# NVIDIA’s 9-year-old Tesla V100 outperforms M5 Ultra in AI benchmarks with four GPUs

Digest AI · Hardware & Compute · published 2026-10-13T04:03:55Z

Canonical: https://digestai.news/story/nvidias-9-year-old-tesla-v100-outperforms-m5-ultra-in-ai-benchmarks-wi

## Summary

A YouTuber named Jun Yamog tested four NVIDIA Tesla V100 GPUs—each with 32GB of HBM2 memory—mounted on an NVLink carrier board to compare their AI inference performance against modern hardware. The setup delivered up to a 230% speed increase over Apple’s M5 Ultra and a 167% boost over NVIDIA’s DGX Spark cluster in tasks like processing long prompts with Qwen3.8-Flash-Next (4-bit quantized). Yamog modified an A100 heatsink to handle the V100s’ 300W power draw per chip, connecting them via PCIe x8 adapters for the test.

The results show the V100’s enduring relevance for inference workloads, though Yamog notes Apple’s advantage stems from software optimizations. Another user previously paired a V100 with an RTX 4080 to run 27B-parameter models locally. Used V100s with 32GB VRAM now sell for around **$744 each** on eBay, offering a cost-effective alternative to cloud-based AI services for hobbyists or small-scale deployments.

## Key points

- Four NVIDIA Tesla V100 GPUs on an NVLink board outperformed Apple’s M5 Ultra by up to 230% in AI benchmarks
- Jun Yamog modified an A100 heatsink to cool four V100s, each drawing 300W, for the test setup
- Used 32GB V100s cost **$744 each** on eBay, enabling local AI inference without cloud fees

## Why it matters

The results highlight how older NVIDIA GPUs like the Tesla V100, when repurposed with modern interconnects, can still compete with newer hardware for AI inference. This matters for cost-conscious users building local AI systems without relying on cloud providers.

## Sources

1. [NVIDIA’s 9-Year-Old Tesla V100 Still Packs Some Inference Muscle; 4 GPUs On An NVLink Board Can Outperform M5 Ultra, DGX Spark In Various AI Benchmarks By Up To 230%](https://wccftech.com/nvidia-tesla-v100-inference-m5-ultra-dgx-spark-ai-benchmarks) (wccftech.com, 2026-10-13)

## Cite

Digest AI, "NVIDIA’s 9-year-old Tesla V100 outperforms M5 Ultra in AI benchmarks with four GPUs", 13 October 2026, https://digestai.news/story/nvidias-9-year-old-tesla-v100-outperforms-m5-ultra-in-ai-benchmarks-wi

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/nvidias-9-year-old-tesla-v100-outperforms-m5-ultra-in-ai-benchmarks-wi.json