EU regions