SpaceX：从猎鹰 1 号到星舰

为了理解宇宙的本质和这些根本性问题，我们必须扩展意识的范围和规模。距离上次公司演讲已经有一段时间了，我以后会尽量每三个月在财报电话会议之后做一次，因为基本上只有那时候我才被允许谈论这些事情。所以，我们首先从 SpaceX 的历史讲起，从最初一直讲到现在。我觉得偶尔回顾整个故事是好的，然后思考你希望故事的下一个章节是什么，你希望故事的下一本书是什么。

In order to understand the nature of the universe and these fundamental questions, we must expand the scope and scale of consciousness. So, it's been a while since the last company talk, and I'll try to do these now every three months after our earnings calls, which is when I'm allowed to talk about things essentially at this point. So, we're going to start off just by going through the history of SpaceX from the beginning and then to present day. I think once in a while it's good to look at the whole story and then think about what you want the next chapters to be in the story, what you want the next book to be in the story.

我们一开始就有一个明确的使命：让生命成为多行星生命，将意识扩展到地球之外，去理解宇宙。而只有当你走出去探索，看到其他行星，最终看到其他恒星系统，你才能真正理解宇宙。我认为，人们能想象的最鼓舞人心的未来之一，就是我们置身于群星之间，或许遇到外星文明，或者探索那些延续了百万年却在千万年前消亡的文明的遗迹。我觉得这就是让我早上醒来思考什么才是鼓舞人心的东西，因为这个世界需要鼓舞人心的事物。生活不能只是解决一个又一个问题。必须有让你期待未来的东西，而成为一个太空文明是我们能为未来做的最有趣的事情。这就是让《星际迷航》成为现实，你知道的。

So we started off with a clear mission, which is to make life multiplanetary, to extend consciousness beyond Earth, to understand the universe. And you can only understand the universe if you go out there and explore and see other planets and ultimately other star systems. And I think one of the most inspiring futures one could possibly imagine is one where we're out there among the stars and perhaps meeting alien civilizations or exploring the ruins of civilizations that lasted a million years but died out 10 million years ago. I think that's what makes me wake up in the morning and think about what is inspiring, because we need things in this world that are inspiring. Life cannot just be about solving one problem after another. There have to be things that make you look forward to the future, and being a spacefaring civilization is the most interesting thing we could possibly do for the future. It is making Star Trek real, you know.

嗯，我们就是这样起步的。所以，我不知道，也许我应该再请一支马里亚奇乐队什么的。你知道，为了这次聚会的结尾，我们会请很多马里亚奇乐队来庆祝。但这就是 SpaceX 最初的全部，只是埃尔塞贡多的一间小仓库，然后是猎鹰 1 号。我们在 SpaceX 起步时真的不知道如何制造火箭。显然，我们炸了三枚火箭才最终进入轨道。现在看起来几乎有点可爱了。

Well, this is how we started out. So, I don't know, maybe I should have another mariachi band or something. You know, for the end of this party, we're going to have a lot of mariachi bands in honor of this. But this was all of SpaceX in the beginning, just a small warehouse in El Segundo and then Falcon 1. And we really had no idea at the start of SpaceX how to actually make rockets. And obviously we blew up a rocket three times before finally reaching orbit. It looks almost really cute at this point.

然后我们经历了漫长的努力，试图让轨道级助推器着陆。所以，弄清楚如何让轨道级助推器着陆并重新飞行并不容易。但过了一段时间，我们变得擅长了。这需要一支非常有才华和敬业的团队进行多次尝试。现在我们已经达到了每隔几天就能起飞、着陆并回收助推器的地步。

And then we went through the long effort of trying to land an orbital booster. So, it wasn't easy figuring out how to land an orbital booster and then refly them. But then we got good at it after a while. It took many attempts from a very talented and dedicated team. And now we've gotten to the point where we take off and land and recover a booster every few days.

顺便问一下，我身后的噪音是不是有点太大？你们能听到我说话吗？

By the way, is the noise behind me a little too loud? Can you guys hear me?

好的，好吧。所以，反思实现目标的困难总是重要的，因为当我们试图让助推器着陆时，没有人真正相信我们能成功。航空航天业的许多专家都告诉我，你不会成功的。即使你成功了，也不会有人买一枚基本上是用过的火箭。这真的没有道理，因为飞机一直在重复飞行。坦白说，我宁愿坐一架飞机的第一千次飞行，也不愿坐它的第一次飞行。所以我们把术语改成了“经过飞行验证的助推器”，这是事实。

Okay. All right. So, it's always important to reflect upon the difficulty of achieving things, because when we were trying to land the booster, no one really believed that we could succeed. And I was told by so many experts in the aerospace industry that you're not going to succeed in doing this. And even if you do, no one's going to buy a rocket that was, you know, a used rocket essentially. Which really makes no sense because airplanes are reflown all the time. And frankly, I'd rather be on the thousandth flight of an aircraft than its first flight. So we changed the nomenclature to 'flight proven booster,' which is true.

所以我们成功造出了猎鹰重型，这是 Starship 之前世界上威力最强的火箭，至少是已投入运营的。想到我们即将迎来第 700 次飞行，真是令人惊叹。这是对猎鹰团队、发射团队和制造团队出色工作的巨大证明。我想为每个人鼓掌，感谢他们做了如此出色的工作。

So we managed to do Falcon Heavy, which is the most powerful rocket in the world before Starship, operational at least. And it's amazing to think that we've got our 700th flight coming up. So it's a huge testament to the incredible work of the Falcon team, the launch team, and the manufacturing team. I would just like to give everyone a hand for doing such an amazing job.

当然，这很经典。顺便说一句，我完全预料到这次发射会失败。所以那是我在洛杉矶开了大约三年的那辆车。因为我想，嗯，这是猎鹰重型的首飞，很可能会失败。但我们不想在上面放一些无聊的东西，比如我觉得波音或别的公司会放一块混凝土块，那太无聊了。所以我们放了 Starman 在飞行，他还在那里。那辆车在地球和火星之间运行，我想它至少会在那里待上 1000 万年。

And of course, this is classic. I fully expected this light to fail, by the way. So that's the car that I was driving around LA for like three years. And because I thought, well, it's first flight of Falcon Heavy, probably going to fail. But we don't want to put something boring on it, like I think Boeing or whatever would put like a concrete block, which is extremely boring. So we put Starman flying around, and he's still up there. The car is orbiting between Earth and Mars, and it'll be up there I think for at least 10 million years.

然后当然是龙飞船，第一艘与空间站对接的私人开发航天器，以及第一位商业宇航员。我们现在已经执行了许多宇航员任务。我想用不了多久我们就会超过 100 名宇航员进入轨道并返回。所以龙飞船一直是一款了不起的飞行器。到目前为止，我们已经搭载了 78 名机组人员，这是一个史诗般的成功。所以我也想祝贺龙飞船团队。干得漂亮。

And then of course Dragon, first privately developed spacecraft to dock with the space station, and first commercial astronaut. We've now done many astronaut missions. I think it won't be long before we get past 100 astronauts to orbit and back. So Dragon has been an amazing vehicle. We've flown 78 crew members to date, and it's been an epic success. So I'd like to say congrats also to the Dragon team. Excellent work.

然后是星链，就像，嗯，如果你有一枚能频繁飞行的火箭，你在上面飞什么呢？所以星链团队从零开始构建了一个完整的全球互联网系统。这是有史以来第一个高带宽的全球互联网系统。我实际上认为，将来它可能会承载大部分互联网流量。事实上，我认为它最终可能会承载超过 90% 的互联网流量在星链上。这真的很不可思议，近 11,000 颗卫星在轨，比所有其他卫星总和的两倍还多。随着星链 B3 及以后的版本，我们将达到 100,000 颗卫星。所以这真的非常了不起。在太空中构建互联网非常困难，就像在太空中重建整个互联网，所有的硬件，所有的软件。然而团队做到了。所以向星链团队表示巨大的祝贺。

And then Starlink, it's like, well, if you've got a rocket that can fly a lot, what do you fly on it? And so the Starlink team has built from nothing an entire global internet system. And it's the first high-bandwidth global internet system ever to exist. And I actually think that down the road it will probably do a majority of the internet. In fact, I think it might end up doing over 90% of all internet traffic on Starlink. Which is really incredible to think, almost 11,000 satellites in orbit, that's more than twice as many satellites as everyone else combined. And with the Starlink B3 and beyond, we'll go to 100,000 satellites. So that's really quite remarkable. It's quite difficult to build the internet in space, the entire like rebuild the entire internet in space, all of the hardware, all of the software. And yet the team has done it. So huge congrats to the Starlink team.

我们现在覆盖 167 个国家，有 2200 万移动用户，1300 万高带宽用户，而且就像我说的，预计这些数字会随着时间的推移增长超过 10 倍。长期来看，事实上 Cloudflare 最近提到，根据目前的增长，他们认为五年内 AI 流量将比人类互联网流量多出一千倍。让这一点沉淀一下。你知道，这是大量的 AI。

We're now in 167 countries, 22 million mobile subscribers, 13 million high-bandwidth subscribers, and like I said, expect these numbers to increase by more than a factor of 10 over time. And long-term, probably the fact is Cloudflare just recently mentioned that based on current growth, they think AI traffic will be a thousand times more than human internet traffic in five years. Let that sink in for a minute. You know, it's a lot of AI.

然后我们还有星盾，它已经成为国家安全的重要组成部分。所以有限制，出于显而易见的原因我们不能对星盾说太多，因为它是绝密的等等。但星盾团队做了惊人的工作，它现在已经成为美国国家安全的重要组成部分。所以祝贺他们。

And then we've got Starshield, which has become a vital part of national security. So there's limited, we can't say too much about Starshield for obvious reasons because it's top secret and stuff. But the Starshield team has done amazing work, and it is now a vital part of America's national security. So congrats.

是的，Starship 太惊人了。哇。很难相信人类造出了这样的东西。我认为未来的 AI 会回顾并说：“对一群人类来说，干得不错。”所以，是的。所以 Starship 显然是一款不可思议的飞行器。

Yeah, Starship is amazing. Wow. Hard to believe that humans made such a thing. I think the future AI will look back and say, 'Not bad for a bunch of humans.' So, yeah. So Starship is obviously an incredible vehicle.

它将把我们的年入轨运力从大约 2500 吨提升到超过 100 万吨，我认为甚至可能达到每年 1000 万吨。这些数字令人难以置信。但我敢说，Starship 极有可能实现每年超过 100 万吨的入轨运力，甚至可能接近 1000 万吨。这显然令人震惊。但要在月球、火星及更远的地方建立文明，就需要这样的运力。如果你无法将数百万吨物资送入地球轨道，就无法将所需的数百万吨物资送到月球和火星，使我们成为一个可持续的多行星物种。所以这是一款不可思议的飞行器。就像我说的，我认为这可以说是人类有史以来尝试过的最困难的事情。Starship 团队在发动机、航空电子、结构、发射操作等各个方面都做得非常出色。让我们为 Starship 团队热烈鼓掌。

And it's what will take us from roughly 2,500 tons to orbit per year to over a million tons to orbit per year, and I think maybe even to 10 million tons per year. These are mind-boggling numbers. But I'd say it's highly likely we do over a million tons a year to orbit with Starship, and probably close to 10 million. Which is, you know, mind-boggling obviously. But that's the kind of tonnage that you need in order to build a civilization on the Moon, Mars, and beyond. If you can't deliver millions of tons to Earth orbit, you can't get the millions of tons that you need to the Moon and Mars to make us a sustainably multiplanet species. So it's an incredible vehicle. Like I said, I think it's arguably the hardest thing that humanity has ever tried to do. And the Starship team has done an amazing job across the board with the engines, avionics, structures, launch operations. So big hand for the Starship team.

那么 SpaceX 的 AI 业务将走向何方？令人惊讶的是，我们的 AI 收入——不是可能，而是肯定——将在 9 月（也就是下个月）超过 SpaceX 所有其他收入，并在第四季度大幅超过其他所有收入。所以 AI 已经成为 SpaceX 未来极其重要的一部分。我认为我们在 AI 上取得成功至关重要，不仅在硬件上，也在软件上。所以，未来从根本上说是 AI 和机器人的。假设文明继续进步，我认为数字智能的数量可能会超过生物智能数量的万亿倍。所以不是与人类智能相当，而是可能高出万亿倍。我认为 10 亿倍是一个容易的预测，可能达到万亿倍。当你利用越来越多的太阳能时，将其转化为智能的能力从根本上说是数字计算机的事情，因为人类很难在没有支持的情况下在深空生存。我的意思是，有些数学问题很有趣：如果你将文明能源使用量（即我们作为文明所利用的能量）增加一百万倍，你仍然只使用了太阳能量的一百万分之一。显然，其中大部分在深空，但未来很大程度上是 AI 的。我想这本来就会是这样，但重要的是我们有一个关心人类、培育人类、帮助我们前往其他行星和其他恒星系统的 AI。

So where are things headed in terms of SpaceX AI? The wild thing is that probably our AI revenue—not probably, definitely—will exceed all other SpaceX revenue probably in September, like next month, and will significantly exceed all other SpaceX revenue in the fourth quarter. So AI has become an extremely important part of SpaceX's future. I think it's actually vital that we succeed in AI. Not just in hardware but also in software. So yeah, the future is fundamentally AI and robots. Assuming civilization continues to progress, I think the amount of digital intelligence will probably be more than a trillion times the amount of biological intelligence. So not equal to human intelligence, but probably a trillion times higher. I'd say a billion times is an easy prediction, and probably a trillion. So as you harness more and more of the sun's energy, the ability to turn that into intelligence is fundamentally a digital computer thing, because it's pretty hard for humans to go live in deep space without any support. I mean, some of the math is interesting to think about: if you increase civilizational energy usage, or the amount of energy that we have harnessed as a civilization, by a factor of a million, you would still be using much less than one millionth of the sun's energy. Now, most of this is in deep space, obviously, but the future is very much AI. And I suppose it was always going to be that way, but it's important that we have an AI that cares about humanity, that fosters humanity, and helps take us to other planets and other star systems.

这就是为什么我认为 SpaceX 不仅要在 AI 硬件上成功，还要在 AI 软件上成功，这一点至关重要。我认为一个可以说是 SpaceX 的“孩子”的 AI 将是一个非常优秀的 AI。我认为 SpaceX 汇集了地球上一些最优秀的人才，无论是在能力、道德还是善良方面。我希望我们拥有一个诞生于这种能力、道德和善良的 AI。我认为这非常重要。所以我想鼓励大家在地面的 AI 硬件、AI 软件以及当然还有 AI 卫星方面提供帮助。太阳能将是其中非常重要的一部分。Terra Fab 将是其中非常重要的一部分。但我们也必须在软件方面取得成功。我们将用 SpaceX 的所有信息来训练 Grok。所以从某种意义上说，它将基于你们进行训练。你们实际上将成为这个 AI 的父母。它将继承你们的思想、想法和信念，我认为这是一件好事。所以，我们必须在太空、AI 硬件和 AI 软件方面取得胜利。这可能是最重要的信息。我认为这实际上是我今天想传达的最重要的信息：我们必须在 AI 上获胜，因为未来主要是 AI 和机器人。我们最终将无法控制 AI；它太聪明了。但就像如果你有一个天才儿童，你仍然可以向这个孩子灌输你认为好的、正确的价值观和信念。所以这就是为什么我们在 Grok 上取得成功极其重要。

And this is why I think it's essential that SpaceX succeed not just with AI hardware, but also with AI software. I think an AI that is sort of SpaceX's baby will be a very good AI. I think SpaceX is a collection of some of the very best humans on Earth, both in capability and morality and goodness. And I think we want to have an AI that is born of that capability and that morality and that goodness. I think that's very, very important. So I'd like to encourage people to help out actually with the AI hardware on the ground and the AI software, and of course with the AI satellites. Solar is going to be a very important part of that. The Terra Fab will be a very important part of that. But we must also succeed on the software front. And we're going to be training Grok on the sum total of all SpaceX information. So in a way, it will be trained on you. You will effectively be the parents of the AI. It will inherit your thoughts and ideas and beliefs, and I think that's a good thing. So yeah, we got to win here on the space, on the AI hardware and the AI software. That maybe is the most important message. I think that is actually the most important message I wanted to convey today: that we must win on AI, because the future is overwhelmingly AI and robots. We won't ultimately be able to control the AI; it'll be too smart for that. But just like if you have a child that is a super genius child, you can still instill in that child the values and beliefs that you think are good and right. And so that's why it's incredibly important that we succeed with Grok.

所以，是的，Grok 4.5 你们可能已经试过了。我们大约一周后推出 4.6，然后 4.7 应该会非常特别。所以我想鼓励 SpaceX 的每个人使用 AI 并让它变得更好。我们已经建造了世界上最强大的 AI 训练集群。这是我们预计到明年年底要做的：大约是迄今为止的 10 倍。这将是太空 AI 非常重要的实践，我确实认为 AI 训练将继续在地面进行，但 AI 推理——AI 的日常使用——将在太空中进行。所以我们的目标是在明年年底达到 10 吉瓦。这取决于每瓦的价值。如果每瓦的价值可能在 30 到 50 美元，这意味着如果我们在明年年底上线 10 吉瓦的 AI，年收入将达到 3000 到 5000 亿美元。大数字。

So yeah, well, Grok 4.5 you've tried probably. We've got 4.6 coming out in about a week, and then 4.7 should be really pretty special. So I'd like to encourage everyone at SpaceX to use AI and to make it better. So we've already built the most powerful AI training clusters in the world. And this is what we expect to do by the end of next year: about 10 times more than what we've done thus far. This will be very important practice for space AI, and I do think that AI training will continue to be on the ground, but AI inference—the day-to-day use of AI—will be in space. So we're aiming to get to 10 gigawatts by the end of next year. This is depending on the value per watt. If the value per watt is probably going to be $30 to $50, which means if we bring 10 gigawatts of AI online by the end of next year, it will be $300 to $500 billion a year in revenue. Big numbers.

有时我会想，我们的目的是什么？可能人类的目的是孕育一个“星灵”，一个具有感知能力的太阳。说实话，我不知道 AI 会怎么想。这是大量的算力。但我想这是文明演化的下一步：利用太阳的力量。我的意思是，有一位叫卡尔达舍夫的物理学家认为，你可以根据一个文明是否能够利用其行星的大部分能量、其太阳的大部分能量，或其星系及更远的大部分能量来对其进行分类。所以这就是将要发生的事情，我们将做到这一点。星灵的工厂将建在附近，基本上就在我们旁边。它们来自太空。太神奇了。

And sometimes I wonder what is our purpose? And it could be that the purpose of humanity is to give birth to a star mind, a sentient sun. I don't know what the AI is going to think about, honestly. It's a lot of compute. But I guess this is the next step in the evolution of a civilization: to harness the power of the sun. I mean, there was a physicist called Kardashev who argued that you could consider any given civilization on the basis of whether it has been able to harness most of the power of its planet, or most of the power of its sun, or most of the power of its galaxy and beyond. So that's what's going to happen, and we're going to do it. The factory for Star Mind will be located nearby, adjacent to us basically. They just come from space. It's amazing.

所以，但从长远来看，在不久的将来，可能在四五年内，AI 将占 SpaceX 价值的 99%。我敢说五年后肯定如此。AI 将占 SpaceX 价值的 99%，而 SpaceX 的价值将是一个天文数字。

So, but long term, not that far into the future, probably in four or five years, AI will be 99% of the value of SpaceX. I'd say five years for sure. AI will be 99% the value of SpaceX, and the value of SpaceX will be some astronomical number.

这才叫有趣的未来。一个非常有趣的未来。就像罗伯特·海因莱因的《严厉的月亮》，但用质量驱动器在月球上建造 AI 计算机并将其送入深空。我真的很想看到那东西投入运行。

Now that's what you call an interesting future. A very interesting future. It's like Robert Heinlein's 'The Moon Is a Harsh Mistress', but mass drivers building AI computers on the Moon and sending them into deep space. I'd really like to see that thing in operation.

这样做的一个好处是，SpaceX 的任何员工，只要想去月球或火星，将来都能去。

One of the nice benefits of this also is that anyone at SpaceX who wants to go to the Moon or Mars will be able to go in the future.

我向你保证。在月球上开派对会很有趣，你知道的，到处蹦跳。所以，是的，那将是一个极其激动人心的未来，我们成为太空文明，太阳系的每个角落都有人类，最终还会前往其他恒星系统。所以我认为这一定会实现。

You have my word. It'd be pretty fun to have a party on the moon, you know, jumping around. So yeah, it will be an incredibly exciting future where we're a spacefaring civilization and there are people on every part of our solar system and eventually going to other star systems. So I think this will happen.

所以，是的，你会让它成真。谢谢你。

So yeah, and you're going to make it happen. So thank you.