Trending Now - Aging NVIDIA V100 GPU Outperforms Newer Consumer Cards in AI Tasks
WHO:
Wccftech - The NVIDIA Tesla V100 GPU is an 8-year-old model.
VideoCardz.com - Hardware Haven conducted the testing of the NVIDIA V100 GPU.
WHAT:
Techmeme - Nvidia has committed over $40 billion to equity investments in 2026, including a $30 billion stake in OpenAI, $3.2 billion in Corning, and $2.1 billion in IREN.
TechRadar - Skymizer introduced a low power PCIe AI accelerator that runs 700 billion parameter language models locally, consuming just 240W using decade-old DDR4 and 28nm chips.
Wccftech - The NVIDIA V100 GPU is currently priced at around $100.
Wccftech - The V100 outperformed the 5-year-old RTX 3060 and the 3-year-old RX 7800 XT in AI LLM workloads.
Wccftech - The V100 produced around 130 Tokens/s in testing with GPT-oss, while the RX 7800 XT produced around 90 Tokens/s.
Wccftech - The V100 was 42% faster than the RTX 3060 in token generation speed.
VideoCardz.com - The V100 setup cost around $200, including the SXM2-to-PCIe adapter and cooling solutions.
VideoCardz.com - The V100 produced 95 tokens/s at a 100W power limit, while the RTX 3060 produced 68 tokens/s under the same conditions.
WHERE:
Wccftech - The V100 is primarily used in data centers.
WHEN:
Wccftech - The V100 was originally priced at over $10,000 when released.
VideoCardz.com - The full setup cost, including cooling parts, came to around $235.
WHY:
Wccftech - The testing aimed to evaluate the performance of older GPUs in modern AI tasks.
HOW:
Wccftech - The V100 was tested using a custom setup that included an SXM-to-PCIe adapter and a 3D-printed cooling solution.
VideoCardz.com - The V100 was installed onto a PCIe adapter card to enable use in a standard desktop system.
(UPDATED May 11, 2026 13:51 UTC)
Nvidia has made a significant commitment of over $40 billion to equity investments this year, which includes a $30 billion stake in OpenAI, along with $3.2 billion in Corning and $2.1 billion in IREN. In another tech development, Skymizer has unveiled a low power PCIe AI accelerator capable of running 700 billion parameter language models locally while consuming only 240 watts, utilizing decade-old DDR4 and 28nm chips. These advancements highlight the ongoing evolution in the tech industry as companies push the boundaries of artificial intelligence and hardware efficiency.
--
(UPDATED May 11, 2026 11:57 UTC)
The NVIDIA Tesla V100 GPU, an eight-year-old model, is currently priced at around $100 and has shown impressive performance in AI workloads, outperforming both the five-year-old RTX 3060 and the three-year-old RX 7800 XT. In testing, the V100 achieved approximately 130 Tokens per second with GPT-oss, significantly surpassing the RX 7800 XT's 90 Tokens per second and being 42% faster than the RTX 3060 under similar conditions. This testing, conducted by Hardware Haven, aimed to evaluate how well older GPUs can handle modern AI tasks, utilizing a custom setup that included an SXM-to-PCIe adapter and a 3D-printed cooling solution.