发布会·中文精读

马斯克与伊恩·达尔:SpaceX 愿景更新,2026年6月8日

Elon Musk and Ian Dahl, SpaceX Vision Update, June 8, 2026

中文精读由 AI 基于原始视频与文稿改写,非官方翻译,可能存在偏差或错误;表述以原视频为准。说明

太阳占了太阳系总质量的 99.86%。剩下的 0.14% 里,木星拿走绝大部分,地球小到可以忽略。人类关于能源、算力、增长的所有争吵,都发生在这粒尘埃上。

2026 年 6 月 8 日,马斯克和 SpaceX 星链团队的伊恩·达尔(Ian Dahl)做了一次愿景更新。开场白里,主持人把过去一年概括成"典型的 SpaceX 一年":首飞全新飞行器、收购 xAI 并改名 SpaceX AI、宣布太瓦级芯片工厂。马斯克接了一句,从来没有无聊的时刻。

这场对话想回答的问题只有一个:这些事怎么串起来,让人类变成多行星物种,顺便在卡尔达肖夫指数上往上挪一格。

一把外星人也认的尺子

卡尔达肖夫指数(Kardashev scale)由苏联物理学家卡尔达肖夫提出。它衡量文明进步的方式非常朴素:你能支配多少能量。

一型文明用尽本行星的能量,二型文明用尽母恒星的能量,三型文明用尽整个星系的能量。三个档位都是可测量、可比较的数字,不用考虑文化、制度、审美。假如有外星文明路过,这大概是他们唯一愿意拿来给人类打分的指标。

人类现在的位置很尴尬。按地球能接收到的太阳能算,我们利用的比例小到写不进表格。马斯克的原话很直接:

我们基本上哪儿都不在。我们甚至没有登记在册。(马斯克)

主持人补了一句,连"一微太阳"都算不上。马斯克回应,一微太阳相对现在已经是史诗级成就,那就是目标。

太阳占 99.86%,人类连零头都没摸到

先看质量。太阳占太阳系 99.86%,剩下 0.14% 里木星是主力,地球属于"其他"那一栏。伊恩说,把地球和太阳摆在一起,地球是一粒尘埃。

太阳的质量大约是地球的 33 万倍。这个倍数很难有体感,换个说法:如果太阳是一头蓝鲸,地球大概相当于它身上的一只跳蚤,而且还不是最大的那只。

再看能量。地球截面上接收到的太阳辐射,大约只有太阳总输出功率的五亿分之一。而这五亿分之一里,绝大部分还没法用,因为地球表面 70% 是水。马斯克顺口吐槽,外星人来了大概会问,明明大部分是水,为什么叫 Earth。

剩下 30% 的陆地也不乐观。南极、西伯利亚、加拿大北部这类地方占了一大块,没人愿意住,两极也拿不到多少太阳能。真正能铺太阳能板、又有人用的土地,面积小得可怜。

结论很硬:想在任何有意义的比例上使用太阳的能量,必须上天。

现在人类的用量是多少?不到太阳输出功率的万亿分之一。万亿是百万乘百万。粗略换算一下,人类一次能源总功率在 20 太瓦上下,太阳输出约 3.8 乘以 10 的 26 次方瓦,两者差着十几个数量级。访谈里说"不到万亿分之一",其实还算是客气的说法。

想做到百万分之一,也就是马斯克说的"一微太阳",文明的能量利用要提升远超百万倍。他把 1% 单独拿出来做参照:一个能用到太阳 1% 能量的文明,会比人类强大到难以描述。

这里有个容易被忽略的层次。从一型到二型的难度跳变,比从零到一型大得多。一型只需要把地球接收的那点太阳能吃干榨净,二型要在恒星尺度上做工程。马斯克说,三型文明怎么搞,现在根本不知道,可能得等 AI 自己想出来。

太空散热比地面便宜

过去几十年没人往太空铺太阳能板,原因简单,没有需求。太空发电的度电成本远高于地面,除非你有非做不可的理由。

理由来了:AI 要吃电,而且胃口越来越大。地面建电厂要面对选址、并网、冷却水、审批、居民投诉。太空里散热反而轻松,热量直接辐射进真空,不用冷却塔,不用河流,不用地下水。

散热在物理上有个残酷的约束。辐射功率和温度的四次方成正比,温度不够高,散热板就要铺得极大。地面机房靠风扇和水冷把热量搬运走,太空没有介质可搬,只能靠面积和温度硬扛。这也是为什么 AI 卫星最显眼的部分是两块大板子:一块收太阳,一块扔热量。

马斯克和伊恩把路径拆成三个瓶颈:把大量质量送上轨道、在太空发出足够的电、造出足够多的芯片。三件事同时成立,轨道数据中心才成立。少一件,整条链就断。

第一块短板:把质量送上天

Starship 解决第一件事。马斯克对它的定位下得很重:这是第一个有能力实现完全、快速重复使用的火箭设计。

重复使用为什么是前提?他打了个所有人都懂的比方:

如果每次飞完就把飞机扔掉,飞行会贵到几乎没人坐飞机,大家会开更多的车。(马斯克)

汽车、飞机、船、马、自行车,所有交通工具都能重复使用。火箭做不到,是因为地球引力井深、大气层厚,复用只是勉强够得着可行性边界。历史上不少团队试过完全复用,多数中途放弃,理由是不相信自己能成。

要让复用它成立,每个环节都得对:发动机、结构、航电、推进剂选择,还有极端的减重。着陆腿是死重,飞上去是负担,落下来还要再背一路,所以 Starship 改成用发射塔的机械臂直接夹住火箭。省下的每一公斤,最后都会体现在入轨成本上。

目前还没实现完全复用,马斯克希望在今年晚些时候做到。做到之后还有更难的一步:快速复用。火箭落地、被塔夹住、放回发射架,不做翻修、不做繁琐检查,像飞机一样再飞。

翻修才是成本黑洞。航天飞机当年也能重复使用,但每次回来要拆开大修,成本压不下来,最后退役。快速复用的意思是把检查变成例行保养,把大修变成换件。做到这一步,火箭才真正变成交通工具。

这也是 Starship 的分量所在。它是史上最大、最重的飞行物,也是推力最大的运动物体。V3 的推力是土星五号登月火箭的两倍多,到 V4 差不多是三倍。马斯克给出的长期节奏是,未来每小时飞一次以上。

第 12 次试飞送上去的载荷,是 SpaceX 历史上最重的一次,而那只是 V3 能力的一小部分。

眼下的份额已经很夸张。伊恩说,靠猎鹰 9 和猎鹰重型,SpaceX 承担了全球入轨质量的 85% 到 90%,剩下的大部分来自中国,包括美国在内的其他国家加起来只有 5% 到 7%。

Starship 要做的,是把每年约 2500 吨的入轨质量,抬到每年数百万吨。时间上,他们的估计是大约 3 年做到每年 100 万吨。

从 2500 吨到 100 万吨,是 400 倍。这个跨度没法靠多造几枚火箭实现,只能靠同一枚火箭反复飞。所以马斯克反复强调复用,理由不在技术炫技,而在算术。

太空数据中心长得不像机房

很多人对"轨道数据中心"的画面是错的,以为要把一栋楼装上发动机发射上去。实际要上天的东西小得多,核心是一批芯片,加上供电和散热。

伊恩说,很多人连地面数据中心的内部长什么样都没见过,脑子里是一团云。拆开看,机房就是一批芯片,加供电,加散热。搬到太空,问题变成两件:怎么发电,怎么把废热辐射出去。

有个反直觉的结论:AI 卫星比星链卫星简单。星链卫星有巨大的相控阵天线、抛物面天线、大量激光链路,结构复杂。AI 卫星基本上是一大片太阳能电池、一块散热器、少量激光链路,把尺寸放大就行。

伊恩提到,这些太阳能技术本来就要用在 V3 星链卫星上,AI 卫星相当于接着往上加尺寸,不需要凭空发明什么。马斯克也强调,这里没有魔法,绝大部分是已经做出来的东西。

一颗卫星就是一台机架

第一版设计的目标是峰值 150 千瓦、持续 120 千瓦算力。太阳能阵列按每平方米 250 瓦估算,散热器按每平方米 1400 瓦估算。散热器双面辐射,以刀刃姿态对着太阳,尽量少晒、多散热。伊恩说这两个数字是很务实的目标,未来还能往上提。

150 千瓦这个量级,差不多对应一台英伟达 GB300 机架。72 颗 GPU 的峰值功率约 140 千瓦,实际运行几乎不可能一直贴着峰值,平均 120 千瓦更现实。所以可以把它理解成一台飘在轨道上的机架。

这个类比值得停一下。现在地面数据中心的功率密度已经让供电和散热部门头疼,一台机架一百多千瓦,一个机房上万台机架就是上百兆瓦。太空方案把"机房"拆成一颗颗卫星,每个卫星自带电源和散热,靠激光链路组网。等于把集中式的电力问题,改成了分布式的制造问题。

机架之间用激光链路互联,单颗卫星的对地激光链路带宽在 1 Tb 量级。也可以直接接入星链星座,由星链用现有的 KA、KU 天线把数据传回地面,另外还有激光对地链路。

延迟不是问题。轨道高度 600 到 800 公里,光速每毫秒 300 公里,单程大约 3 毫秒。马斯克对"太空延迟高"的担心不太在意,光跑得挺快。这个延迟甚至优于不少跨洋光缆路径,因为光在光纤里的速度只有真空中的三分之二左右,还要绕路。

规模上,散热器翼展大约 70 米,和 V3 星链的太阳能阵列差不多大。听起来很大,放到太空尺度里什么都不算。伊恩说哪怕上千颗、上百万颗卫星,太空都装得下。马斯克补了一句,卫星贴近看很大,放到地球旁边小到看不见。

运营层面,SpaceX 有底气。现在在轨的星链卫星约 1 万颗,是唯一在这个量级上有实际运营经验的运营商。伊恩说,怎么把卫星排得足够密又飞得安全,是他们做星座时的头号目标。

这句话的分量在事故率。一万颗卫星在轨,意味着每天都有避碰决策,每次决策都要在燃料、寿命、通信连续性之间取舍。这套运营经验没法从论文里抄,只能靠飞出来。后来者想追,得先补上这一万颗的功课。

产能落在巴斯特罗普

制造放在得克萨斯州的巴斯特罗普(Bastrop)。现在所在的厂房已经在满负荷生产星链用户终端,新的终端产线也在开。太阳能制造设施已经开工,AI 卫星生产大楼很快跟上。

伊恩给出的节点是,明年底,AI 卫星、太阳能、终端几条线都跑到合理产量。

马斯克对终端的长期判断是,最终可能有几亿台星链终端在使用。直连手机(direct to cell)的星座会让手机直接连上太空,拿到高带宽,不需要额外的锅。这条线如果跑通,星链的用户上限就从"愿意装天线的人"变成"所有有手机的人"。

把这几件事放在同一个园区里,逻辑清楚:太阳能板、散热器、卫星总装、终端制造,共享供应链和工程团队。造卫星和造终端看起来是两门生意,实际都在练同一套功夫,也就是大批量、低成本、高可靠地造硬件。

第三块短板:芯片

芯片这件事分两步。第一步是把现成的芯片送上去。当前的参考设计是英伟达的 Rubin,也可以选 GB300,另外会做 TPU 的参考设计。理论上任何现有芯片都能上轨道。

问题是供给。伊恩估算,全球 AI 算力产能大概会走到每年 100 吉瓦的量级。这个数字听起来已经很大,放到太瓦面前还差一个数量级。

所以有了 terafab。马斯克给出的规模是大约 1 亿平方英尺,是特斯拉得州超级工厂的 10 倍。1 亿平方英尺差不多是 9.3 平方公里,相当于一千多个标准足球场连成一片。主持人开玩笑说,从这头走到那头可能需要 Starship 点对点运输。

主持人问,除了大,它和地球上其他芯片厂有什么不同。伊恩的回答是,长期看会有很多技术演进,但底层是规模。哪怕没有任何根本性的技术突破,硬把现有芯片制造技术放大到每年 1 太瓦的产出,也能走通,只是非常难。

换算一下,1 太瓦每年的逻辑芯片产出,相当于每年 10 亿颗功耗 1 千瓦的全光罩(full reticle)等效芯片,另外还要配上大量内存。全光罩是光刻机能一次曝光的最大面积,用它做单位,等于在说"每颗芯片都做到物理尺寸的上限"。

10 亿颗芯片每年是什么概念?现在全球先进制程的 AI 芯片年产量,大概在几百万到千万颗量级。要从这里跳到 10 亿颗,不只是多建几座厂,还要把光刻、封装、内存、供电、测试全部放大一个数量级。这也是马斯克说"相当困难"的原因。

时间表:每年一个数量级

很多人说轨道数据中心还要十年。马斯克给了一个明确但带免责声明的版本:这是他们要去试、并且认为自己大概率能做到的节奏,不是承诺。

明年底,太空 AI 算力的年化产能做到 1 吉瓦。

之后每年提升一个数量级。大约两年半,年化 10 吉瓦。三年半,可能到 100 吉瓦。

再往上,靠 terafab 和全球芯片制造的进步,冲到每年 1 太瓦,也就是 1000 吉瓦。这个数字是美国当前总用电量的两倍。马斯克说,需求应该会有,走着看。

主持人评价了一句,这是很多卫星。马斯克接话,可以做很多模拟。

每年提升一个数量级,这个斜率放在制造业里非常激进。航天行业过去六十年的常态是线性增长,偶尔阶跃。要用指数曲线,前提是每一环都不掉链子:火箭复用率、卫星产线节拍、太阳能板产能、芯片供给、发射窗口、监管许可。任何一环卡住,曲线就变成 S 形。

但换个角度,这个斜率在半导体行业并不陌生。摩尔定律当年也是每年翻倍,靠的是整个产业链协同压缩。马斯克显然想把芯片行业那套节奏搬到航天制造里。

月球上的电磁炮

太瓦还不够。马斯克说,别把目标定这么小,太瓦其实很小。

从每年 1 太瓦再往上提三个数量级,也就是 1000 倍,地球上做不到。他认为唯一的路径在月球,靠质量驱动器(mass driver)。

思路是这样:光伏、太阳能板、散热器尽量在月球本地生产,芯片从地球运,或者将来在月球造。大部分质量必须在月球解决,不能全靠地球运过去。月球没有大气,重力只有地球的六分之一,AI 卫星可以被加速到足够速度,直接抛进深空,根本不需要火箭。用的设备本质上是一台线性电机,也可以理解成电磁炮、轨道炮。

这套方案在物理上成立,在工程上是地狱级。月球表面要建光伏工厂、要采矿、要冶炼、要组装、要一条几公里长的加速轨道,还要承受每次发射的冲击和磨损。但它的好处是绕开了火箭最贵的部分:把质量从引力井里抬出来。

发布里放了一段动画,马斯克说,如果这都激不起你对未来的兴趣,那真没什么能了。他自己也很想看到月球上的质量驱动器建成。

还有个副产品。为了这个项目要往月球运那么大的质量,意味着任何想去月球的人都能去。马斯克说,每个人一生至少该去一次月球,想住那儿也行。

这套叙事该怎么读

把它当成一份工程路线图看,会发现几个值得琢磨的地方。

第一,需求端的逻辑变了。过去太空基建靠政府预算和通信市场,两个池子都不大。现在 AI 算力愿意为电和散热付高价,这是第一个能撑起大规模轨道基建的客户。算力是少数对成本不敏感、对供给极度敏感的商品。

第二,三个瓶颈的排序有讲究。质量入轨靠 Starship,电源靠星链太阳能技术的放大,芯片靠 terafab。前两个是 SpaceX 自己的强项,第三个要跨到半导体行业,难度最大,时间也最难估。

第三,时间表要打折看。马斯克给出的历史节点,方向大多对,时间大多乐观。判断进度不用听发布会,看三个硬指标:Starship 什么时候实现完全不翻修再飞、巴斯特罗普的 AI 卫星产线什么时候出第一批量产件、terafab 什么时候真正动土。

第四,卡尔达肖夫指数这套框架的价值不在预测,在换算。它逼着人把所有讨论落回瓦特和吨。文明进步可以被测量,这句话本身就有力量。

第五,风险一直都在。轨道碎片、频谱协调、发射失败、芯片供应、资本开支、监管审批,任何一项出问题都会拖慢节奏。马斯克自己也说,这是他们要去试的东西,不是承诺。

人类现在连一微太阳都不到。这个事实不太舒服,但也说明天花板有多高。

  • 想跟踪这条线,先盯三个可验证的里程碑:Starship 首次实现不翻修复用、巴斯特罗普 AI 卫星产线量产、terafab 开工。三个都动了,时间表才值得信。
  • 做能源或散热相关工作的,把辐射散热、功率密度、热管理这些基本功重新捡起来。太空数据中心最贵的部分在散热,不在算力。
  • 做芯片和封装的,关注全光罩等效产能这个概念。它决定 terafab 的规模是否成立,也决定 AI 芯片的长期供给曲线。
  • 做网络和协议栈的,研究激光链路组网和星间路由。上万颗卫星的运营经验目前只有一家公司有,这是稀缺知识。
  • 评估任何"轨道数据中心"项目时,先问三个数:峰值功率多少、持续功率多少、散热面积多大。答不上来,多半是 PPT。
  • 对时间表习惯性打三折,但别因此忽略方向。方向对了,晚几年到达,结论不变。
  • 如果对航天制造有兴趣,巴斯特罗普是个值得关注的地点。太阳能、卫星总装、终端制造正在往一个园区里聚。
YouTube 平台字幕(自动或人工),可能有识别错误,人名与数字请以视频为准。已按语句合并成段落,并去掉 um、uh 等口头填充词。 文稿来源:YouTube(自动/人工字幕)。

0:00All right, well, hello everybody and welcome. Hey, now I got Elon and Ian Ellis our Starlink team. Figured we'd check in. It's been a typical SpaceX year. Launched a brand new vehicle. Yeah.

0:12Acquired XAI now SpaceX AI. Announced a terra-sized chip-building project. And so Yeah, never dull moment.

0:20Yeah, never a dull moment. Typical year. And so let's kind of wanted to connect some of the dots on how this all feeds into making life multi-planetary, starting to climb up the Kardashev scale. Maybe show off some cool new AI stats stuff. Kind of start galaxy-sized. And bring people in with the Kardashev scale. What's the big picture?

0:41What's the big picture? What is the Kardashev scale?

0:45Like how do you decide what progress a civilization has made that's the most objective metric that any alien species, say visiting us, would calibrate how how much progress we've made as a civilization. And one of the most objective ways to do that is the amount of power that is any given civilization has been able to harness.

1:16and there was a Russian physicist actually by the name of Kardashev, who thought about this and it's it's good I think it's good way to characterize it, which is you can have you can you can assess how well a civilization is harnessing the power available on the planet. That's a type one. And then type two would be how much of the stars' power are you harnessing? And then type three would be how much of the galaxy's power are you harnessing?

1:51these are very objective and measurable numbers. so right now we're very low on the Kardashev one scale. Like if you say like what proportion of our planet's power are we harnessing? It's a very, very tiny number. and and and basically we've we're harnessing almost nothing of our star's power.

2:17So the the the sun is truly an immense thing. it is it's difficult with words to characterize just how immense the sun is, but this gives you sort of a sense of scale.

2:29Yeah, it's it's it's a big difficulty jump going from level one to level two.

2:35Very big difficulty jump, yes. And level three and we don't even know how to do level three really. We'll get there.

2:40Yeah, yeah, exactly. AI will figure it out, I suppose. One way to appreciate the the size the size of the sun is to think about how heavy is the sun compared to all the rest of the mass in the solar system. So the sun is about 99.86% of all mass in the solar system. It's ev- everything and and then all the remaining one, you know, 0.14% most of that is Jupiter. One planet.

3:07So we're still lightweight.

3:09Yes. the entire mass of Earth is in the tiny miscellaneous category. We're ex- we're like a Earth is a tiny dust mote compared to the sun.

3:20Well, but how much energy are we talking like coming from the sun especially compared to what we're using here on Earth? Cuz it feels like

3:29Yeah. The incident solar energy on the cross-section of the Earth is roughly a half billionth of the sun's power output. and and the vast majority of that we we cannot use because you know, 70% of Earth is water. Yeah.

3:46which technically our planet should be called water. because it is 70% water and I think an alien civilization visiting us would be like, "Why are they calling it Earth when it is mostly water?"

3:56Where are the Where are the Greenlands not green of the of the gala of the solar system?

4:00Yeah. a bunch of the the Exactly. You know even We're 70% water. And then of the 30% that's land, a bunch of it is I you know, Antarctica or you know, Siberia type of thing. Very northern Canada type of thing. Very difficult to Not not not places people typically want to live. And you're not going to get a lot of solar power in the at the poles.

4:25So, the actual usable area of land that where you can get solar power is quite small. Anyway, in order to ascend the Kardashev scale or in order to get to any meaningful percentage of the sun's energy harnessed, you have to go to space. If you wanted to get to say a millionth of the power output of the sun, you would have to increase civilizational energy harnessed by much more than a million.

4:58So, we currently use much less than a trillionth of the power output of the sun. and a trillion is a million times a million. so So, basically there's We're We're basically practically nowhere on on the sort of on the Kardashev two scale. Practically nowhere.

5:16So, on Kardashev scale, we're all still

5:20We're non-existent. We're We're non-existent. We're We're like not even We're not even Yeah. We're We're We're so We're Not We're not registering. Not even a micro soul. Yeah. We're No. And so, to actually

5:34One micro soul would be an epic epic achievement relative to where we are right now. Something to aspire to.

5:40Yeah. Yeah, that's our goal. And like this is I think both simultaneously an incredibly adventurous goal relative to where we are and and yet not particularly adventurous as a percentage of the sun's energy to try to achieve power harnessed being 1 millionth of what the sun outputs. And so to actually start A micro solar.

6:04To actually start getting there though, we're not just going to throw solar arrays in space, try to soak up a bunch of the sun. Like there has to be a need. Like you want to go up there and do something meaningful. And obviously until this point in human history like there hasn't really been a need. What has changed to make us think that like maybe now's the time to start trying to notch a percentage point or two?

6:28I mean getting to a percent of the sun's energy

6:32Maybe not a percent. Let's go like move the decimal point back a couple of years.

6:36extremely kick ass civilization if you get to 1% of the sun's energy. And I'm like, wow. That civilization is going to be vastly more powerful than us to say the least. Yeah.

6:49So in in order to start to make some progress on the Kardashev scale, we need to launch satellites to to to orbit Earth and capture solar power and that avoids the need to build massive power plants on Earth and deal with cooling cuz cooling is actually much easier in space than it is on Earth. you can just radiate to the vacuum.

7:17and and so what what we're proposing here and what we intend to do is to try to climb the Kardashev scale to I don't you kind of like a respectable civilization. so when the aliens hopefully there are aliens out there and they maybe finally decide to talk to us, you know, where we have where we where we have some respectable amount of the sun's energy being used. Yeah.

7:43that's not like totally pathetic. [laughter]

7:47Which is the current situation.

7:48And so before we start sending data centers, sending all of this to space, there are some limiting factors that we got to get there that would traditionally make it so like this is almost impossible.

8:00Yeah, what does it take to scale? Yeah.

8:02so things it takes to scale are you need to have a large mass to orbit capability, which is what Starship will give us. that large mass so Yeah, we ultimately need to send millions of tons to orbit and beyond. And you need the power associated with that. So if you want to put a 100 gigawatts or ultimately a terawatt into space from Earth, you need you will you will at some point need a terawatt of solar.

8:30and then you're going to need a terawatt of AI chips. So the three things on you need are mass to orbit, a lot of solar power, and and radiators of course. And a lot of chips.

8:42All right, well let's start ticking down the list. So mass to orbit, that's where Starship comes in. Yeah, we just had first flight V3, it's awesome. I know you were there. It was crazy to see that rocket launch yeah, and like long time coming. What's kind of what's Starship's kind of purpose of being?

9:02What is it going to be doing?

9:05Yeah, so Starship is going to it's going to revolutionize space really. It's it's the first rocket design that is capable of full and rapid reusability. Now reusability is the fundamental breakthrough that is necessary to make life multi-planetary as well as to ascend the Kardashev scale. You you simply cannot ascend the the Kardashev scale unless you have a re- reusable spacecraft, and you cannot extend life to the moon, to Mars, and to the rest of the solar system without a reusable rocket. the the the cost is simply prohibitive. You You can't You can't make enough rockets. Yeah.

9:42unless you fly unless you can re-fly them. just like any other mode of transport, you can imagine that if if we had to throw away airplanes every time we flew, flying would be far too expensive and basically no one would be flying airplanes. You'd be doing a whole lot more driving. Rapid reusability.

9:58Yes. [laughter] every mode of transport is reusable without which it's simply not viable as as a transport system. so cars, planes, boats, horses, bicycles are all obviously reusable. Yeah.

10:14with rockets, it's much harder to make a rocket reusable because Earth has a deep gravity well and a thick atmosphere. and these make it just barely possible to achieve reusability with a rocket. and there've been you know, many prior attempts to create a re- a fully reusable rocket.

10:35and they Most of those attempts have been abandoned partway through because they they didn't think they could succeed. in order to achieve full reusability, everything's got to be perfect. You got the engines, the structure, the avionics, the choice of propellants, You You've got to You've got to go to extreme measures for mass optimization, which is why we have the tower catch the rocket instead of putting on landing legs, which are heavy.

11:02the the rocket can simply be caught by the tower. And we haven't achieved full reusability yet, but we do expect to achieve that hopefully later this year with Starship. And then you you've got to achieve full reusability, then you've also you've you've got to go step beyond that, which is make it rapidly reusable such that the rocket lands, it gets caught by the tower, gets put back on the launch stand and can be flown again without any refurbishment or laborious inspection like an aircraft. Yeah.

11:32this is incredibly difficult. This is the first time that there's ever been a rocket where that is possible. That's what makes Starship so profound. It also happens to be the the largest flying object ever made, the heaviest flying object ever made, the most powerful moving object of any kind.

11:51Starship B3 is more than double the thrust of the Saturn 5 moon rocket. By version 4 will be pretty much three times the thrust of the Saturn 5 moon rocket. And we expect this expect Starship to be flying more than once per hour down the road.

12:08One of the fun facts from flight 12 that was actually the heaviest payload SpaceX has ever flown and that's still just a fraction of what V3 can do. So, Yes.

12:19I mean, once we're flying massive amounts really rapidly. I mean, we already fly the majority of payload to space with Falcon. Do people even really understand what mass orbit becomes when Starship is flying?

12:39It's it's many orders of magnitude greater than what is the case today. So, even with Falcon 9 Falcon Heavy SpaceX delivers almost 90% of all Earth mass to orbit. I think it's somewhere between 85 and 90% right now. And then most of the remaining mass I think is is launched by China and then the rest of the world including the rest of the US is the remaining I don't know, 5 to 7%.

13:09Now with with Starship, we're we'll be aiming to go from somewhere around 2500 tons a year to orbit to millions of tons per year to orbit. And to do so in a pretty short period of time. So, we think probably we can get to a million tons to orbit per year in in about 3 years thereabouts.

13:35Starship Starship is going to take care of the mass to orbit limiting factor. Yes.

13:41And then power generation. So first and and maybe you can help Sure.

13:47People probably struggle to visualize a little bit when you say like data center in space. Like we're not going to slap engines on a building and fly it up. They're like these actually look like pretty different. And so kind of walk through how you take something that's in a giant building on the ground and turn it into something that's functional in space.

14:06Yeah, I I think it's it's pretty interesting. A lot of people don't actually know what what the inside of a data center even looks like, right? Yeah.

14:12And it's some like mythical place where the the internet's in the cloud or something.

14:16Yeah, some people envision wires, some people envision boxes, but like if actually it comes down to a set number of chips and and the things that we need to launch into space are actually quite small when we look at it. The more challenging part is figuring out how to get how do you get the power for it? And and that's where a lot of what we've worked on for existing like Starlink technology this solar arrays are what we want to utilize that expertise to be able to build a satellite that can actually launch the critical components of the data center into space itself.

14:53We like to look at this and say like what is what is the actual engineering problem here? And and it's it's really a combination of delivering power and then taking the waste heat and energy away and sending it into the vacuum of space as you mentioned.

15:07Yeah. now the the AI satellite is actually much simpler than a Starlink satellite. It's a Starlink satellite has has gigantic phased array antennas. It's got you know, parabolic antennas. It's got let you know, a lot of laser links. it's a it's it's much more complicated than an AI satellite. An AI satellite is essentially a lot of solar cells, a radiator, and you still need some laser links, but you don't have all of the the super complex antennas that you have on a Starlink satellite. So, I mean, given the two, the easier one to design for is the the AI satellite. Yeah.

15:53It's just a little bit bigger. It's bigger.

15:56Just make stuff bigger, yeah. I was like, so we've got this is our AI one. If you guys want to walk us through.

16:03Yeah. Yeah, so so the first thing that we're we're really looking at here is like you first you got to make something compelling, right? And and we thought that the right place to start is around the 150 kW like peak power level. but as we look at the workloads with with our experience with XAI, we get to actually see the the we can also support about 120 kW of average compute. There's a difference.

16:27What we're showing here is kind of a a draft version of the version one of the of the SpaceX AI satellite. The AI one, I guess you could call it. and seems like a reasonable place to start is 150 kW peak power, 120 kW sustained power. And and to give you a sense of what does that actually look like in terms of the size of the radiators, size of the solar panels.

16:53the assumptions here are 250 W per square meter for the solar array. And about 1,400 W per square meter for the radiators. So, the radiators these are double-sided ra- radiators are radiating both sides. They're oriented knife-edge to the sun. And and and it's 1,400 W per square meter is a very achievable goal. over time we think we can probably do above 250 W per square meter and above 1400 W per square meter for the solar panels and radiators respectively.

17:27but this gives you like a The appearance is pretty much what the satellite's going to look like. It's a lot of solar panels, radiator, and then everything else is pretty small by comparison.

17:37And these are like evolutions of of things that we have actually already launched in in in our Starlink constellation to date. Yeah. That's that's really I think the the cool part to me is that we're we're looking at solar technology that we already are are going to use on on the V3 Starlink vehicles. So, I I'm like really excited to then just take those and make it bigger.

18:00Yeah. Part of what we want to want to convey here is that this this not some magic that's necessary that doesn't exist for the AI satellites. As Ian said, this is a lot of this is technology that we've we've already made for the Starlink V3 satellites. so, it's it's basically We don't think this is a a super hard problem compared to things we already do.

18:26There would also be probably something on the order of a terabit of connectivity of laser link connectivity from the from the satellite. The 150 kW peak power level is roughly matches what's a an Nvidia GV300 rack would do. So, if you you've got a GV300 with 72 GPUs, its peak power I think is around 140 kW.

18:50but it's rarely it just it's almost impossible to get it to to be at that peak power. A more reasonable operating envelope would be around 120 20 kW average power. But but it can peak up to 150. So, that's it's basically think of it as a a rack of compute in space. And then you can connect the these these racks of compute to either each other by the laser links or directly to the Starlink constellation. So, you can close the link with the Starlink constellation, and then Starlink can then send that data to the ground using the existing KA and KU antennas on the on the vehicle.

19:34it also has laser to laser links to the ground as well. So, And this this would not be at a particularly high latency. You know, we're we're talking about you know, maybe being around 6 to 800 km above the Earth. and light travels 300 km per millisecond. So, that's it's about you know, 3 milliseconds away, basically. It's not not very far.

20:02Won't worry about that too much, then.

20:03It's not So, many people worry and think there's going to be some like high latency. I'm like, yeah, no, speed of lightning moves pretty fast. Light moves pretty fast. [laughter] Yeah.

20:13I think the cool thing also is the the radiators themselves are about the same size as the existing solar arrays for the V3 vehicle. kind of kind of in that that realm where we're flying today. Yeah.

20:26So, I mean, they got they got about a 70-m wingspan, so these are fairly large. We're talking about building a lot of them and putting them up there, but they you like say like space is in the name, like there's there's a lot of space up there. And so, even when you're talking thousands or even, you know, up to a million satellites, Yeah.

20:47you got plenty of room to move around up there.

20:48Yeah, space is really big. So, it's not like it's not like space is going to get crowded. space is is enormous. Like, if you zoom in close to the satellite, it looks big, but if you actually look at it relative relative to the Earth, the satellites are so tiny, you can you can't even see them.

21:06So, they're they're very, very tiny compared to us.

21:11And I mean, we have 10 about 10,000 Starlinks in orbit right now. We've got a pretty good idea of how to operate just really large constellations and do it safely now, right?

21:23We We are the only operator that has any experience at that scale. it's it's a great thing that we you know, we have this background, so we know how tightly we can pack the satellites and and and fly them safely. That's the that's the our number one goal when when we look at the constellation.

21:38We're going to be building a lot of satellites, and we're going to be building them here in Bastrop, right? So, we've we've got this, which Yeah. So, we're we're in that building kind of in the middle, which

21:49Yeah, we're sitting in that building right now.

21:51This is my first time here. The building is massive. Like you you come around the corner, you see it through the trees, and you're like, "Oh, wow." But, we're about to kind of put this building to shame, aren't we?

22:03yes, we're going to In fact, we already have the solar manufacturing facility it's under construction already. And and then we will be building out the AI satellite production building soon. and yeah, so we expect to have the the AI satellite AI satellite production, the solar production, and all of that operating at some reasonable volume by the end of next year.

22:33So, if anybody wants to work on my AI satellites, this is kind of going to become the hub of that. We're also So, I mean, like right behind us, the machines are humming. They're we're still making all of our user terminals for Starlink here. That's not going anywhere. In fact, we're turning on new production lines for new units, right?

22:51yes. in fact, these are the new Starlink terminals, which we made in much higher volume than than the current terminals. you know, ultimately, we think there's probably going to be a few hundred million Starlink terminals out there. And then our the Starlink direct to cell constellation will connect directly to people's cell phones and enable high bandwidth communication directly from your phone to space.

23:17All right. We're We're two limiting factors down. We've got mass to orbit. Got putting solar in the queue. Third one's chips.

23:26Yes. so at least in the in the beginning we can obviously launch the the chips that are already being made. So our current reference design is for Nvidia Rubin chips or could be either GB300 or or Rubin chips. So and we'll also have a reference design for TPUs and and essentially you can put up put any any existing chips into into orbit. but the the current industry seems to be it's it seems like it's going to I don't know get to maybe around a hundred gigawatts a year of of AI compute.

24:11But it that that doesn't [clears throat] answer the question of well, how do you get to a terawatt? That's why you need the terafab.

24:20Always looking a step bigger. Yeah.

24:22Yeah, in order to get to the next order of magnitude you need a gigantic chip factory. to give you a sense of scale here we expect that the terafab is going to be around a hundred million square feet. which is ten times the size of the the Tesla gigafactory Texas.

24:44And what aside from just, you know, I'm going to need Starship point to point to get from one end to the other. Aside from just the size, what's going to make this unique different from any other chip building operation on the planet?

24:58Well, I think over time there's going to be a lot of technology evolution with the Terra fab. But fundamentally it's about scale. So even if there were no fundamental technology breakthroughs, it's and and you simply you you could simply scale the existing chip making technology

25:18with a lot of difficulty to a terawatt of chip output per year. that's if you look at it just from the logic die standpoint, that's that's equivalent that's like having a billion chips per year with a a kilowatt per radical. So there's a a billion full radical equivalent chips each doing a kilowatt. And then you're going to need a lot of memory to go with that.

25:43A lot of people today have been saying orbital data centers were like a decade away.

25:49Yeah, I think we want to try to give people a sense of of the time frame. we at least the time frame we're aiming for. I mean, you know, people should take this with a grain of salt to some degree because this is this is just our best guess. So this is not a this is not a promise of what we'll do. This is what we what what what we're going to try to do and think we probably can do.

26:11which is to get to roughly an annualized rate of a gigawatt per year by the end of next year in terms of space AI compute. and then aspirationally scale that by an order of magnitude per year. So in 2 and 1/2 years hitting an annualized rate of 10 gigawatts a year to space, in 3 and 1/2 years maybe 100 gigawatts. And then depending upon what progress there is in chip making in the rest of the world and with the Terra fab going beyond that to scale to a a terawatt per year, which is 1,000 gigawatts.

26:48Which is that that's twice the the current electricity consumption of the United States. Yeah.

26:53I think there'll will an appetite for that, but we'll see.

26:56It's a lot of satellites.

26:58I don't know what you're going to think about, but Maybe do a lot of simulations or something.

27:02Yeah. So, after we've, you know, worked out there are all the limiting factors, we've kind of topped out what we can do on Earth, what is the next step to again, try and actually notch maybe some percentage points towards becoming Kardashev level two.

27:22Why stop there? Why think small? Cuz a terawatt actually is very small. small.

27:28Let's not think small. So, there is in order to get to another three orders of magnitude to 1,000 X from a terawatt per year, the the only way that we can really see see that you can achieve that is on the moon with a mass driver. Essentially, where you do local production of photovoltaics and solar and radiators on the moon.

27:54Maybe you bring the chips from Earth or you could conceivably make the chips on on the moon. And but you need most of the mass to be made on the moon, so you don't have to transport it to the moon from Earth. And and then because the moon has no atmosphere and only 1/6 Earth's gravity, you can you can get you can accelerate the AI satellites into deep space without a rocket. So, you can basically shoot them into space using an electromagnetic gun, like a like a rail gun type I mean just it's basically a linear electric motor is a way to think about it.

28:31Is that a way I think we can show people? [music] [music]

30:21I mean if that doesn't get you excited for the future, I don't really know what will.

30:26I'm fired up to sit to see a mass driver on the moon. That would be very cool. Yeah. Sci-fi future. Yeah.

30:33it would also mean that if if we're if we're bringing that amount of mass to the moon, it would mean that anyone who wants to go to the moon will be able to go to the moon. And I think that'll would pretty cool.

30:42Yeah. I'm I'm going to be jumping first in line to get up there.

30:47Yeah, I mean you know, everyone should go to the moon at least once, I think. I know. Just once. Yeah.

30:51You can move there if you want. You can go live on the moon. We'll see. [laughter]

30:55Thanks, guys, for chatting with me for a little bit. Make excited to see whole new type whole new kind of satellite whole bunch more Starship launches more chips, more solar, more more everything. It's It's a big future, but I'm excited to see everybody at this company go out and build.

31:13All right, sounds good. It's exciting. Thanks, guys.

资源与出处

相关视频

同频道、同主题或相邻年份