InferenceX Research
Benchmark write-ups on agentic inference, AgentX results, and chip and serving-stack economics.
New to the terminology? Browse the AI inference glossary.
6 articles
10 min read
Jensen Sandbagging Performance Again, 2x more Annual Profit Per GigaWatt, The More you Buy, The More you Earn, AgentX, InferenceX, Extreme Co-Design
3 min read
The Dynamo router becomes the bottleneck, no submission runs context parallelism, and AMD leaves KV offload on the table
23 min read
Kimi K3's architecture: compressed memory, attention across depth, latent expert routing, and serving performance
20 min read
Rubin LUT Based Tensor Core, Feynman, Rack Scale, Perf Per MegaWatt, Perf Per Dollar, Software Improvements, Public Rubin Software, PyTorch, vLLM, OpenAI Triton
10 min read
DSv4-Pro FP4 8K/1K, Dynamo+vLLM, disaggregated on both racks. GB300's 50% extra HBM (288 vs 192 GB/GPU) unlocks a wider prefill+decode recipe GB200 can't fit — lifting middle-of-curve perf/$ by 2.31x despite a 20% per-GPU TCO premium.
8 min read
DeepSeek R1 FP4 1k/1k. NVL72's 72-GPU NVLink scale-up fabric lets decode run wide EP up to EP=32, where B200's 8-GPU NVLink island caps out at EP=8 over RoCEv2
