四个人，公司卖了7亿 。 作者/王露 报道/投资界PEdaily 又一场AI造富神话。 成立仅一年、产品尚处内测、用户数未披露——AI初创公司To r c h却以超1亿美元（约 合7亿元人民币）的价格被Op e nAI收购了。 令 人 意 外 的 是 ， To r c h 团 队 只 有 四 个 人 ， 其 中 还 有 一 位 华 人 面 孔 ， 将 全 员 加 入 Op e nAI。粗略计算，人均回报约2 5 0 0万美元，早期投资人亦全身而退。 尤为引人注目的是，To r c h仅成立短短一年。公司产品是一款应用程序，旨在把用户分 散在医院、实验室、可穿戴设备和消费级检测机构的医疗健康信息集中到同一处，形成 可供AI调用的个人医疗档案。 简单来说，To r c h会基于用户上传的病历、检查报告及问诊内容进行分类整理，并将晦 涩的医学术语转化为通俗表述，提炼关键信息，并在多源数据中识别趋势与潜在关联， 最 终 输 出 更 易 理 解 ， 也 更 具 可 执 行 性 的 健 康 洞 察 。 不 过 ， To r c h 本 身 依 赖 包 括 Op e nAI 在内的大模型底层支持，目前产品仍处于小范围内测阶 ...

Core Viewpoint - The article discusses the recent acquisition of the Canadian AI startup CentML by NVIDIA for over $400 million, highlighting the ongoing competition for AI talent in the industry [2][11]. Group 1: Acquisition Details - NVIDIA acquired CentML for over $400 million, which includes a base purchase price estimated at over $300 million, with additional payments tied to performance targets [6]. - CentML was founded by a team led by a 95-year-old Chinese PhD, Wang Shang, who has a strong background in AI and GPU efficiency optimization [4]. - The core technology of CentML, called "Hidet," acts as a translator between AI models and chips, significantly improving AI training efficiency and reducing costs [5]. Group 2: Funding and Valuation - CentML completed multiple funding rounds, including a $3.5 million pre-seed round in June 2022 and a $27 million seed round in October 2023, achieving a post-money valuation of $300 million [5]. - Prior to the acquisition, CentML had raised approximately $30.9 million in venture capital, indicating a strong return for its investors [7]. Group 3: Industry Competition - NVIDIA's acquisition strategy is part of a broader trend in the AI industry, where companies are aggressively acquiring talent and startups to enhance their capabilities [11]. - Other tech giants, such as Meta, are also actively recruiting top AI talent, with reports of significant salary offers and multiple high-profile hires from competitors like OpenAI [12][13]. - The article emphasizes that the competition for AI talent is fierce, with a focus on individuals under 30 with top-tier lab experience, particularly highlighting the prominence of the Chinese talent pool in the AI narrative [14].