a16z 播客：Mike Volpi 谈为何旧 GPU 仍在保值
Volpi 在 Cisco 时代做过 70 多起收购，这次聊 AI 基建周期挺有干货。核心结论很反直觉：H100 能效上不如 Vera Rubin，但算力太紧缺，旧卡照样有人抢着用，做 GPU 融资和算力租赁的人应该听听。
a16z 发布播客，Hanabi Capital 合伙人 Mike Volpi 与 Raghu Raghuram、Guido Appenzeller 对谈，核心观点是旧 GPU 的使用寿命比预期更长。Volpi 指出，原本市场认为 GPU 资产只有五年生命周期，融资方若资产寿命延长就会成为受益者。理论上有足够的 Nvidia Vera Rubin 后，继续运行 H100 是把电力转化为美元的低效方式，但当前需求过高，即便能效转换差，H100 仍会被继续使用。
他认为根本原因在于稀缺性来自半导体制造本身，只要制造产能受限，资产寿命就会持续拉长。Volpi 曾在 Cisco 主导 70 多起收购，这些收购后来贡献了公司一半收入，他将 AI 基础设施建设潮与当年的互联网建设周期对比，认为这一轮更有吸引力。节目还涉及对 Nvidia 的建议、定制 AI 芯片的机会、AI 时代数据中心网络的改造，以及为什么现在是投资的最佳时机。
Mike Volpi explains why older GPUs are holding their value: "We thought this asset had a five-year life ... If you're in the business of financing an asset and that asset ends up having a longer useful life, then you're the beneficiary of that." " In theory, at some point when you have sufficient Vera Rubins, it doesn't really make sense to run H100s anymore because it's just a worse way of converting electricity into dollars." "The problem is right now that the demand is so high that even if the watt-to-dollar conversion is low, you're going to still use the asset." "Ultimately the scarcity is in the manufacturing of the semiconductor itself... And so then scarcity persists, then the life of the asset continues to be long." @mavolpi @RaghuRaghuram @appenz
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Nvidia, H100, Vera Rubin, a16z.
Source attribution: curated from a16z (@a16z).
