以下是所提供转录内容的总结，涵盖了所有信息： 发言人强调了两个具有巨大增长潜力的主要业务领域： 1. **人工智能计算业务 (AI Compute Business):** * 其潜在规模被形容为“惊人”。 * 预计将成为一个“每年数千亿美元”的业务。 * 这被认为是一个“极其保守的估计”。 * 这些数字指的是营收，或者“可能指利润，但姑且称之为营收”。 * 该业务“完全独立于地面计算”，并且“与他们的发射业务毫无关系”。 2. **星链（全球通信业务 Starlink Global Communications Business）:** * **当前增长：** 星链目前“每季度新增约两百多万客户”。 * 这种增长速度正在“加快”，‘下一季度可能显示其增长将超过三百万人’。 * 当前客户增长的保守估计是“每季度两百万客户”。 * **星舰的影响：** * 如果星舰的发射频率“能达到目前猎鹰9号的水平”（相同的节奏），它将部署星链V3卫星，而非V2。 * V3卫星将“每次发射提供超过20倍的星链带宽”。 * 星舰“旨在实现更快速、更可靠的可重复使用”，这表明它“实际上应该随着时间推移变得更加频繁”。 * 这意味着，如果星舰成功并能以猎鹰9号的频率发射，星链可能“每季度增加4000万客户”（这是通过将当前保守的200万客户增长量乘以20倍带宽提升计算得出的）。 * 这使得他们能够“在相同时间内增加比现在多20倍的客户”。 * **未来客户预测：** * 发言人的论点是，“星链的直销模式将绝对超越家用宽带的增长速度”。 * 这是因为直销“可以服务更多的客户”，而这些客户“不需要接近家用宽带那么多的带宽”。 * 它可能带来“至少一个数量级，甚至可能接近一个半数量级”的客户增长（指的是每单位带宽的直销用户数量与家用宽带用户数量的对比）。 * 未来，全球可能拥有“1亿、2亿、3亿甚至4亿星链客户”，这被认为是可行的，这些客户将是“家庭宽带和其他互联网用途”的结合。

Here's a summary of the provided transcription, including every piece of information: The speaker highlights two main business areas with significant growth potential: 1. **AI Compute Business:** * Described as "stupid" in its potential scale. * Projected to be a "multi-hundred billion dollar a year business." * This is considered an "extremely conservative estimate." * The figures refer to revenue, or "probably profits, but let's just say revenue." * This business is "completely independent of the terrestrial compute" and "nothing at all to do with their launch business." 2. **Starlink (Global Communications Business):** * **Current Growth:** Starlink is "growing at about 2 point something million customers a quarter." * This growth rate is "accelerating," with the "next quarter might show it's growing at over 3 million." * A conservative estimate for current customer additions is "2 million customers a quarter." * **Impact of Starship:** * If Starship launches "as frequently as Falcon 9 today" (the same cadence), it will deploy Starlink V3 satellites, not V2. * The V3 satellites will provide "more than 20 times as much Starlink bandwidth per launch." * Starship is "meant to be more rapidly, reliably reusable," suggesting it "should actually be happening more often over time." * This means that if Starship is successful and launches at the Falcon 9 cadence, Starlink could add "40 million customers a quarter" (calculated by multiplying the conservative 2 million current growth by the 20x bandwidth increase). * This allows them to add "20 times more customers per time period than they are today." * **Future Customer Projections:** * The speaker's thesis is that "Starlink direct-to-sell is going to absolutely blast past growth in the home broadband." * This is because direct-to-sell "can serve many more customers" who "don't need anywhere near as much bandwidth." * It could enable "at least an order of magnitude, maybe close to one and a half more customers" (referring to the number of direct-to-sell users per unit of bandwidth compared to home broadband users). * It is considered "feasible in the future that there could be 100 million, 200 million, 300, 400 million Starlink customers globally," a combination of "home broadband and internet."