GeForce RTX 5080 16GB
16 GB GDDR7 at 360 W. The lot-of-20 card for render farms and small inference fleets.
In stock. Lead time 72 hours.
Volume pricing
| Units | Discount | Unit price |
|---|---|---|
| 1 unit | List | $1,199 |
| 2 to 7 units | 3% off | $1,163 |
| 8 to 31 units | 6% off | $1,127 |
| 32+ units | 10% off | $1,079 |
Unit price at 1: $1,199
Request a bulk quote
The full sheet.
Figures from the manufacturer datasheet. Price is an indicative street price, shown as from.
- VRAM
- 16 GB
- Memory
- GDDR7
- Bandwidth
- 960 GB/s
- FP8
- 0.9 PFLOPS
- FP4
- 1.8 PFLOPS
- Board power
- 360 W
- Form factor
- PCIe dual-slot
- NVLink
- No
- Architecture
- Blackwell
- GPUs
- 1
- Per GB of VRAM
- $74.94
- Availability
- In stock, ships 72h
- Sale
- Buy online
- SKU
- RTX-5080
What it is for.
Workloads, sizing and what ships in the box.
16 GB GDDR7, priced for lots of 20.
16 GB of GDDR7 at 0.96 TB/s, 0.9 PFLOPS FP8, 360 W on a dual-slot partner board. Built for render farms and small inference fleets where you buy twenty, not one. 3 year warranty handled by us. In stock, ships in 72 hours. $1,199, about $75 per GB, the lowest in the catalog. 8 units save 6%.
Good for
- Llama 3.2 3B at FP16 (7 GB with headroom) at very high batch.
- Llama 3.1 8B at FP8 (10 GB).
- Mistral Small 24B at 4-bit (14 GB), short context.
Fit rule: FP16 weights take 2 bytes per parameter, FP8 1 byte, 4-bit 0.5 bytes, plus 20% headroom for KV cache at short context.
Pairs with
- GeForce RTX 5090 32GB, for the heavier nodes in the farm.
- L40S 48GB, when the job moves to a datacenter rack.
Partner boards.
How much VRAM per box.
Card count is capped by slots and by the chassis power budget.
Parts bought alongside.
Same workload, a different trade between VRAM, bandwidth and price.