商业产业 The Verge AI 2026-10-22T11:50 正文 ok 553 words

OpenAI、Anthropic 与 Nvidia 季度游说支出创历史新高

英文标题:OpenAI, Anthropic, and Nvidia had record lobbying quarters

美国联邦披露文件显示,今年第三季度(7月至9月)Nvidia、Anthropic 和 OpenAI 的游说支出均创下历史新高,分别达到 220 万美元、210 万美元和 150 万美元。这些 AI 头部企业密集就前沿模型安全评估、出口管制、版权、数据中心能源成本及反垄断等关键政策议题展开行动。这表明随着监管审查收紧以及政策环境的快速变化,积极参与并塑造政策规则已成为 AI 领军企业维系竞争壁垒和行业话语权的核心战场。
打开原文相关 2 · 机会 0
media390.4 KB media739.7 KB media3.4 KB media546.8 KB media351.4 KB media462.5 KB media215.9 KB media171.7 KB media335.6 KB media470.9 KB media262.7 KB media797.3 KB media137.6 KB media174.9 KB media324.9 KB media223.5 KB

中文

这个夏天对 AI 游说人员来说异常忙碌。联邦披露文件显示,Anthropic、OpenAI 和 Nvidia 在 7 月至 9 月期间的游说支出均创下了各自的单季度历史新高。

其中 Nvidia 支出最高,达 220 万美元;Anthropic 紧随其后,为 210 万美元;OpenAI 则达到 150 万美元。与在华盛顿长期布局的其他大型科技巨头相比,这些数额并不算异常。例如,自 2021 年以来,微软每个季度的游说支出至少达到 200 万美元,而谷歌通常每季度支出超过 300 万美元(包括本季度)。但这释放出了一个明确信号:这三家 AI 公司已经深刻意识到,参与和重塑政策制定变得何等重要。

在此期间,科技行业 CEO 已经频繁现身华盛顿,其中许多人出席了 9 月下旬与唐纳德·特朗普总统举行的白宫午餐会,并在会上同意达成一项具有“道德约束力”的 AI 安全协议。游说披露信息显示,在特朗普任命其 AI 负责人及特遣工作组之前的这段时间里,幕后进行了更为频繁的沟通与讨论。随着民主党预计将在今年的中期选举后至少赢回国会的部分控制权,对于那些视整顿打击该行业为关键政治筹码的议员们,AI 公司可能会发现与他们打交道变得愈加重要。

Anthropic 列出了一份详尽的清单,记录了其在第三季度针对国会和白宫开展游说的法案、行政命令及其它议题。其中包括《儿童在线安全法》(Kids Online Safety Act)中的 AI 相关条款、《AI Overwatch Act》中的出口管制规定、特朗普将 AI 更名为“超级智能”(Super Intelligence)的行政令,以及旨在让开发商承担数据中心能源成本的《纳税人保护法》(Ratepayer Protection Act)。

AI 公司在政策制定中愈发活跃

尽管 OpenAI 的披露文件仅列出了宏观议题,但其游说举措似乎是对 Hugging Face 遭攻击事件余波的一种应对。除了就版权等议题进行接触外,OpenAI 还向国会、白宫、小企业管理局、商务部和国务院进行了游说,涉及前沿模型的网络安全评估、AI 安全监管框架,以及用于信息共享的反垄断豁免保护。Nvidia 则针对 6G 网络、“负责任的 AI 设计”、竞争政策、量子计算和出口管制政策等问题,向国会以及多个联邦机构和白宫办公室展开了游说。

苹果公司也迎来了一个破纪录的季度,其游说支出达到 310 万美元,涉足的议题包括知识产权、关税、儿童在线安全、竞争和本土半导体制造。例如,苹果参与了旨在将核实用户年龄的责任转嫁给应用商店的相关法案。这是苹果首次报告单季度游说支出超过 300 万美元。

这些披露信息表明,在临近中期选举的这个季度里,领先的 AI 企业在政策制定中正变得越来越积极。OpenAI 直到今年单季度的游说支出才突破 100 万美元,而在 2024 年底之前甚至从未超过 50 万美元。Anthropic 去年单季度最高支出也仅略高于 100 万美元。而本季度也是 Nvidia 游说支出首次突破 200 万美元大关。随着数据中心开发和 AI 安全等议题成为众多选民关注的焦点,无论明年由哪个政党掌控局势,AI 游说者的工作量大概率只增不减。

English / 原文

It was a busy summer for AI lobbyists. Anthropic, OpenAI, and Nvidia each had their highest-spending lobbying quarter between July and September, federal disclosures show.

Nvidia spent the most of the group at $2.2 million, followed by Anthropic’s $2.1 million and OpenAI’s $1.5 million. Compared to other Big Tech companies with long-established presences in Washington, the sums aren’t out of the ordinary. Microsoft, for example, has spent at least $2 million a quarter going back to 2021, and Google regularly spends over $3 million, including this quarter. But it’s a clear signal that the three AI firms recognize how important shaping policy has become.

Tech CEOs were already visible fixtures in Washington during this period, with many of them attending a White House lunch with President Donald Trump in late September, which resulted in a “morally binding” AI safety accord that they agreed to. The lobbying disclosures show that even more discussions were happening frequently behind the scenes, during a time just before Trump named his AI czar and task force. With Democrats expected to win back at least some control in Congress after this year’s midterm elections, AI companies may only find it more important to engage with lawmakers who see cracking down on the industry as a winning issue.

Anthropic included a detailed list of bills, executive orders, and other topics it lobbied Congress and the White House about in the third quarter. Among them were AI-related provisions in the Kids Online Safety Act, export control provisions in the AI Overwatch Act, Trump’s executive order renaming AI “Super Intelligence,” and the Ratepayer Protection Act, which aims to put energy costs for data centers on their developers.

AI companies are becoming increasingly active in policymaking

OpenAI’s lobbying looks like a response to some of the fallout from the Hugging Face attack, though the disclosures only list high-level topics. In addition to engaging on issues like copyright, OpenAI lobbied Congress, the White House, the Small Business Administration, and the Commerce and State Departments over cybersecurity evaluations for frontier models, AI safety oversight frameworks, and antitrust protections for information sharing. Nvidia lobbied Congress, as well as several different federal agencies and White House offices, on issues including 6G networks, “responsible AI design,” competition policy, quantum computing, and export control policy.

Apple also had a record-breaking quarter, spending $3.1 million and engaging on issues including intellectual property, tariffs, kids online safety, competition, and domestic semiconductor manufacturing. It engaged on bills that seek to put the onus on app stores to verify users’ ages, for example. It’s the first quarter in which Apple has reported spending more than $3 million on lobbying.

The disclosures show that leading AI companies are becoming increasingly active in policymaking, in a quarter that marks a lead-up to the midterm elections. OpenAI didn’t even spend $1 million on lobbying in a single quarter until this year, and it didn’t spend more than half a million until the end of 2024. The most Anthropic spent in a single quarter last year was just over $1 million. And this quarter is the first in which Nvidia has spent more than $2 million. With issues like data center development and AI safety on many voters’ minds, AI lobbyists’ work will likely only increase no matter whichever party has control next year.