NVIDIA Tesla V100
Hacker ApprovedRanked 2 of 11 datacenter GPUs by recommenders
People recommend the used V100 as cheap VRAM and FP64 compute for local AI and physics work. Caveats: no bfloat16, loud fans, high idle power.
People recommend the used V100 as cheap VRAM and FP64 compute for local AI and physics work. Caveats: no bfloat16, loud fans, high idle power.
20 people gave an opinion
- recommend it
- 16
- criticize it
- 4
- approval share
- 80%
Recommended in 8 of 13 months
- November 2025: not recommended
- December 2025: recommended
- January 2026: not recommended
- February 2026: not recommended
- March 2026: recommended
- April 2026: not recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 16 people recommend itNeeds 10 or more, met.
- 4 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 8 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Cheap VRAM per dollar
“V100S is a very good card for some workloads. And they were cheap as well.”
Comment from 26 Oct 2026 in “IBM CEO says open weight models are now good and cheap on-prem” “You can run the same model with less context on a single £200 V100”
Comment from 1 Jun 2026 in “I Put a Datacenter GPU in My Gaming PC for £200” “I think the V100 16GBs are a better deal at $250”
Comment from 13 Jul 2026 in “Benchmarking 15 "E-Waste" GPUs with Modern Workloads” - Comment from 2 Mar 2026 in “Zclaw – The 888 KiB Assistant”
- Comment from 3 Jun 2026 in “Gemma 4 12B: A unified, encoder-free multimodal model”
- Comment from 23 Sep 2026 in “GPT-6 Sol and Luna”
Fast local model inference
“I have a single V100 running around 900t/s prefill and 100 t/s decode during DeepSWE runs on 27B”
Comment from 2 Oct 2026 in “Show HN: Janus – Go binary that runs GGUF models via Vulkan on AMD/Intel/Nvidia” “already at around 900 t/s prefill, 90-100 t/s output in Pi on coding tasks”
Comment from 30 Sep 2026 in “The AI Race Just Got Awkward” “Get a couple or V100S GPU. They will run it /fast/ with the right engine.”
Comment from 27 Oct 2026 in “New iMac with M6” - Comment from 29 Aug 2026 in “Run Qwen3.8 27B locally: real numbers from my Mac Studio”
Strong FP64 for scientific computing
“perfectly capable of running and delivering the physics calculation optimizations”
Comment from 9 Jun 2026 in “AI Is Slowing Down” “V100s are also quite useful for many applications that use FP64”
Comment from 23 Jul 2026 in “Nobody knows what a used GPU cluster is worth” “the FP64 performance (7+ TFLOPS!) is fantastic given their age”
Comment from 31 May 2026 in “I Put a Datacenter GPU in My Gaming PC for £200” Scales across multiple cards
“8xV100 will outperform most of what you can buy for that price simply by way of memory and nvlink bandwidth.”
Comment from 10 Dec 2025 in “Mistral releases Devstral2 and Mistral Vibe CLI” “Get a couple or V100S GPU. They will run it /fast/ with the right engine.”
Comment from 27 Oct 2026 in “New iMac with M6” “combined them without an issue”
Comment from 1 Jun 2026 in “I Put a Datacenter GPU in My Gaming PC for £200”
What people warn about
Each mark is one comment
Lacks bfloat16 and modern data types
“the V100 doesn't support bfloat16”
Comment from 31 May 2026 in “I Put a Datacenter GPU in My Gaming PC for £200” “horribly inefficient, lacking modern data types and engines, requiring screaming server fans”
Comment from 22 Jul 2026 in “Nobody knows what a used GPU cluster is worth” High idle power and loud fans
“horribly inefficient, lacking modern data types and engines, requiring screaming server fans”
Comment from 22 Jul 2026 in “Nobody knows what a used GPU cluster is worth” “v100s idle super high! 25-35watt with nothing loaded and easily 50w when a model is loaded”
Comment from 31 May 2026 in “I Put a Datacenter GPU in My Gaming PC for £200”
Other datacenter GPUs people recommend
The four most recommended other datacenter GPUs, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| NVIDIA H100 | 20 | 16 | 56% | 11 | — | — |
| NVIDIA Tesla V100This page | 16 | 4 | 80% | 8 | — | Approved |
| AMD Instinct MI355X | 14 | 5 | 74% | 7 | — | Approved |
| Nvidia A100 | 10 | 4 | 71% | 8 | — | Approved |
| Intel D270 GPU AI Accelerator | 10 | 10 | 50% | 1 | — | — |
What people said
24 of the 27 comments that mention it, newest first. Each links to the original comment.
“Get a couple or V100S GPU. They will run it /fast/ with the right engine.”
“I just really like this as an option.”
“V100S is a very good card for some workloads. And they were cheap as well.”
“I have a single V100 running around 900t/s prefill and 100 t/s decode during DeepSWE runs on 27B”
“already at around 900 t/s prefill, 90-100 t/s output in Pi on coding tasks”
“I find it far from useless”
“they do all the work overnight”
“they are the cheapest way to get a 32GB gpu”
“So 32gb v100 sounds good right? Maybe...”
“that's _really_ cheap inference, and it's not a modified card”
“Nvidia drivers stopped a while ago (before V100 were considered old also for x86_64).”
“V100s are also quite useful for many applications that use FP64”
“horribly inefficient, lacking modern data types and engines, requiring screaming server fans”
“I think the V100 16GBs are a better deal at $250”
“perfectly capable of running and delivering the physics calculation optimizations”
“You can get SXM V100s for like $100 off ebay”
“You can run the same model with less context on a single £200 V100”
“combined them without an issue”
“the FP64 performance (7+ TFLOPS!) is fantastic given their age”
“the V100 doesn't support bfloat16”
“v100s idle super high! 25-35watt with nothing loaded and easily 50w when a model is loaded”
“the 32GB V100 seems like a pretty good VRAM/$”
“8xV100 will outperform most of what you can buy for that price simply by way of memory and nvlink bandwidth.”
“you would be using used v100 Teslas and not 4 year old single chip gaming GPUs”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.