Extropic launches Z1T on sparse probabilistic hardware, claims up to 140x GPU energy efficiency
beffjezos · x · 2026-09-06
- Extropic founder/CEO beffjezos (Guillaume Verdon) unveiled Z1T in a full MTS interview: the first family of transformer-like models built for sparse probabilistic hardware.
- Claims up to 140x the energy efficiency of a GPU at chip level, 100x at data center level.
- Unlike dense GPU models, Z1T is sparse; a pbit uses 10,000 fewer transistors.
- Proposes a new scaling law: scale connectivity, not just flops.
- Wins are concentrated in decode, not prefill; long-term vision is sun-powered chips in space.
Related event: Extropic Unveils Z1T, Claiming Up to 140x GPU Energy Efficiency(7 posts)→
More from Infra
- Corporate token spending growth collapses from 42% to 0.2% as hyperscaler ROIC falls to 20% — SumitGup · 2026-10-21
- Qualcomm plans worldwide performance-based layoffs as it chases $5B AI data center revenue — rohanpaul_ai · 2026-10-21
- vLLM at PyTorchCon 2026: KV cache and kernel optimizations across TPU, AMD, and NVIDIA — PyTorch · 2026-10-21
- antirez to port DeepSeek Infer kernels to DGX Spark's SM121 for DwarfStar — antirez · 2026-10-21
- Ternary-Bonsai-2-9B: 2-bit ternary vision-language model trends on Hugging Face for on-device use — prism-ml · 2026-10-21
- General Compute x Positron: GPUs Handle Prefill, Memory-First Chips Tackle Decode-Bound LLM Inference — cneuralnetwork · 2026-10-21