风萧萧_Frank
以文会友
埃隆·马斯克 CNBC 主持人大卫·费伯采访
CNBC 实录：埃隆·马斯克今日接受 CNBC 主持人大卫·费伯的现场采访
CNBC Transcript: Elon Musk Sits Down for Secondary Interview with CNBC’s David Faber Live on CNBC Today
May 20, 2025
CNBC 对特斯拉首席执行官埃隆·马斯克、SpaceX 首席执行官、xAI 首席执行官兼 X 创始人以及 CNBC 主持人大卫·费伯的采访实录，采访于5月20日（星期二）上午10点至下午4点（美国东部时间）播出。这是大卫·费伯今日第二次采访埃隆·马斯克。视频链接：
https://www.cnbc.com/video/2025/05/20/tesla-ceo-elon-musk-theres-no-need-for-tesla-to-buy-uber.html
https://www.cnbc.com/video/2025/05/20/tesla-ceo-elon-musk-no-plans-to-merge-tesla-and-xai.html
David Faber 今天首次采访 Elon Musk 的文字记录：
大卫·费伯：乔恩，谢谢。是的，能有这么一个精彩的中场休息真是难得。埃隆·马斯克再次来到奥斯汀的超级工厂与我见面。谢谢你回来，我们之前讨论过特斯拉，但很多事情都没谈及。显然，我很少遇到这种空档期。所以，让我来回答一下上次谈话后一些人开始提出的一个问题。特斯拉为什么不收购优步？
埃隆·马斯克：没必要，因为我们有大量的汽车。我们有数百万辆能够自动驾驶的汽车。我应该说，这既包括特斯拉自有车队，也允许特斯拉车主在车队中增减车辆，这样现有的特斯拉车主就可以通过增加或减少车辆数量来赚钱。我认为，特斯拉车主或许可以通过允许车辆加入自动驾驶车队来赚取更多收益，而不是在租赁中支付费用。这就是为什么我认为，这有点像Airbnb和Uber的结合体。就像Airbnb一样，你可以出租闲置的卧室或房子（如果你不使用的话），并从中赚钱。这正是我们期望特斯拉客户能够做到的。
费伯：对。随着规模的扩大，你显然，你之前说过，你觉得自己拥有了物流能力，也拥有了开发应用程序的能力。当然，还有其他需要改进的地方吗？
马斯克：是的，我认为我们没有遗漏任何东西。特斯拉拥有在一夜之间打造庞大自动驾驶车队所需的所有要素。
FABER：是的，虽然我谈到了交易，但这实际上会让我去XAI的一些地方，但你知道，我之前在今天早些时候的一次采访中提到过，不是跟我。你很忙，伙计。你有时白天要面对很多不同的媒体。
MUSK：是的，有很多事情要做。
FABER：是的，你曾表示希望在特斯拉拥有更多控制权。
MUSK：是的。我的意思是，真的，只要有足够的控制权，以免在未来某个时候被激进投资者赶下台就行了。
FABER：是的
MUSK：这就是我所说的，就像某种……
FABER：你之前说过。
MUSK：是的，有些新东西，也就是大约25%的水平，你知道，这意味着如果我发疯了，我肯定会被赶出去。但这确实需要，你知道，我感觉有点像这个数字，因为这就是我拥有的控制权，但控制权还没有大到让我不会被赶走。
FABER：好的，我明白你想要25%的原生控制权。在那个层面上，你根本不可能拥有。
马斯克：不，你可以，如果你真的，如果你表现糟糕，并且正在摧毁公司的价值。或者，好吧，或者我已经彻底疯了。但如果你认为这不是，那不是，那不是真正的控制权。我的意思是，如果你拥有真正的控制权，就像谷歌的拉里和谢尔盖那样，你知道，任何拥有多数投票控制权的人，而我没有。
FABER：不，明白，扎克伯格拥有它，拉里和谢尔盖。我知道有些公司拥有绝对投票权。
马斯克：即使你犯了谋杀罪，身陷囹圄，你仍然能够控制公司。
费伯：在这种情况下，你的经济状况和你的投票权是一样的。
马斯克：在这种情况下
是的，我很乐意把这分开，但我基本上是交易所。他们不被允许这么做。
FABER：我知道，我知道他们会这么做。你知道，你提到要毁掉公司，显然你不会这么做。但还是回到我们之前讨论的话题吧。不过，你不觉得你损害了公司。你知道，我再次回到XAI的问题上。但去年息税前利润为140亿美元，而2022年的息税前利润为76.59亿美元，你的营业利润率显然下降了，上个季度的销售额和所有市场都下降了。你看到转机了吗？你会不会照照镜子，然后说，伙计，我得做得更好。我得变得更好。
马斯克：你看。我认为人们可以看看个别季度的业绩，但这无关紧要。正如我之前提到的，我们为了生产Model Y，对全球所有生产线进行了大规模的改造。Model Y几乎占据了我们汽车销量的绝大部分，而Model Y是全球最畅销的车型。所以，显然，为了改造工厂，
FABER：而且还有一款外观漂亮的新车型。
MUSK：是的，它实际上看起来像个变形金刚。
FABER：当然，我知道，我们的相机，抱歉，上面那个，就是镜头。就是它，没错，擎天柱次贷，没错。
MUSK：所以，但真正衡量公司未来价值的最佳指标是股价和市值，而目前股价非常强劲。
FABER：不，它的市值已经反弹到超过1万亿美元。显然，它在过去一年里大幅上涨。不过，正如我们之前所说，这很大程度上与自动驾驶的Robotaxi有关。我一直指着身后，我们的观众看不到它。但是，你最近说擎天柱机器人会发展到数百亿个，但这还需要几十年的时间。
马斯克：至少十年后。
费伯：肯定不止这些。
马斯克：它会增长得非常快。
费伯：你为什么这么认为？
马斯克：我认为人形机器人将成为有史以来最热门的产品。它的需求将是无止境的。基本上每个人都想要一个，谁不想拥有自己的C3PO或R2D2呢？每个人都会想要一个。
费伯：你说过。你知道你的目标是生产一百万台机器人，我想到2030年，这就是我之前提到的。
马斯克：我认为这是一个合理的目标。
费伯：然后，开始迈向可持续的富足，这是你可以探讨的。你知道，我想知道我们一直在讨论自动驾驶，以及需要多长时间，需要训练汽车才能达到或超过人类的能力。但是这些机器人呢？它们需要多少训练才能真正完成各种不同类型的任务？这需要很长时间吗？
马斯克：是的，需要大量的训练，这需要大量的计算机资源，而且需要时间。我认为我们可以实现某些门槛上的突破，比如，如果擎天柱可以观看视频、YouTube视频、操作视频等等。基于那段视频，就像人类可以学习如何做这件事一样，那么你的任务扩展性就会非常显著，因为它可以非常快速地学习任何东西。所以我认为我们很快就会实现这个目标。
费伯：不过，我们还没到那一步。
马斯克：我们还没到那一步。
费伯：你依赖于学习和训练方面的显著提升。
马斯克：是的，不，那是——
费伯：还有通勤。
马斯克：这就是为什么我称之为一个非常重要的门槛——
费伯：明白了。
马斯克：通过观看视频学习的能力，就是这样——
费伯：对。
马斯克：但就目前而言——
费伯：这与观看人类视频，或者让人类现在用某个任务来训练它不同。
马斯克：是的，没错。目前，我们正在训练擎天柱执行一些基本任务，需要人类穿着一种叫做“动作捕捉”（Mocap）的服装，头上装有摄像头，机器人会像人一样移动，比如捡起物体、开门，或者做一些基本任务，比如扔球、跳舞等等。我认为我们需要这样做，才能提升它的智能，让它具备基本功能。接下来，我觉得它非常有趣，也非常像人类的地方在于，你希望机器人能够自我游戏。你会问，孩子是如何学习的？嗯，孩子有玩具，然后玩玩具，玩积木。在某个时候，通过反复练习，弄清楚如何把三角形放进三角形的洞里，把圆形放进圆形的洞里。这就是自游戏。一旦你拥有了很多机器人，你就可以进行这种自游戏，也就是把机器人放在一个有玩具的房间里，让它玩玩具。
费伯：它会学习。
马斯克：是的。而且你必须有一个奖励机制。比如说，机器人的目标是，你知道，那种经典的儿童
玩具。你把圆形放进圆形孔里，正方形放进方形孔里，三角形放进三角形孔里，一直重复，直到它起作用，奖励函数成功为止。
FABER：现在不需要任何进步就能实现这一点吗？我的意思是，不需要人工智能或计算方面的进步，或者诸如此类的事情？
马斯克：需要一些进步，但我认为这些进步并非不可克服。我认为我们可以，我们可以在未来几年内解决这些问题。
FABER：好的，所以到那时，当数百万辆这样的汽车下线时，就像我们刚才看到的汽车一样，它们将完全自动驾驶，它们将会……
马斯克：哦，是的。
FABER：我的意思是，就像你所描述的……
马斯克：当然是五年后。
FABER：来我家吧，我就能说，好了，洗碗了。现在我需要你去遛狗。
马斯克：当然。
费伯：帮我抱抱孩子。
马斯克：事实上，你真的不需要它，它会自动知道你可能想要什么，然后按照你可能想要的方式去做，甚至你都不用开口。
费伯：你需要多少个 GPU？
马斯克：嗯，相当多的 GPU。我们确实有自己的训练程序 Dojo，我认为它会很有帮助。
费伯：目前，它对自动驾驶的贡献大约为 5%，对吧？
马斯克：是的，所以——
费伯：是在纽约吗？在纽约州？
马斯克：是的，在纽约。所以我们预计仍然会从英伟达购买大量 GPU，一些从 AMD 购买，也可能从其他公司购买。只要英伟达的产品比我们生产的更好，我们就会继续从英伟达购买。
FABER：现在是这样吗？
马斯克：是的，是的。
FABER：是的。我的意思是，你显然是为了xAI，孟菲斯——
马斯克：规模很大。
FABER：规模很大，对吧？
马斯克：是的。xAI正在构建最强大的，我们拥有最强大的，我认为我们目前拥有世界上最强大的训练集群，它由超过20万个GPU组成，用于进行连贯的训练。
FABER：现在已经有20万个了。
马斯克：是的。
FABER：在孟菲斯，那个设施——
马斯克：是的。
FABER：好的
马斯克：是的。
FABER：你要去哪里？
马斯克：然后，我们将在孟菲斯附近的某个地方部署百万级 GPU。
费伯：等一下，在新地点部署一百万个 GPU，或者再部署 80 万个 GPU。
马斯克：一百万，一百万个下一代 GPU。
费伯：那么布莱克韦尔。
马斯克：嗯，是的。
费伯：等等，你现在在建造这个吗？
马斯克：是的。
费伯：你是怎么为它供电的？
马斯克：这是一个千兆瓦级的系统。
费伯：什么，你为它准备了电力吗？
马斯克：是的。
费伯：有。天然气还是——
马斯克：是的。这是一个难题。
费伯：这是一个难题。找到电力？
马斯克：获取千兆瓦的电力，将千兆瓦的电力投入使用，并确保这些千兆瓦的电力可靠运行，因为电网中存在电力波动等等。因此，实际上，我今天刚在网上发布了一些信息，其中提到了一堆特斯拉超级电池组（megapack），这些电池对于调节电网的电力至关重要，这样 GPU 就不会受到电力波动的影响。它们需要稳定的电力，所以，如果偶尔出现轻微的电压下降，或者像这次停电一样，你需要一个不间断电源来应对这种情况。所以我们准备了很多超级电池组来支持——
费伯：为了支持这一点。你明白吗——
马斯克：这将是世界上第一个千兆瓦级的训练集群，也是最强大的训练集群。
费伯：什么时候能建成？
马斯克：希望在六个月或九个月左右。
FABER：这主要是为 Grok 提供动力吗？
MUSK：是的。
FABER：是的。
MUSK：这只是为 Grok 提供动力。
FABER：对。Grok 还在不断进步。我的意思是，我经常使用它。
MUSK：是的。
FABER：我看到你在 Joe 身上用过这个模式，是和 Rogan 一起用那个模式吗？那个模式是什么？就是那种总是调皮捣蛋、满嘴脏话的模式。
MUSK：嗯，我的意思是，如果你想在派对上找点乐子，Grok 的疯狂模式会很有趣。
FABER：是的。
MUSK：这是下一个层次。
FABER：是的。功耗会成为我们继续推进人工智能的门槛吗？
马斯克：是的，我想几年前我就做过一个非常明显的预测，那就是人工智能的限制因素将是芯片。现在仍然是芯片，某种程度上是芯片，然后是电气设备，因为你需要将电压从30万伏降到400伏，供计算机使用。所以，你需要降压变压器，大量的变压器，大量的电缆、线路和保险丝。你知道，本质上就是大量的变压器。而电力变压器行业并不习惯需求的大幅变化。
费伯：现在变压器短缺。
马斯克：字面意思是变压器短缺。然后，有趣的是，人工智能算法被称为变压器。
我们的擎天柱机器人是以擎天柱命名的。那是一个变形金刚机器人。
费伯：数量很多。
马斯克：所以我们是变形金刚的变形金刚，变形金刚的变形金刚。
费伯：对，对，但这台变压器，就是其他变压器都需要的，而这台变压器短缺。
马斯克：当我们解决变压器短缺问题时，就会出现根本性的电力短缺。
费伯：我们到了吗？或者说我们很快就会到达？
马斯克：我们很快就要到了。
费伯：我们正在。
马斯克：我猜人们可能会在发电方面开始面临挑战，大概在明年年中或年底。
费伯：即使放松管制，并努力加快步伐，情况也可能如此。
马斯克：有多少发电厂正在建设中？一座发电厂能建多久？
费伯：对，对。中国似乎正在快速建设这些设施。
马斯克：中国正在建设——中国已经建成和正在建设的发电厂数量非常多。我认为人们并没有完全意识到这一点。我在我的X账户上发布了一张美国发电量与中国发电量的对比图。中国的发电量看起来就像一枚火箭升空，而美国的发电量则持平。
费伯：对。
马斯克：所以我认为到今年年底，中国的发电量将达到美国的2.5倍左右，并且有望达到美国的3到4倍。
费伯：有趣的是，当你想到中国时，我的意思是，电动汽车、自动驾驶汽车，我们谈到了电池、太阳能、发电——顺便说一下，甚至最近还谈到了生物技术。我不知道你是否看到了辉瑞公司授权的抗癌药物。
马斯克：是的。
费伯：他们似乎是，我不知道。我来问你个问题。他们在某些重要领域领先于我们吗？
马斯克：美国在突破性创新方面仍然占有优势，但——所以是美国——我认为这在某种程度上是一种文化因素，也就是说，要想实现突破性创新，你必须质疑权威。从根本上说，当你进行突破性创新时，你就是在质疑传统观念。
费伯：对。
马斯克：你知道，中国通常不喜欢质疑自己的权威，或者说，在质疑权威方面，中国不像美国那样鼓励这种做法。
费伯：不。但他们似乎确实擅长发现一些东西，然后使其变得更好，并不断扩大规模。
马斯克：是的，我确实想强调的是，中国拥有如此多聪明、才华横溢、努力工作的人才，数量之多令人惊叹。人才的数量之多，简直是惊人的。我的意思是，我钦佩中国的实力。我认为，我认为大多数中国以外的人并不了解中国的实力。它确实很特别，很特别。
费伯：是的，是的。你知道，在我们剩下的时间里，我想继续关注xAI。首先，你知道，你之前提到希望在特斯拉拥有更多控制权。你会考虑将xAI并入特斯拉吗？显然，这是一种他们可以向你发行股票的方式，而且可以想象这会提升你的整体经济效益。有可能吗？
马斯克：嗯，我想一切皆有可能，但很难去猜测——你知道，特斯拉是上市公司，所以——
费伯：是的，我认为要获得少数股东的多数票，或者任何你可能需要的票数，可能并不容易。
马斯克：是的。这并非不可能，但这必须是特斯拉股东会投票赞成的——他们愿意投票支持的，所以——
费伯：明白了。所以，但这不是你正在考虑做的事情？
马斯克：目前还没有。目前还没有——没有计划这么做。
费伯：对。
马斯克：这并非不可能，但显然需要特斯拉股东的支持。
费伯：另一个你显然可以增强控制权的方法是让那项薪酬计划在特拉华州或特拉华州最高法院获得通过。或者制定一个新的方案。我相信你从2018年起就没领过工资了。
马斯克：是的。
费伯：那么你现在在哪里——
马斯克：七年了。
费伯：是的。你现在在哪里——
马斯克：七年零工资。
费伯：不过，公平地说，如果交易成功，你会赚得盆满钵满。
马斯克：是的，没问题。当然。
费伯：没问题？你会赚得比任何人都多的钱。
马斯克：是的，我想是的。但我的意思是，如果任何一位财富500强企业的CEO同意像我同意的这样的计划，你应该立即购买该公司的股票。
费伯：是的。不，我记得当时，公平地说，股价远高于目标价，而且显然已经达到了。
马斯克：是的。
费伯：据你所知，董事会是否正在制定另一个潜在计划，如果那个计划最终无法通过特拉华州，最高法院裁定反对州长的裁决？
马斯克：我的意思是，我不想……
我代表董事会发言，但你知道，我相信他们正在考虑这个问题。不过，你知道，我无法评论特斯拉董事会的讨论。
费伯：你之前说过。我的意思是，你想继续担任首席执行官。为什么不——你知道，埃隆，考虑到你在很多方面都承受着巨大的压力，我不禁想，为什么不在特斯拉成为像埃里森那样的人物呢？显然，你年轻得多，但他在甲骨文仍然很有影响力。但你知道，他不是首席执行官，而且他没有得到应有的关注。
马斯克：他不是首席执行官？
费伯：不，他不是。萨弗拉才是首席执行官。
马斯克：哦，好的。
费伯：是的。至少在头衔上不是。你提得好。
马斯克：不，不，我想也许更贴切的说法是——我是拉里·埃里森的超级粉丝，我的好朋友。老板。
费伯：再说一遍？
马斯克：老板，我想应该是——他是甲骨文的老板。
费伯：是的，是的，他是。而且他有很多股份。最后，我想——
马斯克：他不是CEO，他只是老板。
费伯：他是老板。是的，他的股份比例很高，比你在特斯拉的股份比例高得多。
马斯克：是的——甲骨文的价格和比例。
费伯：最后，我想说，我的意思是，有很多事情我们还没谈到，但我们尽量不耽误你的时间。我们准备好迎接人工智能将给社会带来的变化了吗？你知道，我每天都在努力工作，关注这些变化。显然，我们——他们在某种程度上——被当今的议题推到了一边。但这一切来得很快。
马斯克：是的，来得非常快。我的意思是，我感觉我们正处于智能爆炸的大爆炸之中。就像我们身处其中——我们正在观看。我们坐在场边，见证着智能爆炸的大爆炸。有一点是肯定的，它不会无聊，所以——
费伯：不。你似乎不再像以前那样频繁地谈论毁灭的可能性是20%了。
马斯克：嗯，我认为我们应该始终考虑到可能存在一些不好的结果，并努力避免这种不好的结果。我们不想自满，说一切都会好起来。不可能出现不好的结果。你知道，你知道，我有点用电影术语来思考这个问题，比如，我们是在《星际迷航》电影里，还是在吉恩·罗登贝里的电影里，或者詹姆斯·卡梅隆的电影里？我们究竟是哪一部电影？结果可能是罗登贝里式的，也可能是卡梅隆式的。我认为在这种情况下，我们想要罗登贝里式的结果。
费伯：是的。至于随之而来的丰厚回报，埃隆，我们就到此为止吧。非常感谢你的再次采访。显然，我们还有很多话要说，所以我希望很快能再见到你。
马斯克：很高兴。
费伯：谢谢你抽出时间。
马斯克：谢谢，非常感谢。
费伯：当然，特斯拉的首席执行官埃隆·马斯克——他仍然想继续担任首席执行官。
马斯克：插入公司名称的首席执行官。
费伯：你现在会更频繁地来这里吗？
马斯克：我的意思是，这里主要是——奥斯汀，特斯拉总部，也是我的主要住所。
费伯：你会完全想念白宫吗？
马斯克：我——我大致的计划是每隔几周去白宫待几天。尽我所能提供帮助。
费伯：谢谢。好的，奥斯汀超级工厂的报道就到这里。
CNBC Transcript: Elon Musk Sits Down for Secondary Interview with CNBC’s David Faber Live on CNBC Today
May 20, 2025
CNBC's “Closing Bell: Overtime”
Following is the unofficial transcript of a CNBC interview with Elon Musk, Tesla CEO, SpaceX CEO, xAI CEO & X Owner, and CNBC’s David Faber on “Closing Bell: Overtime” (M-F, 4PM-5PM ET) today, Tuesday, May 20. This is David Faber’s second interview with Elon Musk today. Following are links to video on CNBC.com:
https://www.cnbc.com/video/2025/05/20/tesla-ceo-elon-musk-theres-no-need-for-tesla-to-buy-uber.html,
https://www.cnbc.com/video/2025/05/20/tesla-ceo-elon-musk-no-plans-to-merge-tesla-and-xai.html.
Transcript of David Faber's first interview with Elon Musk today:
MANDATORY CREDIT: “CNBC”
1. Mandatory credit to “CNBC” on first reference.
2. The onscreen “CNBC” logo must be clearly visible and unobstructed at all times in any image, video clip or other form of media.
3. Embedded web video must stream from the CNBC.com media player with the unobstructed credit as described above.
###
DAVID FABER: Jon, thank you. And yes, it is a rarity to have sort of a twofer a nice little intermission. Elon Musk joining me again here at the Gigafactory in Austin. Thank you for coming back so many things we didn’t get to we were discussing Tesla. Obviously, it’s rare that I get this sort of interregnum. So let me just come to a question some people started to have after our last conversation seemed to come up. Why doesn’t Tesla buy Uber?
ELON MUSK: There’s no need, because we have a large number of cars. We have millions of cars that will be able to operate autonomously. And I should say that it’s a combination of a Tesla owned fleet and also enabling Tesla owners to be able to add or subtract their car to the fleet, so that existing Tesla owners will be able to earn money by adding their car to the fleet for autonomous use. And I think it’s maybe possible for Tesla owners to make more in how that in allowing the car to be added use the self driving fleet that then they then they cost them in the lease. And that’s why, I’d say, consider it sort of, sort of a business model that’s similar to some combination of of Airbnb and Uber. Just like Airbnb, you can rent out your spare bedroom or your your house if you’re not using it, and make money on it. And that’s what we expect Tesla customers to be able to do
FABER: Right And as this scales up, you obviously, you said earlier, you feel like you have the logistics ability and the ability to create the app as well. Of course, anything else around that that needs to occur?
MUSK: Yeah, I don’t think we’re missing anything. Tesla has all the ingredients necessary to offer a vast self driving fleet overnight.
FABER: Yeah, while I’m on the subject of deals, and this will take me actually, to XAI go a few places, but you know, I did note earlier in an interview you did earlier today, not with me. You’re busy, man. You do a lot of different media during the day sometimes.
MUSK: Yeah, a lot going on.
FABER: Yeah, you had indicated that you would like more control at Tesla,
MUSK: Yes. I mean, really, just enough control to not get ousted by activist investors at some point in the future
FABER: Yeah
MUSK: So that’s what I’ve said, is like some something
FABER: You’ve said that before
MUSK: Yeah something new, which is that around the sort of 25% level, you know, that means I, I can certainly be thrown out if I go crazy. But it really requires, you know, that’s the number that I that I feel kind of, you know, because it’s that that’s, that’s where I have, I have some control, but not so much control that I can’t be thrown out,
FABER: Right I understand you want 25 like, native control. There’s no real way at that level that you could be.
MUSK: No you can, if you, if you’re really, if you’re terrible, and, destroying the value of the company. And, well, or I’ve just gone flat, flat out crazy. But if you say it’s not, it’s not, it’s not real control. I mean, if you real control, would be like, say, you know Larry and Sergey at Google, but they, you know, anyone who has a majority voting control, which I wouldn’t have.
FABER: No understood, Zuckerberg has got it, Larry and Sergey. I know there are these companies where there is an absolute voting.
MUSK: You can commit murder and be in prison and still control the company.
FABER: Your economics are the same as your vote level, in this case.
MUSK: In this case yes, but I’d be happy to separate that, but I’m basically the exchanges. They’re not allowed to.
FABER: I know, I know they do. You know, you mentioned destroying the company, and obviously you’re not doing that. But sort of finish where we were talking about earlier. You don’t feel like you’ve damaged the company, though. You know, again, I come back to I want to move on to XAI as well. But 14 billion in EBIT in 2022 last year was 7.659 in EBIT, your gap operating margins, obviously are down, sales and all the markets last quarter were down. Do you see a turnaround coming? Do you look at yourself in the mirror and go, Man, you know, I got to do better. I got to be better.
MUSK: Look. I think the one can sort of look at individual quarterly results, and that’s neither here nor there. As I mentioned earlier, we did a massive global retooling of all the factory lines for the model Y, which is almost it’s majority of our vehicle sales is the model Y, and model Y is the best selling car on Earth any kind. So obviously, you know, for retooling the factory,
FABER: And a nice looking new, new model here too.
MUSK: Yeah, it looks like a transformer actually.
FABER: I know, of course, our camera, sorry, up there, that’s the shot. There it is, yeah, Optimus subprime, yeah.
MUSK: So, but the really, the best indicator of the forward value the company is the share price of the market cap, and that’s very strong right now.
FABER: No, and it’s and it’s rebounded to over a trillion dollars. Obviously, it’s up sharply over the last year. A lot of it has to do, though, as we said earlier, with autonomous the Robotaxi, and I keep pointing behind me, our viewers can’t see it, but, but Optimus, the robot, you said recently 10s of billions of robots, but that’s decades away,
MUSK: At least one decade away.
FABER: It’s got to be more than that.
MUSK: It’s going to grow very fast.
FABER: Why do you think that?
MUSK: I think, I think humanoid robots will be the biggest product ever. The demand will be insatiable. Everyone’s gonna want one, basically, who wouldn’t want their own personal C3PO or R2D2. Everyone’s going to want one.
FABER: You’ve said that. And you know your your goal, you’ve also said is to produce a million robots, I think by 2030 that’s what I had you on the record as saying,
MUSK: I think that’s a reasonable target.
FABER: And then, and then start towards sustainable abundance, which you can get into. You know, I wonder we’ve been talking about autonomous and how long it takes, it takes to train the automobile to be able to be the equivalent of or exceed human capabilities. But what about these robots, to the extent that how much training are they going to need to actually be able to do various different types of tasks? Is that something that is going to take a long time.
MUSK: A lot of training yes, it’s going to take a lot of computer resources, and it’ll take, it’ll take time. I think there’s certain threshold breakthroughs that that we think we can achieve, where, if, if optimus can watch videos, YouTube videos, or how to videos, or whatever. And based on that video, just like a human can learn how to do that thing, then you really have task extensibility that is dramatic, because then it can learn anything very quickly. So I think we’ll get there in the next.
FABER: We’re not there yet, though.
MUSK: We’re not there yet
FABER: You’re relying on a significant uptick in terms of learning and training.
MUSK: Yeah, no, that’s—
FABER: And commute.
MUSK: That’s why I’m calling it a very significant threshold would be—
FABER: Understood.
MUSK: The ability to learn from watching a video, just so—
FABER: Right.
MUSK: But at the point we’re—
FABER: As opposed to watching a human right, which is, or having a human sort of train it right now with a task.
MUSK: Yeah, right. Right now, we’re really, we’re training an Optimus to do like primitive tasks where a human in a kind of a, what’s called a Mocap suit is and sort of cameras on the head is moving in the way that the robot would move to, say, pick up an object or open a door, or the basic tasks, throw a ball, dance, and we need to that I think that’s needed to sort of bootstrap the intelligence so you can have the basic functions. Then where I think it gets very interesting, and very much like humans, is that you want the robot to self-play. So you say, how does a child learn? Well, a child has toys and a child plays with the toys. Plays with the blocks. At some point, figures how to put the triangle in the triangle hole and the circle in the circle hole by doing it over and over again. And this the self-play. Once you have a lot of robots, you can do this self-play, which is that you just put the robot in a room with toys and have the robot, literally have the robot play with toys.
FABER: And it will learn.
MUSK: Yeah. And you have to have a reward function. Say, like, okay, the goal of the robot is to put, you know, that classic child’s toy. You put the circle in the circle hole, the square in the square hole, triangle in the triangle hole, and keep doing it until it works and the reward function is succeeding.
FABER: And there are no advances needed to accomplish that now? I mean, no advances in AI or just compute and things of that nature that can happen?
MUSK: There are some advances needed, but I don’t think these are insurmountable. I think we I think we can, we can solve these things in the next few years.
FABER: Okay, so at that time when millions of these things are coming off a line like we just, you know, saw with cars, they’re going to be fully autonomous, they’re going to come off and—
MUSK: Oh yeah.
FABER: I mean, what you’ve described—
MUSK: In five years, of course.
FABER: Come in my house, and I’m gonna be able to say, alright, do the dishes. Now I need you to walk the dog.
MUSK: Absolutely.
FABER: Hold the baby.
MUSK: In fact, you really won’t even need to it’ll it’ll figure out what you probably want and and do what you probably want without you even having to ask.
FABER: How many GPUs are you going to need for that?
MUSK: Well, quite a few GPUs. We do have our own program called Dojo for training, which I think will be helpful.
FABER: Right now, it’s contributing about 5% towards self-driving, right?
MUSK: Yeah, so—
FABER: Is that the one in New York? In New York State?
MUSK: Yeah, yes, in New York. So we expect to still buy a lot of GPUs from Nvidia, some from AMD, and maybe from others. And as long as Nvidia is better than what we make, we’ll keep buying from Nvidia.
FABER: Is that the case right now?
MUSK: It is, yeah.
FABER: Yeah. I mean, you’re obviously buying them for xAI, the Memphis—
MUSK: Big time.
FABER: Big time, right?
MUSK: Yeah. xAI is building the most powerful, we have the, we have the most powerful, I think we have the most powerful training cluster in the world right now, which is over 200,000 GPUs, training coherently.
FABER: You’re at 200,000 already now.
MUSK: Yes.
FABER: There in Memphis, that facility that—
MUSK: Yes.
FABER: Okay
MUSK: Yes.
FABER: And where are you going?
MUSK: And the, we’ll be at the million GPU level in a location just near Memphis.
FABER: Wait a million GPUs for a new location, or 800,000 additional GPUs.
MUSK: A million, a million of the next generation GPUs.
FABER: So Blackwell.
MUSK: Well, yeah.
FABER: Wait, are you building that now?
MUSK: Yeah.
FABER: How are you powering that?
MUSK: It’s a gigawatt class system.
FABER: What, do you have the power lined up for that?
MUSK: Yeah.
FABER: You do. Natural gas or—
MUSK : It is. It is a hard problem.
FABER: It’s a, which is the hard problem. Finding the power?
MUSK: Getting the gigawatts, bringing, bringing a gigawatt online and actually having the gigawatts of power be reliable, because you get the power fluctuations in the grid and whatnot. So we’re using, actually, I just posted something online today which is a whole bunch of Tesla megapacks, batteries that are important for power conditioning the grid so that GPUs do not like power fluctuations. They’re like, they like a power steady so and then, if there are sometimes there’s slight brownouts, or for this blackout, you want to have be able to carry through that like an uninterruptible power supply. So, so we got a lot of megapacks there to support—
FABER: To support that. Do you see—
MUSK: It’ll be, it’ll be the first gigawatt class training cluster in the world, and the most powerful training cluster.
FABER: When is that going to be done?
MUSK: Hopefully, in about six months, maybe nine months.
FABER: And that’s largely powering Grok?
MUSK: Yes.
FABER: Yeah.
MUSK: It’s just powering Grok.
FABER: Right. Which continues to advance. I mean, I use it frequently.
MUSK: Yeah.
FABER: I saw you using that on the Joe, was it with Rogan when you were using that mode? What is that mode? Where it’s all sassy and curses all the time.
MUSK: Well there’s, I mean, if you want to have fun at parties, Grok unhinged mode is pretty funny.
FABER: Yeah.
MUSK: It’s next level.
FABER: Yeah. Is power going to be the gating issue for our ability to continue to advance in AI?
MUSK: Yeah, I think we’re, I mean, a few years ago, I made a very obvious prediction which is that the limitation on AI will be chips. And it’s still chips, kind of chips today, then it will be electrical equipment for the because you need to take power at that might be at 300,000 volts, all the way down to 400 volts for the for the computer. So it’s, you need to step down transformers, and a lot of them, and a lot of, you know, cabling and wiring and fuses. And I, you know, it’s a lot of transformers, essentially. And the electrical transformer industry is not used to big changes in demand.
FABER: Now there’s a shortage of transformers today.
MUSK: Literally the shortage of transformers. And then, funnily enough, the the the AI algorithm, is called a transformer. And then our Optimus robot is named after Optimus Prime. That’s a transformer robot.
FABER: It’s a lot.
MUSK: So we’re transformers for transformers for transformers.
FABER: Right, right, but it’s the one, the one transformer is the one in shortage that the others need.
MUSK: And then as we solve the transformer shortage, there’ll be the fundamental electricity generation shortage.
FABER: And are we there yet? Or are we going to be there soon?
MUSK: We’re getting there soon.
FABER: We are.
MUSK: My guess is people are going to start getting challenges with power generation, maybe by middle of next year, end of next year.
FABER: Even with deregulation and an effort being made to perhaps move -- along more quickly.
MUSK: How many power plants are getting built and how fast can you build a power plant?
FABER: Right, right. China seems to be building them pretty quickly.
MUSK: China is building – China has so many power plants that have been built and are being built. I don’t think people quite realize this. I posted on my X account just the graph of U.S. power generation versus China power generation. China power generation is like, looks like a rocket going to orbit, and U.S. power generation is flat.
FABER: Right.
MUSK: So I think by the end of this year, China will have about two and a half times the power output of the United States, and it’s headed towards maybe three or four times the power output of the United States.
FABER: It’s funny, when you think about China, I mean, EVs, autonomous, we talked about batteries, solar, power generation – by the way, even biotechnology recently. I don’t know if you saw Pfizer’s licensing cancer drugs.
MUSK: Yeah.
FABER: They seem to be, I don’t know. I’ll ask you the question. Are they ahead of us in certain areas that are important?
MUSK: The United States still has an advantage in breakthrough innovation, but – so it’s the United – and I think it’s somewhat of a cultural thing, which is that to have breakthrough innovation, you have to question authority. That fundamentally your breakthrough, you’re questioning the conventional wisdom when you do a breakthrough innovation.
FABER: Right.
MUSK: And you know, China, they don’t generally like to question their authority, or that’s just not as encouraged as it is in the U.S. on questioning authority.
FABER: No. But they do seem to be good at finding something and then making it better. And scaling.
MUSK: Yeah, I do want to emphasize that the sheer number of smart, talented people in China who work very hard is amazing. The amount of – there’s just the sheer quantity of talent. I mean, I’m an admirer of China’s capabilities. I think, I think most people outside of China do not understand the power of China. It really is something, something special.
FABER: Yeah, yeah. You know, in the time we have left, I would like to sort of keep the focus on xAI. First of all, you know, you spoke earlier about wanting more control at Tesla. Would you ever consider merging xAI into Tesla? It would be a way, obviously, that they could issue shares to you and conceivably would increase your overall economics. Is that a possibility?
MUSK: Well, I guess anything’s possible, but it would be difficult to speculate about something – you know, Tesla’s a publicly traded stock so –
FABER: Yes, I would think to get a majority of the minority vote, or whatever you might need, might be not easy.
MUSK: Yeah. It’s not out of the question, but that would have to be something that the Tesla shareholders would vote – would want to vote for, so –
FABER: Understood So, but it’s not something you’re thinking about doing?
MUSK: It’s not currently. It’s not currently – there are no plans to do so.
FABER: Right.
MUSK: It’s not out of the question, but obviously would require Tesla shareholders support for them.
FABER: Another way you could obviously increase your control would be to get that comp plan passed in Delaware or through the Delaware Supreme Court. Or a new one. You haven’t been paid, I believe, since 2018.
MUSK: Yeah.
FABER: So where are you on –
MUSK: Seven years.
FABER: Yeah. Where are you on –
MUSK: Seven years with zero pay.
FABER: Although, to be fair, if it goes through, you’re gonna have an enormous, enormous pay day.
MUSK: Yeah it will be fine. Sure.
FABER: It will be fine? You’ll have more money than everybody’s ever made.
MUSK: Yeah, I suppose so. But I mean, I would, let’s just say that if any CEO of the Fortune 500 were to agree to a plan like the one that I agreed to, you should buy the stock in that company immediately.
FABER: Yeah. No, I remembered at the time, to be fair, the stock was so far – the targets were so far above where the stock was, and obviously they were met.
MUSK: Yes.
FABER: Is the board working to your knowledge, on another potential plan if, in fact, that one never gets out of Delaware, the Supreme Court rules against the Chancellor’s ruling?
MUSK: I mean, I don’t want to speak on behalf of the board, but, you know, I’m sure it’s on their mind. But you know this, I can’t comment on Tesla board deliberations.
FABER: You said earlier. I mean, you want to stay as CEO. Why not – you know, I wonder, Elon, given all the heat you’ve gotten in a lot of ways, why not sort of become like an Ellison-like figure at Tesla? Obviously, you’re a lot younger, but still, he’s very influential at Oracle. But you know, he’s not CEO, and he doesn’t get the attention.
MUSK: He isn’t the CEO?
FABER: No, he’s not. Safra’s the CEO.
MUSK: Oh, okay.
FABER: Yeah. At least not in title. That’s a good point.
MUSK: No, no I think maybe a better term for – I’m a huge fan of Larry Ellison’s, good friend of mine. Owner.
FABER: Say it again?
MUSK: Owner, I think would be – he’s the owner of Oracle.
FABER: Yes, yes, he is. And he has a lot of it. Finally, I want to –
MUSK: He’s not the CEO, he’s just the owner.
FABER: He’s the owner. Yeah, it’s a big percentage, a lot higher than yours, actually at Tesla.
MUSK: Yes – price and percentage at Oracle.
FABER: Finally, I want to, I mean, so many things we haven’t gotten to, but we’re trying to keep to your time. Are we ready for the changes that AI is going to bring to this society? You know, I try in my job every day to sort of follow them. Obviously, we’ve been pushed – they’ve been pushed aside to a certain extent by issues of the day. But it’s coming fast.
MUSK: Yeah. It’s coming very fast. I mean, I feel like we’re, you know, we’re in the big bang of the intelligence explosion. Like we’re in – we’re watching. We have courtside seats to the big bang of intelligence explosion. One thing’s for sure, it won’t be boring, so –
FABER: No. You don’t seem to be saying there’s a 20% chance of annihilation as often anymore.
MUSK: Well, I think we should always consider that there’s some chance of a bad outcome, to try to protect against the bad outcome. We don’t want to be complacent and say that everything’s just going to be fine. There’s no chance of a bad outcome. You know, you know, I sort of think of this maybe in movie terms, as like, are we in a Star Trek movie or like are we in a Gene Roddenberry movie or a James Cameron movie? Which movie are we in here? And you could either have a Roddenberry or a Cameron outcome. And let’s, I think in this case, we want the Roddenberry outcome.
FABER: Yeah. And the abundance that may come with it, Elon, we got to stop there. I certainly appreciate your coming back. Obviously, there’s so much more we can talk about so I hope we’ll see you again soon.
MUSK: Pleasure.
FABER: Thank you for your time.
MUSK: Thank you appreciate it.
FABER: Elon Musk, CEO, of course, of Tesla – still the CEO, wants to be.
MUSK: CEO of insert company name.
FABER: You going to be back here more often now?
MUSK: I mean, this is mainly – Austin, Tesla headquarters, is where –is my main residence.
FABER: Are you going to miss the White House at all or no?
MUSK: I’ll – my rough plan on the White House is to be there for a couple days every, every, every few weeks. And to be helpful where I can be helpful.
FABER: Thank you. All right, that does it from the Gigafactory in Austin back to you guys.
