Tracked topic
… Step Aside NVIDIA & Huawei, China's DeepSeek Prepares Its Own AI Inference Chip DeepSeek, the leading AI firm which is based in China, has started developing its own AI chip. …
… Reuters reports that DeepSeek aims to reduce its reliance on NVIDIA and Huawei, obtaining more control over its hardware. While the plan is in its early stages, DeepSeek has already begun hiring talented chip design engineers in recent months for the same reason. …
… DeepSeek's Wenfeng Says Huawei's Atlas 950 SuperPoD can fully substitute NVIDIA's GB300 NVL rack, but for every GB300 GPU DeepSeek needs 4x Ascend GPUs DeepSeek's Wenfeng disclosed a slew of valuable information in a recent investor conference call, noting that the AI lab's available compute only a… …
… Just hours later, however, DeepSeek launched a refreshed version of its latest Flash-class model, dubbed the V4 Flash 0731. …
… NVIDIA Cost Per Token Narrative Sees Massive Gain In DeepSeek V4 As AI Model Sees 5x Boost On Blackwell GPUs With Continued Optimizations "Cost Per Token" is the fundamental metric for AI TCO, as NVIDIA highlighted this a few months back, and now, the company is delivering the lowest-ever token cos… …
… NVIDIA Blackwell NVFP4 Architecture Delivers Major Speed-Ups In DeepSeek v4 With More Optimizations On The Way With the launch of DeepSeek V4, we saw some major optimizations in compute & memory requirements. …
… On the Artificial Analysis Intelligence Index, DeepSeek V4 Flash 0731 leads with a score of 50 percent , while Inkling-Small trails slightly at 40 percent, matching the older baseline DeepSeek V4 Flash variant. …
… How DeepSeek V4 Slashes Compute and Memory Costs In its release notes for DeepSeek V4, DeepSeek outlines that the new model is capable of using just 27% of single-token inference FLOPs and 10% of key-value KV cache when running a one-million-token context window. …
… My math yields 22.4% MFU tho https://t.co/HD3iNLFxbA pic.twitter.com/TeZexTHyby — Teortaxes▶️ DeepSeek 推特🐋铁粉 2023 – ∞ @teortaxesTex January 25, 2025 Of course, DeepSeek claims that its R1 model was trained on just 2000 H800 GPUs. …
… It would be interesting to see how DeepSeek's R2 AI model, which is expected to debut in a few weeks, turns out. …
