✦   ✦   ✦

breaking news

News, read through The Primal Race
← 全部评论 · all commentary

没有销售员的公司,卖的是什么A Company With No Salespeople, Selling What Never Needed Selling

科技 结构层 · 文化层 CNBC ↗ 2026-11-02 § 链接
新闻摘要Palantir 公布第三季度财报,营收23.1亿美元、净利润11.3亿美元,均超预期,美国商业客户收入同比增长148%至9.84亿美元,首次超过美国政府收入的9.53亿美元。CEO 卡普称公司没有销售团队也能取得这样的成绩,并再次呼吁将 AI 实验室国有化。
The newsPalantir reported third-quarter revenue of $2.31 billion and net income of $1.13 billion, both above estimates. U.S. commercial revenue rose 148% to $984 million, surpassing U.S. government revenue of $953 million for the first time. CEO Alex Karp said the company is succeeding without salespeople and repeated his call to nationalize AI labs.
卡普说没有销售员,可他卖的东西从来不需要销售员。
Karp says he has no sales force. What he sells never needed one.

卡普说这家公司「没有销售团队」也能做出这样的成绩。这句话是真的,但它不是夸奖,是招供。一家不需要销售员就能把美国商业收入做到同比增长148%、首次超过政府部门的公司,卖的显然不是需要推销的东西。它卖的是别人离不开的东西。

Palantir 的老本行是政府。美国政府部门这一季贡献9.53亿美元,增长96%。现在商业客户以9.84亿美元反超,等于说企业界终于也走到了同一个入口前面。这个入口不是软件功能,是数据、是预测、是对人和组织的提前判定。政府用它是为了治理,企业用它是为了定价、筛选、裁员、风控。同一套东西,换一个付款方,暴力的层就从 structural 挪到了 cultural:谁被系统标成高风险,谁被算法提前排除,没有人签署这个决定,也没有人为它负责。

卡普这一季说的另一句话更值得记下来:AI 实验室应该国有化,这是管理责任的「唯一现实的方式」。一个靠政府合同起家的 CEO,在监管呼声最高的时候,主动要求把整个行业收进国家手里。听起来像进步派,实际上是抢在别人定义规则之前先定义规则。国有化这个词一旦由他来提,讨论的起点就不再是「要不要管」,而是「由谁来管、管成什么样」。这就是第三章的东西:夺取认知入口,比夺取市场更值钱。

财报里没有出现的名字,比出现的更说明问题。报道没有写这些商业客户是谁,没有写它们的员工、用户、被评分的人在这套系统里处于什么位置,也没有写那148%的增长对应的是多少份合同、覆盖多少人的数据。Karp 说「未来已经在这里了,它的政治后果也在这里了」——他说得对,只是他没说的是,这些后果由谁承担。

一家不需要销售员的公司,靠的是它的产品已经变成了基础设施;而基础设施的意思是,你可以不同意它,但你没法不使用它。这才是这一季财报真正的数字:不是23.1亿,是它已经站在了那个位置上——站在那个位置上的人,从来不需要说服任何人。

Karp says the business is succeeding without salespeople. That's true, but it isn't a boast — it's a confession. A company that can grow U.S. commercial revenue 148% year over year and overtake its own government segment without a sales force is not selling something that needs to be pitched. It's selling something people can't get away from.

Palantir's old business was government. The U.S. government segment brought in $953 million this quarter, up 96%. Now commercial customers at $984 million have pulled ahead, which means the corporate world has finally arrived at the same doorway. That doorway isn't a software feature. It's data, prediction, the advance judgment of people and organizations. Governments use it to govern; companies use it to price, screen, fire, and manage risk. Same product, different payer, and the violence shifts layers: whoever the system flags as high-risk, whoever the algorithm quietly excludes — nobody signed that decision, and nobody answers for it.

Karp's other line this quarter is worth keeping: AI labs should be nationalized, the only realistic way to manage liability. A CEO whose company was built on government contracts, asking at the peak of regulatory pressure to hand the whole industry to the state. It sounds progressive. It's actually getting to define the rules before someone else does. Once he says the word nationalize, the question stops being whether to regulate and becomes who regulates, and into what shape. That's the third chapter: seizing the cognitive entrance is worth more than seizing the market.

The names missing from the earnings report say more than the ones in it. The report doesn't say who these commercial customers are, doesn't say where their employees, users, and scored subjects sit inside the system, doesn't say how many contracts or how many people's data the 148% represents. Karp says the future is here and so are its political ramifications. He's right. What he leaves out is who carries them.

A company with no salespeople runs on the fact that its product has become infrastructure — and infrastructure means you can disagree with it and still not opt out. That's the real number in this quarter's report. Not $2.31 billion. The position. And whoever holds that position never has to persuade anyone.