# gpu

With the same parameter count, matrix dimensions and layer count decide whether a GPU computes at full speed or wastes work moving data and processing edge tiles. Near-square dimensions aligned to 128, 256, or 512—and often wider, shallower models—fit hardware better without sacrificing accuracy.

SemiAnalysis benchmarked ~1,000 GPUs across NVIDIA and AMD lineups. GB300 NVL72 hits 100x over H100 — Jensen's 30x was an underestimate. AMD FP8 competes, but FP4+disagg+wideEP combo falls apart in software.

SemiAnalysis reports that NVIDIA GPU rental prices are rising rapidly again and capacity is being sold out. Compared to mid-2024 through Q3 2025, customers now have much less negotiating power with Neoclouds — driven by surging agentic coding demand and rising DRAM prices.

Everyone stares at the first 5 years of AI compute profits. But the real game is decided by what happens after the contract expires.