Part.01
Part.02
Part.03


[ARK Invest Big Ideas 2026](https://ark-invest.com/big-ideas-2026/):原始报告,包含ARK引用的所有GPU基准测试表格、推理成本数据及资本开支预测
[LLM推理价格的快速但不均匀下降 (Epoch AI)](https://epoch.ai/data-insights/llm-inference-price-trends):独立追踪大语言模型推理价格趋势,显示同等性能下每年约10倍成本下降
[AMD vs NVIDIA推理基准测试:谁赢了?(SemiAnalysis)](https://newsletter.semianalysis.com/p/amd-vs-nvidia-inference-benchmark-who-wins-performance-cost-per-million-tokens):AMD与英伟达GPU架构每百万token成本的详细对比分析
[AI将推动数据中心电力需求到2030年增长165% (Goldman Sachs)](https://www.goldmansachs.com/insights/articles/ai-to-drive-165-increase-in-data-center-power-demand-by-2030):高盛研究量化AI数据中心能源需求走势
[太空数据中心可用太阳能为AI供电 (Scientific American)](https://www.scientificamerican.com/article/data-centers-in-space/):轨道数据中心能效分析及相关技术挑战
[为什么AI世界突然着迷于杰文斯悖论 (NPR Planet Money)](https://www.npr.org/sections/planet-money/2025/02/04/g-s1-46018/ai-deepseek-economics-jevons-paradox):通俗解读杰文斯悖论如何适用于AI效率与需求动态
[超大规模厂商资本开支2026年超6000亿美元 (IEEE ComSoc Technology Blog)](https://techblog.comsoc.org/2025/12/22/hyperscaler-capex-600-bn-in-2026-a-36-increase-over-2025-while-global-spending-on-cloud-infrastructure-services-skyrockets/):2025年和2026年超大规模厂商资本开支预测数据

夜雨聆风