斯科特·诺兰 - SpaceX、创始人基金与美国铀浓缩的重建 - [像最好的投资者一样投资，第467期]

你对政府和科技的关系有什么新的认识？

What have you learned about governments and tech?

我在SpaceX的经历是，政府其实很想和私营企业合作，一起解决问题。

My experience through SpaceX was the government wants to work with private industry and wants to solve problems.

在SpaceX，我跟着汤姆·穆勒做发动机系统，负责大量结构和热防护工作，确保这些系统能扛住各种极端环境，同时做到低成本和高性能。

At SpaceX, worked on the engine systems under Tom Mueller and did a lot of the structural thermal work on those and helped make sure that they would stand up to the environments and everything and be low cost and high performance.

我们最终追求的，就是这两点。

That's ultimately what we were optimizing for.

等发动机系统稳定下来后，我就转到了龙飞船项目。

And then once those engine systems were working really well, I moved over to the Dragon capsule.

在龙飞船项目里，我们参与的是NASA的商业轨道运输服务计划，也就是COTS。

And in Dragon, we were on the NASA COTS program, Commercial Orbital Transportation Services.

这是一个投入数亿美元的项目，目标是实现两大能力：一是向空间站运送物资，二是最终发展出能与空间站对接的载人飞船。

And it was a multi $100,000,000 program to bring back two things, capability of launch to space station and then a first a cargo vehicle and then ultimately a crewed vehicle that could dock with the space station.

我最后那两年主要专注在龙飞船上，和NASA紧密合作，因为我们即将对接的，可能是人类历史上最昂贵的航天器。

The last year or two I was there was really focused on Dragon and working very closely with NASA because we were gonna go dock with, I guess, the most expensive asset mankind had ever developed.

所以，大家最不希望看到的，就是私营公司的货运飞船在对接空间站时出任何状况。

And so the last thing anybody wants is any sort of issue with a private company's cargo capsule docking with the space station either

比如撞上去之类的。

Smashing it or something.

对，如果是碰撞，那确实是个大问题。但轨道运行的规律是，即使出现偏差，其实还是有一定空间去避免碰撞的。

Yeah, like you would think collision, then yes, that's an issue but the way orbits work, it's actually there's some ability to avoid it even if things are off track.

更难把握的其实是热环境和压力这些方面。比如，飞船上有太阳能电池板，飞船本身会吸收多少热量？这些热量传导回空间站后，空间站能不能承受得住？

The harder thing to really get a handle on was things like thermal and pressure and okay, you've got this vehicle and it has solar panels and what's the heating on the vehicle and what thermal load is that driving back to the space station and can that handle it?

才会有这一大堆对接的技术要求。

So all these interface requirements.

所以我们得和NASA紧密合作，把这些要求都落实好。

And so we were working closely with NASA and those.

而且这些NASA的员工都非常聪明，他们真心热爱航天事业，很多人已经在NASA工作了几十年，哪怕那段时间航天领域的发展并不算活跃，他们也一直坚守着。

And ultimately, these were incredibly smart people who believed in space, who had been working at NASA for decades and cases despite not a ton of growth in space activity.

他们能坚持留在那里，就是因为心中有这份信念，也热爱这份事业。

They were there because they believed it and they loved it.

能跟一家公司合作，一起探讨怎么推进，比如：我们有空间站，也有这个项目，接下来怎么帮你们达成下一个里程碑？

And the opportunity to work with a company to like, okay, we have the space station, we have this program, how do we get you guys to the next step in the milestones?

我们怎么协作，既能确保安全，又能让项目真正落地？

How do we collaborate to like make sure this is safe but that it actually happens?

合作非常顺畅，氛围也很积极。

Super collaborative, very positive.

所以我的体会是，在那些有真正信仰、在政府机构里深耕多年的行业里，他们之所以留在这里，是因为真心希望事情能成。只要公司靠谱，并且像机构一样重视安全和性能，双方就会非常开放、充满热情地一起合作。

So I think my takeaway was in industries where you have true believers who are in the government agencies and who have been doing this for a long time, they're there because they want it to happen and there's a lot of openness and excitement to working together if the company is credible and is cares as much about safety and performance as the agency does.

你在SpaceX工作期间，对于如何快速打造出能稳定运行的出色产品，有什么心得吗？

In your operating time at SpaceX, what did you learn about making great things that work quickly?

其实都是些大家都听过的经典道理。

I think it was all the classic things.

现在都有一套所谓的马斯克做事法则了

There's the Elon algorithm at this point.

但放在那个时候，这套逻辑还完全没成型呢

I don't think it was explicit back then.

就是一句话，咱们得把这枚火箭发射出去。

It was just, hey, we have to get this rocket launched.

先确保它能正常工作。

Let's make sure it works.

别过度优化，也别陷入分析瘫痪，先把目标定清楚，想出个靠谱的方案就行。

Let's not over optimize or have analysis paralysis, just define what the goal is, come up with a good solution.

这个方案看着不错。

This looks like a good solution.

这计划挺好。

That's a good plan.

那就干吧。

Let's run with it.

别为点小事拖上好几个月反复讨论。

Let's not deliberate for months and months and months over things.

干脆点，定了就干。

Let's just decide and move forward.

就算我们搞错了，以后也能回头调整，因为整个过程中有大量的测试。

And if we're wrong, we can always go back because there's extensive testing along the way.

所以关键是用好工程原则，从第一性原理出发，快速推进，先做到95%或99%的完成度，投入运行，之后再慢慢优化。

So it's really use good engineering principles, think from first principles, move fast, get it to 95%, 99%, get operational and you can make it better later.

但如果我们一直不发射，也不投入运行，那这事儿肯定成不了。

But if we just never launch, if we never get operational, this isn't gonna work.

现在让我觉得特别有意思的一点是‘在做中学’，而且现在已经有人仔细研究过这个现象：正是在一次次迭代中，我们才能积累大量经验，而且这些学习过程其实是有规律可循的。

One of the things that's most interesting to me today is this whole learning by doing thing, which has now been carefully studied that it's in the iterations that you gather lots of your learnings and there's like a literal predictable curve to these things.

而且我们把这么多环节都外包出去了，特别是在制造业方面，都外包到了海外。

And that we've outsourced so much of this, especially in manufacturing to overseas.

所以问题不在于我们负责创新、别人负责制造，而是制造过程本身就能催生大量的创新。

And so it's not that we do innovation and someone else does manufacturing, it's more that the manufacturing creates a lot of the innovation.

我挺好奇，你觉得在未来一二十年里，这对美国来说有多重要。

And I'm curious how important you think that is in the next decade or two here in The US.

显然，你正在做的这个项目之前被运到了俄罗斯，现在我们又想把它弄回来。

Obviously, you're doing something that got shipped over to Russia and now we're trying to bring it back.

我想请你再聊聊，把这类产业迁回我们本土的整体部署，你是怎么看的？

Curious for you to say a little bit more about the broader effort to bring more of that stuff back onto our own shores.

好的。

Yeah.

产业回流确实是一方面，但或许这个问题本质更简单，就是垂直整合而已。

Think there's the onshoring piece, but it's maybe even simpler as just vertical integration.

哪怕只看美国国内，核能、航空航天还有国防等很多领域，现在都默认要把业务外包出去。你把一个子系统外包给别家，对方又会把某个零部件再外包出去，这个零部件还涉及不同的上游原料，于是一层层把业务一直外包到底。

So even domestically, have companies in the nuclear space, the aerospace industry, like defense, many different sectors where subcontracting is the norm and you're gonna subcontract a subsystem to somebody who subcontracts a component and then that component has different inputs and they subcontract that all the way down.

我觉得在航天领域，比如SpaceX早期，光一个系统就涉及三十层分包商，这情况跟航天飞机时代差不多。

I think in aerospace at the SpaceX days, it was 30 layers of subcontractors in this one system somehow, I think was the case from the space shuttle.

分包商的数量多到离谱。

Some crazy number of subcontractors.

最近核能领域也有类似情况，有一次分析师电话会上，一家公司还炫耀自己有将近900家分包商，多到他们甚至需要设立区域协调员来管理这些分包商。

And then in the nuclear space recently, there was one analyst call where a company was bragging about having something like 900 subcontractors and so many that they needed regional organizers of the subcontractors.

一方面，这说明你做的事情确实非常复杂，行业壁垒也很深；但另一方面，每个跨公司的接口通常都是固定的，不会轻易变动，所以你得把它们当作固定项来处理。

And on one hand, it's like, okay, clearly what you're doing is really complicated then and there's deep barriers to entry but every one of those interfaces that crosses another company is typically a fixed interface that's not gonna move very quickly and so you have to treat it as fixed.

所以每个人都根据预先设定的接口要求来设计自己的部分，最终整个系统架构变得僵化不堪。

And so everyone's designing their individual piece against preconceived interface requirements and you end up with a really calcified architecture of your system.

这还只是系统设计层面的问题。

This is just at the system design level.

你没法跨层级为整体目标做优化。

And you can't optimize across layers for the overall goal.

但如果你把这些工作都收回来，不再依赖大量分包商，作为工程团队，你就能在每次迭代中灵活调整接口需求。

Now if you bring all that in house and don't have a lot of subcontractors, you can actually, as an engineering team, optimize with every iteration, trade off interface requirements.

你们电气团队之前提的那个需求，对我们机械团队来说实在太难了。

Hey, this thing that you asked of me over in the electrical team, for me, the mechanical team is gonna be really, really hard.

能不能给我一点余地，稍微放宽一点要求？

Do you mind giving me a little bit of breathing room on that?

我可以在别的地方把这部分补回来。

I'm gonna make it up somewhere else.

两个人坐在同一张桌子旁，沟通起来要容易得多，不像两家公司相隔千里那样。

And so you can have those conversations much more easily when two people are sitting side by side at a desk than when you have two separate companies across the country.

所以至少，你得把工程团队整合起来，垂直整合，自己多承担一些工程工作，别总往外包。

And so at a minimum, think you need to pull the engineering together and vertically integrate and just sign up for doing more of the engineering yourself and not outsource it.

接下来第二点是制造环节。

And then part two is the manufacturing.

这还只是设计层面的问题。

That's just on the design side.

如果你真要生产这个产品，而且想设计出便于制造的方案，那最好让制造和工程团队在同一个地方。

If you're actually gonna make the thing and you're trying to design for manufacturing, well then maybe you want your manufacturing actually co located with engineering.

现在所有优秀的硬件公司都是这么做的：至少把首版产品的制造环节和工程团队安排在同一个屋檐下。

And so that's what all the great hardware companies are now doing is you at least have your first of a kind manufacturing co located with engineering under the same roof.

你们可以直接沟通，说你刚设计的这个东西，为什么让我生产起来这么难。

You basically talk it out of here's why this thing that you just designed is hard for me to make.

如果你能这么改一下，我就不用六轴数控机床了。

If you can make it this way, I don't need a six axis CNC.

我用激光切割机就能做出这个零件。

I can use this with a laser cutter and make this part.

这样能把效率提高十倍，成本却只有原来的十分之一。

We're gonna get 10x throughput at one tenth the cost.

所以我觉得，至少要把首件试制环节和工程团队放在同一个地方。更好的做法是，把小批量、中批量的生产也留在这儿，等规模大了再搬到其他地方，比如成本更低的地区，而不是像南加州那样，这么多公司都扎堆在那儿。

And so I think you need to at a minimum put your first of a kind manufacturing co located with engineering and then better yet, small scale manufacturing, mid scale manufacturing, maybe a large scale, you can push it somewhere else and go to a lower cost center as opposed to like Southern California where so many of these companies are.

但这一步是必须的。

But I think it's mandatory.

你必须把这些环节全都整合在一起。

You've got to bring it all together.

关于能源在文明或社会中所扮演的角色，你有哪些心得？

What have you learned about the role that energy plays in a civilization or a society?

挺有意思的，其实很长一段时间里我们都没怎么聊过能源这个话题。

It's funny, like we didn't really talk about energy for a long time.

我觉得在AI、数据中心这些热潮爆发之前，人均能源增长其实停滞了好几十年。

I think the per capita growth of energy was kind of flat for a couple of decades prior to this recent surge because of AI and data centers and everything.

你对于能源整体作用的思考框架是什么？我们先聊聊这个，然后再深入到你具体在做的事情。

What is your framework for thinking about the role of energy in general and then we'll get into the specifics of what you're doing?

关于这个话题，我有两个特别喜欢的图表。

There's two of my favorite charts around this topic.

一个是人均GDP对比人均能源消耗。

One is GDP per capita versus energy consumption per capita.

它的R平方值肯定超过0.8。

The r squared on it is certainly over point eight.

把全球所有国家的数据点都标出来，画出人均GDP和人均能源消耗，你会发现它们之间有一条非常明显的趋势线。

You cluster every country on earth, you plot GDP per capita, you plot energy consumption per capita and there's a very obvious line through them.

虽然有些例外，但这个关联性实在太强了。

There's outliers but it's so predictive.

这就是关键所在。

It is the thing.

能源的使用和生产，本质上就是人类繁荣、经济活动这些方面的终极指标。

Energy use and production is the ultimate proxy for human prosperity, for economic activity, for all these things.

这一点其实早就被大家认识到了，但一直被严重忽视。

This has been, I think, pretty well understood for a while but completely under discussed.

然后你看美国，这是我常看的另一张图：我们曾经长期扩建电网，但从九十年代起，直到今天，几乎就没怎么增长了。

And then you look at The US, this is the other chart I always look at is we grew our grid for a long time and then starting around the nineties, essentially no growth until today.

中国却一直在快速增长。

China was growing a long time.

到2010年，我们俩还旗鼓相当。

2010, we were equal neck and neck.

我觉得今年他们的总能源产量可能会达到我们的三倍。

I think this year they now will be triple us on total energy production.

所以，如果美国还想在经济上保持影响力，就必须做点什么了。

And so The US just needs to do something if it wants to continue to be relevant economically.

要想推动经济增长，就得增加能源生产。

You need more energy production if you wanna grow your economy.

我把制造业全往外一推，等于主动放弃了能力，尤其是当涉及到新型制造技术、如何快速扩大产能这些至关重要的东西时。

And I think just purely outsourcing stuff overseas just gives up capability, especially as it gets to like new manufacturing techniques, how to scale manufacturing rapidly, all these things that are so important.

你觉得因果关系是怎样的？

How do you think the causality runs?

有一种可能性是，如果中国的人均能源产量超过我们，比如说甚至达到我们的两倍，

One idea could be if China surpasses us on a per capita basis and let's say they get to double.

会不会存在某种客观规律，使得他们一定能找到方法把这些能源用起来，推动人均GDP增长，进而带动各方面的发展？

Is there some rule of the way things work that will just mean they'll find ways to use that energy in some way that raises GDP per capita and drives all this progress?

那有没有可能因果关系是反过来的？比如GDP先增长，或者我们先实现新的技术突破，然后能源产能才跟上，去满足这些新增的需求？

Or is the causality the other direction as GDP rises or we find new innovations like energy catches up to rise to meet that demand or something?

这事挺值得琢磨的：如果我随便给一个国家空投——我也不确定具体怎么操作——相当于美国人均三倍的能源产能，这个国家就一定能发展得富裕起来吗？

It's curious to imagine if I just air dropped, I don't know, three times The US per capita capacity of energy on some random country, would that country necessarily become prosperous?

从九十年代一直到2020年，大家争论的焦点差不多是：就算经济增长了，你真的想要那样吗？

The whole argument from the nineties almost till 2020 was something like, even if it would increase the economy, would you want to?

所以呢，我是该在国内建几十家铝冶炼厂，还是说把这些低利润、没意思、对经济加速没多大帮助的事外包出去更好？既然这样，干嘛还要让人干这个，让他们去服务业做点更有趣的事不行吗？

So okay, do I wanna have dozens of aluminum smelting facilities in my country or is that better done overseas and it's low margin and not interesting and not gonna accelerate your economy and why do you have people working on that when they can be in the services world and do more interesting things?

这确实是个很有力的论点。

Certainly, that was an argument.

但我现在看到的是，如果你不主动出击，没有快速扩张的能力，当能源需求突然大增时，你就会措手不及，完全被动。

But I think we're seeing it right now what happens when you're not proactive about it, when you don't have the capability even of expanding rapidly and you don't wanna be caught completely flat footed when there's a large demand for energy production.

现在AI和数据中心就是这样，看看这些趋势图，如果任由这种增长继续下去，到2030年数据中心的耗电量可能占满整个电网。

And now here we have it with AI and data centers and you look at some of these curves and looks like data centers could consume if they were allowed to the entire grid by 2030 at this growth rate.

理论上，你可能会说，我们需要多少能源，就会有多少能源建起来。

And so I think in theory, you would say energy production will come online as we need it.

但现实中，人们不会主动去建设，反而会偏向其他领域，因为我们的能源成本本来就高，供应又紧张，而其他国家产能过剩、补贴充足，他们趁机掌控了产业链前端，慢慢主导了经济的很多方面。

But then in practice, you don't develop it, you may bias yourself towards other things because, well, our energy costs are kinda high or our energy cost is higher than other countries because we're tighter on supply and they've overbuilt and they're subsidizing, but then they take those industries and then they take the front of the supply chain and just start marching forward to where they can dictate a lot of different things about the economy.

过去人们普遍认为，经济是高效的，市场也是高效的。

That was the classic view is economies are efficient, markets are efficient.

只要有能源需求，我们自然会在需要的时候把它建起来。

If there's a demand for energy, we will bring it online when we need to.

但现实中的基础设施建设，都有它自己的时间表。

But physical world stuff has a timeline with it.

在美国，想获批做点事情往往特别难。

And then in The US, it can be very hard to get stuff permitted to go do things.

所以一旦出现突发的、迅猛增长的需求，你就可能措手不及——现在美国的数据中心就是如此，到处都在抢电。

And so when there's unexpected demand that's very rapid in its increase, you might be caught completely off guard, which I think is the situation in The US now with data centers and them scrambling to find power.

几年前，如果你看数据中心选址，会想它们到底会建在哪儿？

And where a few years ago, if you looked at a data center and and you thought, okay, where are they gonna place it?

答案通常是建在有闲置电力的地方。

The answer was they'll put the data center where there's stranded power.

比如西德克萨斯的闲置风电，或者南达科他州、北达科他州的闲置伴生气。

Stranded wind in West Texas, stranded flare gas in South Dakota, North Dakota.

现在这些闲置资源都用完了，得建新产能，但数据中心连燃气轮机都订不到，核电还得几年才能装好。

And now those stranded assets are gone and now it's time to build new capacity and data centers are completely backed up on getting natural gas turbines and nuclear takes a few years to get installed.

所以我觉得，从能源的稳定性和韧性角度来看，你需要具备相应的能力，至少要有按需快速部署的能力。

So I think just from a energy stability, robustness point of view, you wanna have capabilities and at least the ability to stand those up as you need to.

显然，你正在采取措施解决这个问题。

Obviously, you're doing something about it.

不过，你对目前的整体状况有什么看法？

But how do you feel about the state of things overall?

我是说美国整体的能源状况，包括现状和发展方向。

The state of energy in The United States overall, Both state and direction.

以前情况还行，但现在发展方向完全停滞了。

The state's been fine until now and then the direction's completely flat.

对我来说，问题在于几十年来，供应端一点都没增加。

So to me, the issue is that we have not increased at all in decades on the supply side.

发展方向真不怎么样。

The direction don't feel great.

我真希望我们能有更多能源。

The state, I wish we had more.

我真希望我们别只占中国三分之一的水平。

I wish we were not one third of China.

如果能源不仅充足，还便宜，我们本可以做更多事情。

I think we could be doing a lot more things if energy was not just abundant but cheap.

光靠增加高成本能源，根本没法让那些离开的产业回来，也不会让我们重新做起以前的事。

And so just making more high cost energy also won't bring back certain industries, it won't cause us to start doing things we used to do.

你觉得能源是制约我们把那些外流产业拉回来的瓶颈吗？

You think energy is the bottleneck to us bringing back some of these industries that have left?

还是说问题更多出在劳动力上？

Or is it more labor?

我觉得劳动力反而是最容易灵活调整的因素。

I think the labor is actually the one that's probably the most responsive.

如果某个领域的工作岗位真的很有吸引力，大家自然会转行去做这行。

If there's really compelling jobs in a field, people will shift over to that.

最典型的例子就是这几年的电工岗位。

The classic one, last couple years has been electricians.

供应量远远跟不上需求。

Massively undersupplied.

供应量缺口实在太大了，连建所有数据中心的量都凑不出来。

Massively undersupplied, can't get enough to build all the data centers.

这些岗位的薪资水平特别高，赚的甚至比不少上过大学的人还多，对的。

They're commanding very high great wages, can make more than people who went and got a Yeah.

这绝对是份很不错的职业，现在已经有很多人转行进入这个领域了。

It's like an incredible career and now get a lot of people shifting into it.

我觉得劳动力的调整是会响应的，可能需要几年时间，但比建设基础设施要快得多。

I think labor responds, it probably takes a few years, but it's faster than building infrastructure.

基础设施可能是最大的瓶颈。

Infrastructure is probably the bottleneck.

那问题出在能源上吗？

And then is it just energy?

应该不只是能源问题，审批流程也可能拖很久。

Probably not just energy, probably permitting can take a while.

所以我觉得，让项目上线其实面临好多不同的障碍。

And so I think there's a bunch of different bottlenecks to getting things online.

你为什么相信核能？

Why do you believe in nuclear?

为什么偏偏是核能？

Why that specifically?

从数据上看，核能一直都是最安全、最清洁的基载能源。

If you look statistically, it's always been the safest, cleanest baseload.

所以我觉得对经济来说，真正需要的是基荷电力。

So I think for the economy, what you really want is baseload.

基荷电力就是一直供电的意思吗？

Baseload meaning always on?

就是持续稳定地供电，不是时有时无的，而是能作为经济基础，用来规划工业流程的可靠电力。

Always on, very reliable supply of electricity, not intermittent, but something you can actually design an industrial process around as the foundation of the economy.

就连数据中心也需要基荷电力。

Even data centers want baseload.

太阳能配上足够多的储能理论上也能当基荷，但通常储能根本达不到那个量。

And so solar with enough storage could be baseload, but typically it's not anywhere near that amount of storage.

太阳能一般还是用来削峰填谷的，波动性比较大。

Usually solar's more intermittent, more for peak shaving.

所以如果你真想要基荷电力，想要企业跟工业能依赖的稳定电源，那它就得几乎全天候运行。

So if you really want baseload, if you want something that businesses and industry can rely on, it has to be on almost all the time.

它必须非常可靠。

You need it to be highly reliable.

所以核能是最可靠的。

And so nuclear is the most reliable.

这样你就满足了基荷供电的要求。

So you check the baseload box.

过去这十到二十年，人们才开始像现在这样关注气候问题。

People only in the last ten, twenty years have started caring about climate even remotely as much as they do now.

如果你关心气候，我想你大概也在意碳排放。

And if you care about climate, I think maybe you care about carbon.

你很可能也关心颗粒物污染。

You probably also care about particulate.

在这些方面，核能的清洁程度远超其他能源。

Nuclear is the cleanest by far on these dimensions.

所以从环保角度看，核能是基荷供电的最佳选择。

So on the environmental side, nuclear wins for baseload.

至于安全性，这正是人们最常质疑的地方。

And then the safety side, this is the part that people think they question.

在安全性方面，它也是所有基荷能源里最安全的。

The safety side, it's the safest by far of baseload too.

如果是化石燃料发电厂，大家都知道它排放的废气会实实在在损害人们的健康。

You have fossil fuel plant that has known impact on human well-being from emissions.

核电本身不会产生任何碳排放，虽然它确实出过几起备受关注的事故。

And nuclear has none of the carbon emissions and there have been high profile accidents with nuclear.

大家首先想到的就是切尔诺贝利、三里岛和福岛这几起事件，但即便把这些广为人知的事故算进去，核电的实际风险也远远低于任何其他基荷能源。

You think of Chernobyl, 3 Mile Island, Fukushima, but the actual risk associated with nuclear despite those high profile incidents is still far far lower than any other form of baseload energy.

所以我会说，要是你问我哪种基荷能源最安全、最清洁？

And so it's like, okay, if you ask me what's the safest, cleanest form of baseload?

绝对是核电。

Nuclear, absolutely.

我为什么支持核电呢？

Why do I believe in nuclear?

这些因素在我看来都非常关键。

I think those things are important.

好的。

Okay.

那成本呢？

What about cost?

成本上它怎么样？

How does it stack up on cost?

这里有个关键点：从七十年代开始，我们逐渐减少核电建设，越往后越少。

And this is the piece where it gets really exciting to where over time we did less and less and less nuclear probably starting in the seventies.

随着核电建设减少，它的成本反而越来越高。

And you see the cost of nuclear going up as we do less of it.

看看现在核电的成本，已经比化石能源还高了。

You look at where the cost of nuclear is today and it's more than fossil fuels.

除非你特别在意安全性——虽然各种能源的安全性都还说得过去——或者特别在意碳排放，否则真值得吗？

It's like, okay, unless you really care about safety, which is pretty acceptable from all forms, or you really care about carbon emissions, is it really worth it?

建一座反应堆要花十年、十五年，你真愿意这么做吗？

Are you really gonna do nuclear when it can take ten years, fifteen years to build a reactor?

而且最终的花费还可能会超出预算一倍。

And they can be double the expected budget.

这种情况对任何电力公司来说都太难接受了，谁还会说，对，我还要接着建核电项目呢。

That's really hard for a utility to stand up to and say, yeah, I wanna do more of this.

他们得为普通用电用户把成本降下来。

They need to bring costs down for ratepayers.

他们得明确规划好新增发电产能的上线时间，不能含糊。

They need to have predictability in when they're gonna bring capacity online.

从这些角度来看，核电的表现一直不太理想。

On those dimensions, nuclear has not been the best.

它的新增发电产能响应速度不够快，成本也一直不是最低的。

It's not highly responsive new energy generation then it hasn't been the lowest cost.

那可能有人会问，既然这样，我们为什么还要关注核能呢？

So you say, okay, why care about nuclear?

这是因为从基本原理来看，核能本应该是成本最低的能源之一。

It's because on first principles, should be one of the lowest cost.

现在的物理技术强大多了，反应堆可能小很多，但能输出大量能量。

You've got much more powerful physics, much potentially smaller reactors outputting a lot of energy.

你看一小粒核燃料，它所相当于的煤炭量，差不多有一吨那么多。

You look at a pellet of nuclear fuel and the amount of coal that that's equivalent to is like a ton of coal.

所以我能把这些东西、占这么大地方的燃料，压缩到小得多的空间里。

And so I can take all this stuff, all this space, and pack it in so much smaller.

我的反应堆不需要那么多材料。

I don't need as much material for my reactor.

我也不用消耗那么多 mined 的原料。

I'm not gonna be going through as much mined product.

这完全是数量级上的差别。

It's just an order of magnitude different.

所以从基本原理上讲，核能应该非常便宜，甚至可能比化石燃料便宜得多。

And so on first principles, nuclear should be an extremely affordable thing, potentially much more affordable than fossil fuels.

现在确实不是这样，但这才是我们追求的目标。

Now it has not been at all, but that's the goal.

所以现在，有一波先进反应堆正在努力从反应堆层面实现这一点，而我们则想在燃料层面也做到。

And so now you see this whole wave of advanced reactors trying to make that true on the reactor side and then we're trying to do it on the fuel side.

你能对这些先进反应堆做个分类吗？

Can you create a taxonomy of the advanced reactors?

比如，现在都有哪些不同的技术路线在尝试？

Like what kinds of approaches are being tried?

大多数人一想到核能，脑子里浮现的都是那种巨大的核电站和高高的冷却塔，画面特别有冲击力。

Think everyone thinks of nuclear as the massive kind of plants and the stacks that they're used to seeing pictures of and are very evocative.

这种形象已经持续很久了。

And it's sort of been that same way for a long time.

建造一座这样的电站，成本动不动就上百亿美元。

They cost $10,000,000,000 to make or something.

那现在都有哪些新的尝试方向？

What are the things being tried?

咱们一个一个来分析。

Take that one piece at a time.

所以在反应堆这边，大家想到的都是冷却塔。

So on the reactor side, everyone thinks of the cooling towers.

其他工业流程里也有冷却塔，但大家一提到冷却塔，脑子里浮现的就是蒸汽和热水冷却降温的场景，这样水就能在循环开始时重新作为冷却剂使用。

You have cooling towers in other industrial processes too but everyone thinks of the cooling tower which all it is, is steam cooling off, hot water cooling off so that it can be reused as a coolant at the start of the cycle.

很多工业流程都会这么做，但核能行业尤其以这个著称。

So lots of industrial processes do that, but nuclear is known for it.

这通常指的是吉瓦级别的反应堆。

That's typically about a gigawatt scale reactor.

在美国常见的是西屋公司的AP1000反应堆，装机容量是十亿瓦级别的。我们后面可以再聊技术细节，比如它用什么冷却剂、什么燃料，这些内容可能挺有意思的。

Often in The US, an AP 1,000 Westinghouse gigawatt scale reactor design, which we can get into the technical of what type of coolant they use and what type of fuel, which might be interesting.

但在我看来，大家首先要记住的是，这种反应堆属于大型堆，规模是十亿瓦级别的。

But the thing to think about in my mind is that's a big reactor, that's a gigawatt scale.

我觉得按照规模来划分是比较有意思的分类方式，因为规模和它们的应用场景、市场是直接挂钩的

I think the interesting buckets are around the size because they link to applications and markets.

首先就是电网供电这类用途

And that's grid.

所以如果要给电网供电，就需要吉瓦级的规模，这样才能真正给电网带来实质影响，而且只有达到这个规模，才能把成本压到最低。至少从目前来看，这种规模的反应堆是合适的选择。

So you wanna go battle it out on the grid, you need gigawatt scale because it moves the needle on the grid and you're getting the scale that can bring your cost down as low as possible, at least for now, then that's a good format.

而在另一个极端，还有微反应堆，主要服务于那些孤立的、目前靠柴油发电机供电的社区。

On the other end of the spectrum, you have microreactors where you're saying there's a stranded community that's doing diesel generators.

这种做法既不环保，成本也算不上低。

This is both bad for the environment and not that cheap.

我们的成本能比它们更低，所以就可以选择微型反应堆。

We can beat them on cost, and so you go microreactor.

现在我们来说兆瓦级的，不是吉瓦级，是兆瓦级。

Think one megawatt scale, not gigawatt, megawatt.

然后还有中间档的，就是总装机量100到300兆瓦的小型模块化反应堆，也就是SMR。

And then you have the middle, which is a 100 to 300 total megawatts SMR, where SMR is small modular reactor.

这三类规模的反应堆采用的技术路线各不相同，但我认为它们未来都会非常重要。

And all three buckets have different approaches on the technology side, but I think all three are gonna be important.

如果说吉瓦级反应堆是为电网服务，兆瓦级反应堆供给偏远社区或政府设施，那中间规模的反应堆，我觉得会在未来五到十年里，在数据中心的户内供电领域找到自己的发展空间。

And so if gigawatt scale is for the grid and megawatt scale is for remote communities or government installations, then the middle scale I think is gonna find its niche behind the meter with data centers over the next five to ten years.

一个机组可以专门供应一个数据中心。

Where one unit could be dedicated to one data center.

就像我作为当地用户支付的电费不会因为附近有数据中心而上涨，整个系统是完全独立的。

Like something that doesn't interrupt the cost that I'm paying as a ratepayer in whatever district I'm in, even if there's a data center there, like something that's more cordoned off.

对，就是这样。

Yeah, exactly.

所以你可以把这种模式看作是完全脱离电网的孤岛。

So think of it as like the ultimate behind the meter thing is just an island.

想象一下，围起几百英亩地，里面建个数据中心，旁边就是它的核电站，或者其他形式的发电设施——比如峰值电厂、天然气涡轮机，甚至在某些情况下还有太阳能。

Imagine a fence around a few 100 acres, you've got a data center and you've got its nuclear power plant there or whatever other form that they might have, peaker plants, natural gas, turbines, maybe even solar in some cases.

明白了。

Okay.

所有设备都完全隔离。

That's all cordoned off.

它根本不接入公共电网。

It does not even touch the grid.

它不会对电网造成影响。

It's not impacting the grid.

它也不会给社区带来任何负担，比如电费成本，完全是分开的。

It's not doing anything to the communities like rate cost completely separate.

我觉得这其实是个巨大的机会，能借此改善社区。

I actually think there's a huge opportunity to improve communities through this.

如果一个数据中心投入数十亿资金，既用于计算也用于电力生产，假设这两方面各占一半，那我们能不能把电力生产部分再提升10%？这其实只相当于总项目成本增加了5%。

If you have a data center putting billions into both the compute and the power production, let's say they're even fifty fifty on power production and compute, can we increase that power production 10%, which is only a 5% project increase?

对于一个吉瓦级的数据中心来说，给社区额外增加100兆瓦的电力，这影响可不小。

And in the case of a gigawatt data center, bring an extra 100 megawatts to a community?

这确实是个大动作。

That's huge.

这会让那个社区的公用事业费用大幅下降。

That would plummet utility prices for that community.

所以我认为，最少见的情况是人们会采用‘表后’模式，或者至少不完全脱离电网；而最理想的状态是，他们实际上在向电网供电，甚至提供基荷电力。

So I think at a minimum, we're gonna see people going behind the meter or not disconnected from the grid and ideally we see them actually feeding the grid and feeding base load.

这让我想起了你那个BYOE的概念。

That reminds me of your BYOE concept.

也许你可以再详细解释一下。

Maybe just spell that out a little bit more detail.

我们几十年来第一次，对电网进行了如此大规模的投资。

There was just the concept of we're doing all this investment in the grid for the first time in decades.

这些投资都是由私营企业承担的。

Private companies are making the investment.

他们会自带电源、自备能源，这就是BYOE。

They will make the investment with bringing their own power, bring their own energy, BYOE.

而且，不管怎样，他们本来就要做这些投资。

And they're going to be doing that investment anyway.

他们最担心的可能是：我能在社区里建这个数据中心吗？

Their biggest worry is probably something like, will I be allowed to build this data center in this community?

过去一年，我觉得这已经开始成为讨论的焦点。

Over the last year, I think that's been the start of a discussion.

如果企业想把这件事敲定，还能跟当地社区说：‘我们会把项目建在这块闲置的土地上。’

And if they want to get to the right answer on that and they can tell a community, hey, we're gonna be over here on this unused land.

你们是否接受我们在这里运营、提供算力，而且我们还能给当地电网带来净收益？

Are you okay with us being here and running a bunch of compute and being net positive to your grid.

那我们能不能反过来，给社区提供额外电力，换取入驻当地的许可呢？

Can we just plug in and give you some extra power in return for inviting us to be a neighbor?

在我看来，这对所有相关方来说都是稳赚不赔的事。虽然会小幅增加成本，但现在那些建设数据中心的超大规模企业，核心诉求就是尽快接通电力、上线运营——毕竟大家都在比拼谁能先把规模做起来。

To me that's a complete no brainer for everybody at a slight cost increase during a time when hyperscalers, building data centers are all about speed to power and they wanna be online as fast as possible because this is the contest to who can get to the greatest scale first.

跟能顺利进驻一个社区、快速落地项目，而不是被拒之门外的优势比起来，这笔投入的成本非常低。

That is a very low cost compared to the advantage of deploying quickly in a community and being invited there versus being rebuffed.

所以这绝对是个不用多想的好选择，而且现在的新型数据中心完全能实现全闭环冷却。

So I think total no brainer and the modern data centers can do completely closed loop cooling.

所以水资源问题根本算不上问题，真正卡脖子的就是电力这块。

So the water issue is not an issue, it's really just this power thing.

那如果这些新型先进反应堆最后没能投入使用，原因可能会是什么呢？

What would be the explanation for these new advanced reactors ultimately not working?

那有什么事会让你担心得睡不着觉呢？

What keeps you up at night?

毕竟这种情况本来就不是你能掌控的。

Because this is kind of out of your control.

你本人不是这些新型反应堆的设计者，你是在协助他们完成目标。

You're not designing these advanced reactors yourself, you're helping them achieve their mission.

但他们的项目必须要能成功才行。

But they need to be successful.

如果要你分析这些新型先进反应堆有可能行不通的最主要原因，你觉得会是什么？

What do you think if you had to handicap the reasons why they might not be, are the most likely?

我觉得排在第一位的原因就是没有足够的燃料来维持运行。

I'd say number one is they don't have fuel to operate.

那这个项目就彻底停摆了。

That's a showstopper.

所以我们现在要全力排除这个风险。

So we're trying to eliminate that risk.

第二个问题是成本太高了。

And then part two is they're too expensive.

我觉得物理原理层面是没问题的。

I think the physics will work.

我觉得没人会太担心这个问题

I don't think anyone's too worried about that.

现在有几十家实力出众的公司，里面的工程师都很优秀，都在努力攻克先进反应堆的相关难题

You have dozens of incredible companies with great engineers who are trying to solve the advanced reactor problem.

一般来说，这些团队会选定一个规模合适的技术方案，瞄准特定的市场——要么是供偏远地区使用的微反应堆，要么是给数据中心用的小型模块化反应堆（SMR），还有就是给电网供电的千兆瓦级大型反应堆。

And, typically, they'll pick a form factor that is of the right size to focus on a certain market, whether that's microreactor for remote or SMRs for data centers or very large gigawatt scale for the grid.

而在我看来，小型模块化反应堆最终的发展目标，是要比电网用的千兆瓦级电站项目建造成本更低。

And SMRs, I think, ultimately have the plan of being cheaper than the gigawatt scale construction projects for the grid.

现在也有很多研发新型反应堆的团队在努力解决反应堆的成本问题。

But there's many reactors that are trying to solve the reactors cost side of it.

但核心问题还是，他们最终能把发电成本降到足够低的水平吗？

And the question is gonna be, is their energy production cost low enough?

但对于这些先进反应堆来说，问题不仅仅是燃料供应，更是成本问题。

But for these advanced reactors, it's not just a fuel availability problem and question, it's a cost question.

燃料成本可能占到总成本的一半，而且我认为长期来看，铀浓缩费用会占到燃料成本的一半。

It's the fuel can be half of their total cost, and I think long term enrichment will be half of the fuel cost.

你的财务团队并没有因为重大失误而亏钱。

Your finance team isn't losing money on big mistakes.

钱是通过成千上万个没人注意的小决策慢慢流失的。

It's leaking through a thousand tiny decisions nobody's watching.

Ramp能在你消费前就帮你把支出管控好

Ramp puts guardrails on spending before it happens.

实时限额、自动规则，完全不用费心救火

Real time limits, automatic rules, zero firefighting.

欢迎访问ramp.com/invest了解详情

Try it at ramp.com/invest.

随着业务不断发展，Vanta也能同步升级，它可以自动完成合规审核，还能为你整合所有安全与风险相关的核心信息

As your business grows, Vanta scales with you, automating compliance and giving you a single source of truth for security and risk.

可以访问vanta.com/invest了解更多信息。

Learn more at vanta.com/invest.

Ridgeline正在重新定义资产管理技术，它真正以合作伙伴的身份提供服务，而不仅仅是一个软件供应商。

Ridgeline is redefining asset management technology as a true partner, not just a software vendor.

它们已经帮助很多公司的规模扩张到原来的五倍，实现了更快的增长、更高效的运营，还帮企业建立了竞争优势。

They've helped firms 5x in scale, enabling faster growth, smarter operations, and a competitive edge.

你可以访问ridgelineapps.com，看看它们能为你的公司带来什么发展可能。

Visit ridgelineapps.com to see what they can unlock for your firm.

每家投资公司的情况都不一样，通用化的人工智能根本摸不清你这边的业务流程

Every investment firm is unique, and generic AI doesn't understand your process.

但Rogo就可以

Rogo does.

这是一个专门为华尔街打造的人工智能平台，能对接你的各类数据、贴合你的业务流程，还能产出实际可用的工作成果。

It's an AI platform built specifically for Wall Street, connected to your data, understanding your process, and producing real outputs.

可以访问rogo.ai/invest了解更多。

Check them out at rogo.ai/invest.

从OpenAI、Cursor到Perplexity，头部AI和科技公司都在借助WorkOS，只用几天就能满足企业级服务的各项要求，不用拖上好几个月。

The best AI and software companies from open AI to cursor to perplexity use WorkOS to become enterprise ready overnight, not in months.

访问workos.com，把那些枯燥乏味的基础设施搭建工作甩在一边，专心打磨你的产品就好。

Visit workos.com to skip the unglamorous infrastructure work and focus on your product.

今天专门聊聊General Matter这家公司，我很想听听你从三个维度讲讲你是怎么规划它的发展的

For General Matter specifically, I'd love to hear at three levels how you're thinking about building this thing.

分别是产品层面、业务层面，还有公司自身的整体层面

The product level, the business level, and then the company itself.

那我们先从产品说起，你们是怎么摸清市场到底想要什么、需要什么的呢？

So starting with the product, how have you figured out what the market wants or needs?

你们的业务是浓缩铀。

You're enriching uranium.

铀是这套能源分类体系里，用来发电的核心燃料资源。

Uranium is the key fuel source for this different parts of this taxonomy for creating energy.

你能给客户提供的产品和服务其实有很多不同的种类。

There's lots of different kinds of things that you could provide to your customer.

在确定最终产品到底是什么的过程中，你们经历了怎样的摸索？

What has the journey been like to figure out what the actual product is gonna be?

对，这得追溯到对核燃料循环或核燃料供应链的理解。

Yeah, so this goes all the way back to understanding the fuel cycle or the fuel supply chain for nuclear.

整个过程分五个步骤。

So there's five steps.

第一步是从地下开采铀矿。

You mine uranium out of the ground.

之后要把它转换成气态

You then convert it into a gas.

再进行浓缩提纯

You then enrich it.

之后再把它变回固体形态。

You then deconvert it back into a solid.

而浓缩这个步骤本质上就是提炼和分离的过程。

And the enrichment step is really a refining step, a separation step.

然后把它重新变成固体，再压成燃料颗粒。

Then you turn it back into a solid, and then you form fuel pellets.

从地下的矿石到小小的燃料颗粒，再到装进反应堆的燃料棒，一共要走五步。

Five steps to go from ore on the ground into a little pellet into a fuel rod that goes into a reactor.

美国在各个环节都有不错的能力，唯独在浓缩环节不行。

The US has good capability across the board except in enrichment.

这是唯一一个我们完全没有商业化生产能力的环节。

It's the one area that we don't have any commercial scale capability.

虽然有一些研发能力，但远达不到商业规模，也无法跟俄罗斯和欧洲竞争。

There's some r and d capability, but nothing that's commercial scale or commercially competitive with Russia and Europe.

所以我们决定聚焦在浓缩这个环节，因为这是关键瓶颈。

And so we decided let's hone in on the enrichment step, that's the bottleneck.

正是这个瓶颈导致了我们所说的‘核燃料断崖’，目前有三个断崖。

That's the thing that's leading to these what we call the nuclear fuel cliffs where there's three cliffs.

第一个断崖是高丰度低浓缩铀的供应链。

Cliff 1 is the HALEU supply chain.

这个问题我们在所有小型模块化反应堆公司那里都听到了。

This is the same problem we heard from all the SMR companies.

我们没有高丰度低浓缩铀的来源。

We don't have a source of HALEU.

高丰度低浓缩铀是他们需要的燃料。

HALEU is the fuel that they need.

高丰度低浓缩铀是浓缩到大约20%的铀，也就是铀-235，这是裂变材料。

HALEU is uranium that's enriched to about 20%, u two thirty five, which is the fissile material.

这种原料需要把铀浓缩度提高到20%，这样才能把反应堆造得更小。

They want it enriched a little bit higher to 20% because it helps the reactor be smaller.

所以第一个瓶颈就是先进反应堆所需的HALEU原料供应问题。

So that's the first cliff is HALEU for advanced reactors.

目前完全没有相关产能。

They have none.

如果未来几年都拿不到稳定的货源，他们的产能扩张计划根本无从谈起。

If there's no reliable supply in the next few years, it's going to make it very difficult for them to scale up.

到目前为止，他们只从能源部拿到了一小部分，用来证明反应堆能按预期运行，以及进行首批部署，但还不足以规模化到真正对电网产生重大影响。

So far, they have small amounts from the DOE to prove that their reactors work the way they think they will and to do first deployments, but not to scale to where it really moves the needle on the grid.

第二个核能悬崖，也就是核燃料悬崖，出现在2028年，届时美国对俄罗斯铀产品的禁令将全面生效。

The second nuclear cliff, the nuclear fuel cliff is 2028 when The US's ban on Russian uranium ports goes into full effect.

到了2028年1月1日，根据国会通过的一项法案，我们将不再被允许进口俄罗斯铀。

01/01/2028, we are, through an act of congress, no longer allowed to import Russian uranium.

是指禁止浓缩铀，还是铀产品都包括？

Enrich uranium specifically or or or both?

我们不会从俄罗斯进口浓缩铀，也不会从俄罗斯进口任何铀原料。现在我们能获取浓缩铀服务的地方只有欧洲和俄罗斯，

We will not be importing enriched uranium from Russia or uranium from Russia, which the only place we get it enriched is Europe and Russia.

而当前美国进口的铀里，有约25%都来自俄罗斯。

And today, Russia is about 25 of US imports.

而且到2028年，现有的25%储备量会直接耗尽

And so 25% immediately goes away in 2028.

到那时美国的电力公司就只能消耗自身的库存，还要想方设法和欧洲谈合作，从欧洲进口更多的铀

And so utilities in The US will then start eating into their inventories and trying to work out deals to import more from Europe.

但第一个危机是高 Assay 低浓缩铀（HALEU）的供应缺口，缺口占总需求的20%。

But cliff one is HALEU, 20%.

第二个危机是低浓缩铀（LEU），也就是浓缩度在3%到5%之间的铀的供应危机。

Cliff two is LEU, low enriched uranium at 3% to 5% enrichment.

而再过一段时间，美国海军推进系统用的浓缩铀储备也会彻底耗尽。

And then at some point in the future, The US stockpile of enriched uranium for propulsion for the navy and eventually that runs out.

所以这些危急的关口会接连出现。

And so there's these cliffs.

我们主要关注核能这一块，也就是低浓缩铀和高浓缩铀，这些就是我们要生产的产品。

We're focused on the nuclear energy side of it, LEU and HALEU, that's what we're gonna be producing.

我们意识到，眼下最紧迫的问题就是高浓缩铀，特别是针对先进反应堆的需求，必须尽快解决。

We realized, hey, the most urgent cliff that we need to address as soon as possible is HALEU for advanced reactors.

巧合的是，这属于一个小众市场。

And coincidentally, this is the small market.

这是一个新兴的市场，我们可以去开拓，而且我们认为现有的大企业不会在同一时间段内跟进。这样我们就能服务好那些我们认识多年的先进反应堆客户，确保他们拿到所需的高浓缩铀，从而顺利部署并扩大反应堆规模。

This is the emergent market that we can go after that we don't think any incumbent will go after on the same time frame as us, and we can serve those advanced reactor customers that we've known for years and make sure that they have the halo that they need to deploy and scale their reactors.

我们就是从这里开始的。

That's where we're starting.

接下来第二阶段是投入低浓缩铀的生产，为当前电网上的94座反应堆提供所需燃料，目前在美国这个市场价值约202.5亿美元。

And then phase two is go into LEU production, low enriched uranium for the 94 reactors on the grid today and supply them with the fuel that they need, which today in The US is a $2.02 and a half billion dollar market.

这种浓缩工艺和用于武器的浓缩方式之间有什么关系？

What's the relationship between this kind of enrichment and a more weapons oriented type of enrichment?

大部分情况下，所有的浓缩工艺其实都是一样的。

All enrichment is for the most part the same.

你可以把它看作是一种精炼或者蒸馏过程，只是不断提纯而已。

If you think of it again as like a refining process or a distillation process, you're just refining it further.

我们最终提供的是浓缩服务。

Ultimately, the product that we make is enrichment services.

你可以把从地下开采出来的天然铀送入浓缩流程，通过处理得到任意浓度的浓缩产物。

You can run material through from natural uranium as it comes out of the ground and you can run it through an enrichment process to get to any arbitrary level of enrichment.

在商业上主要关注的是低浓缩铀，浓度3%到5%，用于现有反应堆；或者高浓缩铀，浓度20%，用于先进反应堆。

The commercially relevant ones are LEU, three to 5% for existing reactors or HALEU, 20% for advanced reactors.

所以我们一开始做这件事的时候就说，现在市场上最急需的是高富集铀。

And so when we set out to do this, we said, hey, HALEU is what the market needs most right now.

而且我们启动项目的时候，还没发生俄罗斯禁令，所以当时并不存在低富集铀的断供危机。

And we set out pre Russia ban, so there was no LAU cliff.

我们决定开发富集能力，来满足这个市场需求。

We said we're gonna develop enrichment capability to serve that market.

富集能力的基本单位，其实就是完成这种提纯过程的能力。

And the fundamental unit of enrichment capability is it's just the ability to do this refining.

在富集领域，这个能力用一个叫‘分离功单位’的指标来衡量。

It's measured as something called Separative Work Units in the enrichment field.

我们会提供这项服务，卖给核电运营商和先进反应堆，帮他们获得所需的燃料。

We will provide that service and we will sell that service to utilities and advanced reactors to get them the fuel that they need.

所以最终，我们开发的产品就是富集服务，它可以应用于这两种富集水平。

So ultimately, the product that we developed is enrichment, and you can apply it to either of those levels.

至于你问的，这和武器级富集有什么不同，当一些国家走向武器级时，通常是为了突破20%这个国际公认的分界线。

To your question on how is that different from weapons grade, when you see countries going to weapons grade, they're often trying to go north of 20%, which is the internationally agreed limit to where you go to weapons grade.

出于防扩散的考虑，国际社会有强烈的共识，认为我们都应该保持在20%以下。

For nonproliferation reasons, there's strong international consensus that we should all just stay below 20.

说实话，风险远大于收益，没必要让人走得更远。

There's really, the risk is not worth the reward to let people go higher.

我看到有些国家声称要把浓缩度提高到60%，说是为了发展核能产业，但这理由挺可疑的。

You see countries going to 60% saying that it's for their nuclear energy industry and it's pretty suspect.

我觉得他们明显是想尽可能接近武器级水平，也就是超过90%的浓缩度，以便研发核武器。

I think clearly they're just trying to get as close as they can to weapons grade, which is well over 90% and they're trying to develop weapons.

所以归根结底，技术本身并没有区别，只是应用方式不同，而且是在完全不符合国际共识的环境下进行的。

And so ultimately, it's not different technology, but it's applied in a different way and in a completely non international consensus setup.

这个百分比是不是可以理解为纯度之类的概念？

Is the percentage there a way to think about that like purity or something?

对。

Yeah.

铀-238和铀-235是它们从地下开采出来时的自然状态。

There's u two thirty eight, there's u two thirty five as it comes out of the ground.

还有其他同位素，但主要的是铀-238，我们需要把它过滤掉，才能得到铀-235。

There's other isotopes also, but the main one's u two thirty eight and you basically wanna filter that out to get u two thirty five.

铀-235才是容易发生反应的同位素。

U two thirty five is the the isotope that wants to react.

这才是可裂变的同位素。

That's the fissile isotope.

我们对这个了解多少

What do we know

全球的铀矿储量情况如何？

about the stock of ore in the world?

就是原材料。

The raw element.

我们有足够的量吗？

Do we have enough of it?

是的。

Yeah.

储量很充足。

Plenty.

我觉得地下的铀资源至少够我们用好几百年的。

I think we've got certainly hundreds of years of supply in the ground.

美国那边也有供应。

US has supply.

美国现在还在进行铀矿开采。

We have active mining in The US.

加拿大的铀矿储量更丰富，矿石品位也更高，足够供应我们很长时间，澳大利亚也是这种情况。

Canada has even more supply, higher grade ores, enough to supply us for a very long time, same as Australia.

然后哈萨克斯坦也有巨大的铀矿储量，美国也从那里进口。

And then Kazakhstan's a huge deposit and US imports from there as well.

所以在矿石供应方面，没什么问题。

So on the ore side, not an issue.

这部分会逐步推进。

That'll progress.

效率还会进一步提升。

That'll get more efficient.

现在有了新的技术。

There's new techniques there.

采矿业肯定会不断进步。

So mining's gonna make progress.

我觉得采矿成本会下降。

I think the cost of mining will come down.

产量会提高。

Output will go up.

但这根本不是问题。

But it's not really an issue.

你怎么基于这个产品打造一家优秀的企业？

How do you build a great business on top of this product?

是因为这个新领域有巨大的潜在需求，所以做起来很容易吗？还是有其他因素要考虑？

Is it so easy just because there's so much latent demand for it from these this new segment or are there other considerations?

我觉得完全谈不上容易。而且我认为，这种潜在需求在很长一段时间里并不明显，直到最近两三年才显现出来。

I wouldn't say it's so easy at all And I think the latent demand has not been obvious for a long time until the last couple of years.

所以，我认为那庞大的潜在需求，是在过去二十四个月里，因为大家觉得会有更多数据中心从事AI推理或训练工作才出现的。

So any large latent demand I think is in the last twenty four months linked to the thought that there's gonna be a lot more data centers doing AI inference or training jobs.

但在那之前，其实很少听到有人讨论新建核电站的事。

But if you go before that, there wasn't a lot of new nuclear being talked about.

确实有对小型模块化反应堆（SMR）的需求，但那个市场其实挺小的。

There was need for HALO, but that's a pretty small market.

所以我认为很多人觉得，其实我们并不需要新增产能。

And so I think a lot of people have thought that this is not really something where we need new capacity.

你回顾一下2010年代，那十年里几乎没有任何地方在扩建新产能。

If you go back to the twenty tens, really no new capacity building anywhere in the twenty tens.

特别是低浓缩铀的生产，当然还有高浓缩铀，毕竟这些领域最近才刚刚兴起。

Certainly not for LEU production, certainly not for HALEU just because it's been emerging so recently.

所以我觉得，既然这是一个增长中的市场，我们在很多方面依然看好它。

So I think the fact that this is a growing market is we're still betting on that in many ways.

目前我们并没有在美国大规模部署新反应堆，但我们认为这即将到来。

That's not the case that we're currently deploying tons of new reactors in The US, but we think that's coming.

所以，我们深信这个市场，但我不确定整个行业是否都认同需求确实存在。

So I think the market we believe in deeply, I don't know that the entire industry believes that the demand is there.

他们经历过之前的核能复兴，但最后没能成行，所以他们的观点是：等确定有需求了再建。

They've been through nuclear renaissance before that didn't pan out and I think their point of view would be we'll build it when we know that there's a need.

至少这是我们与许多人交流时，他们表现出的态度。

At least that's been the attitude of many people that we've talked to.

至于说这事儿做起来难不难，其实技术上是成熟的。

On the ease of doing this, it's a proven thing.

以前有人做过，但依然不容易。

People have done this before, but it's still not easy.

我可以说，你看有些以前用过的老技术，后来因为太久没碰，很多能力都丢了。

I'd say you see some of these retro technologies that were done in the past and then we lost a lot of the capability because we didn't do it for so long.

所以铀浓缩这块，行业里从九十年代以来就没怎么变过，没什么进展，也没怎么利用新技术。

And so for enrichment, it's been an industry without a lot of change since really the nineties And not a lot of progress, not a lot of leveraging of new technology.

我们需要攻克大量艰难的工程难题，才能重新追上现代技术的发展节奏。

There's a lot of doing the hard engineering work to get back up to speed to modern.

而且就算这些工作都完成了，你还要建上百万平方英尺的设施，这类基建项目规模极大，要投入数十亿美元。

And then even once that's done, you're building facilities that are a million square feet and large large multi billion dollar infrastructure projects.

就好比特斯拉第一代Roadster造出来就已经很难了，但真正的挑战是怎么把产量规模化？

In the same way that like maybe building the first Tesla Roadster was hard, but the real challenge was how do you scale this up?

要建起一座大型工业基地要付出多少心血，这一点大家往往都严重低估了。

I think it's extremely underappreciated how much goes into standing up a huge industrial facility.

我觉得这恰恰是最棘手的环节。

I think that's actually the hard part.

那你会怎么设定这个业务的核心指标呢？

How do you think about the North Star metric for the business?

打个比方，就像SpaceX把每公斤载荷送入轨道的成本那样，把这个成本变化放在时间轴上看，真的很有意思。

Like if I equate it to SpaceX, the cost per kilogram to orbit or something like that is like a really cool thing to visualize over time.

那你这边对应的变化趋势图会是什么样的？

What's your chart gonna look like?

那这个核心指标是什么呢？

What's the metric?

对的。

Yeah.

没错，他们用每公斤入轨成本作为指标，而且特指的是近地轨道的成本。

So they had cost per kilo to orbit, really to a specific orbit of low Earth orbit.

而且实际上你还能把轨道成本拆解到发射所需的速度增量上。

And you can actually decompose the orbital piece into a velocity.

那差不多可以类比成把一公斤载荷送上某一速度对应的单位成本。

So it's almost like cost per kilo to some velocity.

放到我们的业务里，这个指标就是把一公斤铀原料提纯到指定丰度的单位成本。

Our version is cost per kilo of uranium, that of a payload, to some enrichment level.

所以我们的指标和那个类似，就是每公斤的成本降低到3%到5%，或是19%，这个成本也可以用单位分离功的成本来描述。分离功是核浓缩行业的通用计量单位，本质上就是用材料的重量乘以熵减程度，也就是材料的分离或规整度。

And so it's like cost per kilogram to three to 5% or to 19%, which can be described as cost per separative work unit, where separative work unit is the industry measure of enrichment to where it's basically kilos times some entropy reduction or separation or organization of the material.

分离功单位通常写作千克分离功单位，也就是重量乘以分离功。

And so a separative work unit is typically referred to as kg SWU, kilos times separative work units.

我们的核心指标是每千克分离功单位的美元成本。

Our North Star metric is dollars per kilo SWU.

这个指标和客户的价值创造周期对应得上吗？

Does that map onto the value creation cycle of your customer?

这个投入成本是不是决定他们业务成败的关键？

Is that input cost the key determinant of their success as a business?

确实是。

It is.

重要性取决于你用的是哪种反应堆。

The importance depends on what type of reactor you're running.

如果你用的是十亿瓦级的传统AP1000压水堆，用的是3%到5%低浓缩铀，那么燃料成本在总成本里的占比其实很低。

So if you're doing a gigawatt scale classic AP 1,000 light water reactor that uses 3% to 5% enriched fuel, low enriched uranium, LEU, then the cost of that fuel as a percent of your overall cost is quite low.

这个影响不大。

It doesn't really matter.

你真正需要的是运行的稳定性。

What you want is availability.

你需要的是可靠性。

You want reliability.

你需要供应商的多样性。

You want diversity of suppliers.

你得确保他们能持续供货，而且不想看到市场过度集中，从而给你带来供应风险。

You wanna know that they're gonna be there, and you don't want them overly concentrating and creating supply risk for you.

但燃料成本在你整体发电成本，也就是平准化度电成本（LCOE）里占比并不大。

But the cost of the fuel is not a huge input into your cost of electricity, your LCOE.

大部分成本其实是资本支出，毕竟你们的项目规模高达一百亿美元。

Most of that's gonna be CapEx for your huge $10,000,000,000 project.

所以对那些大型项目来说，这影响不大，但对先进反应堆而言，这就非常关键了。

So So for them, it doesn't matter that much, but for advanced reactors, matters a lot.

有些先进反应堆使用的HALEU燃料，浓缩度高达19.75%，提炼过程需要消耗大量原料。随着过滤层层递进，这种燃料的成本甚至能占到其发电总成本的一半以上。

So some advanced reactors, the cost of the fuel, the HALEU fuel that goes into it enriched further to 19.75%, which requires more input material to refine it all the way there as you filter through more and more and more material, that fuel can be more than half of their energy production cost.

因此，对他们来说，这确实是个大问题。

And so to them, it does matter a lot.

目前业内采购这类燃料的流程，刚好和我们的核心指标完全对应，这套流程包含五项不同的服务。

And the way that fuel has been purchased in the industry so far, which maps directly onto our North Star metric is five different services.

电力公司会先采购铀原料，之后在整个供应链流转的过程中，再陆续为铀的升级处理付费。

A utility purchases uranium, and then it purchases upgrades to the uranium as it goes through the supply chain.

所以这个行业本质上就是代加工模式：对方持有核材料的相关许可，整个流程全程只走账面流转，每完成一个环节的工作，他们就给对应服务商按服务单元付费。

So it's really a tolling business where you buy they own license to the material and it's book transfers all the way through, and they are paying different people per unit of service provided to do their work.

他们给我们提供的服务是按“套”来美元计价的。如果梳理一下核燃料全流程的实际成本结构，对于低浓缩铀来说，浓缩环节的成本占比可能不到三分之一，但就算这样，它也是成本占比最高的几个板块之一。

The service that they provide to us is priced in dollars per suite and if you look at the actual cost structure of producing fuel all the way through, for low enriched maybe enrichment's a little bit less than a third but it's one of the largest cost segments.

对于高-assay低浓缩铀（HALEU）来说，这个占比还会更高，我们认为它会成为燃料成本里最主要的影响因素。

For HALEU, it's even more and we think it's gonna be the dominant cost driver of fuel cost.

回到我们最开始聊的创始人基金（Founders Fund）的那套逻辑，你相当于精准找到了整个供应链里成本最高的环节，还把它卖给了那些新兴的、规模尚小的全新客户——这类客户的成本占比极高，你完全可以拿下这个市场。

Tying back to the Founders Fund stuff we were talking about at the beginning, you sort of found the highest cost segment of the workflow sold to the highest percentage of the cost, new customer that's emerging, that's small that you can go own.

这就是你最终能把生意做大做强的逻辑：把所有资源和动力都集中在最核心的那个瓶颈上，全力打通它。

And that's how you then ultimately build a great business is by driving all of this force through kind of the narrowest possible choke point.

如果要用投资术语或者商业战略的说法来总结，刚才说的就是核心道理。

If you wanna put it in investment terms and like business strategy terms, that would be it.

但要是回到最开始的情况来看，最后会发现这个环节其实是最不可或缺的。

But I think if you go back to the very beginning, it just turned out that this was the thing that was most necessary.

正是浓缩铀这个环节出了问题，我们才一直生产不出先进反应堆所需的高丰度燃料。而且你会发现，美国明明在八十年代是遥遥领先全球的第一，现在却已经丢掉了这项能力。

This step of enrichment was why we didn't have the more enriched fuel that advanced reactors needed, and you realize The US had lost the ability despite being the number one in the world during the eighties by far.

在这件事上我们曾经做得非常出色。

It was something we were extremely good at.

后来我们就彻底停止了这项业务。

We completely stopped doing it.

这个瓶颈会成为整个核能行业的瓶颈。

This bottleneck was gonna be the bottleneck to all of nuclear energy.

如果你看好先进核能，这就是我们必须解决的问题。

If you believed in advanced nuclear energy, this is the thing we had to solve.

所以我们做了，因为没人做，而且这事极其重要、非常紧迫。

So we did it because no one was doing it and it was extremely important and urgent.

我觉得这恰好为我们提供了一个非常好的市场切入点。

And I think that ended up corresponding to this very good entry point for the market.

但我觉得这两件事其实是紧密关联的。

But I think the two are really linked.

我觉得这绝非巧合——当你主动去解决一个没人碰的紧急难题，切入的又是一个规模不大但处于上升期的市场，这个切入点本身就完美契合行业的发展逻辑。

I don't think it was coincidence that if you're solving a problem that's urgent that no one else is doing on a small but emerging market, it will completely fit that framework.

那最后一步就是打造一家能够长久发展的优秀公司，我说的可不只是维持现金流运转这么简单，而是要打造一支优秀的团队，还能为国家创造实实在在的价值。

And then the final step is building a great enduring company, by which I mean not just a flow of cash flow, but a collection of people and a set of impacts on the country.

你觉得从现在开始，要做成一家优秀的公司，最该优先做的事是什么？另外，对你来说，做成这样一家公司到底意味着什么？

How do you think about the most important things you can do starting now to build a great company and like what does that mean to you?

你投资过很多优秀的公司，除了好产品和稳定的现金流，一家伟大的公司还具备什么特质？

You've invested in many great companies, what is a great company on top of a good product and a stream of cash flow?

归根结底还是团队。

It comes back to the team.

Naval 说过很多次，你组建的团队就是你打造的公司。

I think Naval has said this a bunch of times, the team you build is the company you build.

所以对我们来说，团队基因才是关键。

And so to us, it's team DNA.

我们得考虑把公司设在哪里。

We want people we thought about where do we put the company?

这在一开始是个大问题。

This was a big question at the very beginning.

你得看看团队里到底需要哪些人。当然，你需要核工程师，但同时也需要优秀的机械、电气、软件、化学等各个领域的工程师。

And you look at who you actually need on the team and of course you need nuclear engineers but you also need great mechanical engineers, electrical, software, chemical, every type of engineering.

所以你得问自己，公司到底该设在哪？

And so you ask yourself, well, where do we wanna put the company?

结果发现，团队里核工程师的比例是个位数。

And it turns out the percent of the team that's nuclear engineers is single digits.

我们并不涉及反应堆。

We're not doing a reactor.

实际上，我们的流程中根本没有任何核反应。

There's actually no nuclear reactions in our process.

我们必须确保过程中不会出现核反应。

We need to make sure that there's no nuclear reactions.

但要做到这一点，其实不需要很多核工程师。

But to do that, you don't need a huge number of nuclear engineers.

所以当时就面临一个问题：是去核工程师集中的地方，还是去其他类型工程师多的地方？

And so there was this question of, do you go to where the nuclear engineers are or do you go to where all those other types of engineers are?

从核工程师这个维度来看，我们确实不需要太多。

And on dimension one, the nuclear engineers, you don't need that many of.

但必须是顶尖的。

You need great ones though.

我们上哪儿去找这样的人？

Where are we gonna find them?

你看看全国范围内，他们分布得零零散散，哪儿都有。

And you look at where they are in the country and they're just scattered everywhere.

根本找不到一个集中的地方。

There's no one place to go.

所以我们只能选另一个方案。

And so we had to go the other option.

那就去其他工程师聚集的地方，特别是硬件和航空航天领域，那就是南加州。

You go to where all the other engineers are, specifically hardware, aerospace and that's Southern California.

我们想打造的团队基因是，这是一家由工程驱动的公司。

The team DNA that we wanted to set up was this is an engineering driven company.

我们不是在做科学实验。

We're not doing a science project.

不涉及任何新物理理论。

No new physics.

不能走那种充满不确定性的、长达数年的研发路线。

No going down a multi year R and D path that's uncertain.

我们必须尽快让项目投入运营。

We need to get this operational as fast as possible.

这本质上是个工程问题。

This is an engineering problem.

我们要做的所有事情，都得按工程思路来考量。

Everything we have to do has to be thought of as engineering.

我们正通过工程手段把成本降下来。

We're engineering cost out of the system.

我们正通过工程手段把性能提上去。

We're engineering performance up.

我们正通过工程手段降低资金成本。

We're engineering cost of capital down.

我们正通过工程手段把进度安排得尽可能紧凑。

We're engineering schedule to be as tight as possible.

就连我们的办公楼都是自己用工程化思路打造的。

We're even engineering our own buildings.

这套思路其实是从特斯拉的成功经验里借鉴来的。

This is taken from this Tesla playbook.

千万别找总承包商把所有事都外包出去，那样你会彻底失去控制权。

Don't hire a GC and outsource everything, you lose total control.

项目进度就完全不在你自己掌控里了。

The schedule could be it's not in your hands anymore.

你得组建一支包含内部工程师、采购和施工团队的EPC公司，来负责你的建设项目。因为最难的地方不在于单纯的工程设计或制造，而是要在预算内、按计划完成数百万平方英尺的施工。

You need to build your own in house engineer, procure, construct EPC firm as a team to go run your construction projects because it's one of the hardest things about this is not just engineering, not just manufacturing, but actually construction of millions of square feet on schedule, on budget.

所以，公司的整个DNA都围绕着尽快为行业交付产品这一目标，同时确保安全、可靠，满足其他所有你期望的维度。

And so the whole DNA of the company is oriented around drive to deliver for the industry as quickly as we can while being safe, reliable, every other dimension you would want.

在这些约束条件下，快速行动，因为行业需要它。不要陷入分析瘫痪，不要在无关紧要的事情上反复纠结。

Subject to those constraints, go quickly because the industry needs it, don't have analysis paralysis, don't deliberate over things that don't matter.

让我们先上线一项能为美国行业提供服务的产品，然后再逐步降低成本。

Let's get a service live that can deliver for The US industry and then let's bring the cost down over time.

所以所有事情都围绕进度和成本展开。

And so everything's oriented around schedule and cost.

你是怎么管理公司的？

How do you run the company?

说真的，你的时间是你刚才描述的那些事情能否发生的关键投入之一。

Like literally your time as one of the key inputs into everything you just described happening or not.

你观察过很多垂直整合的公司，它们从头到尾掌控整个流程，并且做得非常出色。

You've watched a lot of vertically integrated companies that own a process end to end do that really really well.