Hardware: GPU servers
HGX B200 SXM6: 8-GPU Server System
First-generation Blackwell. 180 GB HBM3e per GPU, NVLink 5, NDR800. New-build only, no secondary market yet.
Live lots: available now
Currently available HGX B200 systems
New build via OEM factory order only. Prices shown against the Rack Rate Index median, deal score is automatic.
8×B200 SXM6 HGX — OEM, factory sealed
8× B200 SXM6 180GB · liquid-cooling ready · NDR800 fabric
Market pricing: Rack Rate Index
What HGX B200 systems trade at
| Series | Min | Median | Max | N | MoM |
|---|---|---|---|---|---|
| New build, factory order | $540,000 | $565,000 | $590,000 | 7 | stable |
Full server system pricing: chassis, NDR networking, host CPUs, and memory included. Indicative market ranges, updated Jul 18, 2026. Verified series publish once they clear our N-quote threshold: median of verified quotes, supplier-blind.
System specifications
HGX B200: full system specs
| GPUs | 8× NVIDIA B200 SXM6 |
| GPU memory | 180 GB HBM3e per GPU · 1.4 TB total |
| Memory bandwidth | 8.0 TB/s per GPU · 64 TB/s total |
| GPU interconnect | NVLink 5 · 1.8 TB/s bisection · NVSwitch |
| FP4 (system) | ~144 PFLOPS dense |
| FP8 (system) | ~72 PFLOPS dense |
| GPU TDP | 1,000 W per GPU |
| System power draw | ~14–16 kW · plan 18–20 kW with PUE |
| Host platform | 2× Xeon SP or EPYC · 2 TB DDR5 · PCIe Gen 5 |
| Common chassis | Supermicro SYS-821GE-TNHR · Dell XE9780 |
| Network fabric | 8× ConnectX-8 800G NDR800 InfiniBand |
| Form factor | 10U, increased power and cooling vs H100 |
Buyer decision guide
Should you buy a HGX B200?
Buy HGX B200 if
Building a frontier training cluster
B200's 4× FLOPS improvement over H100 and 2.25× memory bandwidth justify the price for large-scale training. The roadmap is Blackwell.
Buy HGX B200 if
You can absorb 14–20 week lead time
New Blackwell is the right call if you're planning a cluster that runs for 3+ years and timing isn't the constraint.
Consider alternatives if
Consider H100 for immediate capacity
H100 deploys in 1–3 weeks at $250K/system. If you need GPUs now, the secondary market wins on speed.
HGX H100 pricing →Used vs refurb: what the premium buys
| Factor | Used, tested | Refurb, warranted |
|---|---|---|
| Warranty | None / seller DOA window | 90-day seller warranty + DOA |
| Burn-in report | Available (ask for it) | Included and certified |
| Price premium | — | +8–12% over used median |
| Right for | Buyers who can evaluate burn-in logs | Teams who want turnkey assurance |
Pre-purchase verification
What to verify before buying
At this price point, skipping due diligence is expensive. Every item below should have a documented answer before you wire funds.
- Verify OEM factory order is confirmed, not just quoted
- Lock allocation terms in writing: delivery window, cancellation terms
- Power infrastructure: B200 needs ~14–16 kW per node vs ~10 kW for H100
- Confirm NDR800 switch compatibility if expanding an existing NDR cluster
Export control notice
B200 carries advanced export controls, stricter than H100/H200. Verify end-user and destination against current BIS entity lists before placing an order.
Looking to buy HGX B200 systems?
Submit your requirements: quantity, condition, timeline, and budget. We benchmark against the Index and match you with verified supply.
Have HGX B200 inventory to sell?
List your lot: describe what you have, condition, and quantity. We benchmark against the Index and surface it to qualified buyers.
Compare models