2026年重大构想：物理人工智能与工业栈

这一转变伴随着真正的风险。

This shift carries genuine risks.

同样那些可以检测野火或预防工地事故的工具，也可能引发反乌托邦式的噩梦。

The same tools that can detect wildfires or prevent job site accidents could actually enable dystopian nightmares as well.

软件影响物理世界的方式，是通过这些具身化的通电组件。

The way that software will affect the physical world is through these sort of embodied electrified components.

我们看到初创者试图将这些问题分解为一组可模块化的部分，以便将流水线的原理应用于社会规模的问题。

We're seeing founders try to reduce these problems into kind of a decomposable set of modular parts such that you can apply the principles of an assembly line to society scale problems.

数据混乱的问题并非新问题，它正是这一更广泛运动的核心。

The problem of messy data is not a new one, and it's at the heart of this broader movement.

下一波的赢家将是那些真正赢得公众信任的人，他们构建了保护隐私、可互操作的原生AI系统，使社会更加透明，同时不减少自由。

The winners in this next wave will be those that really earn public trust, building privacy preserving, interoperable AI native systems that make society both more legible without making it less free.

定义下一年建设的关键是什么？

What will define the next year of building?

我们的2026年构想反映了我们的投资团队认为将塑造技术未来演进的主题。

Our twenty twenty six ideas reflect the themes our investing teams believe will shape how technology evolves next.

本集围绕四个核心理念展开，探讨人工智能如何走出屏幕，进入实体经济。

This episode is built around four big ideas about AI leaving the screen and entering the physical economy.

当人工智能进入工厂、建筑工地、供应链和关键基础设施时，规则将发生变化。

When AI moves into factories, construction sites, supply chains, and critical infrastructure, the rules change.

可靠性至关重要，现实世界的约束会迅速显现，优势将从能够构建系统而非仅开发软件的团队身上体现。

Reliability matters, real world constraints show up fast, and the advantage shifts from teams that can build systems, not just software.

你将听到三种视角，探讨推动这一转变的因素：以工厂为先的思维、电工业栈、物理可观测性以及工业数据前沿。

You're going to hear three perspectives on what enables that shift: a factory first mindset, an electro industrial stack, physical observability, and the industrial data frontier.

首先，我们需要一个运营模式。

To start, we need the operating model.

亚伦·普赖斯·赖特认为，我们正进入美国工厂的复兴时期，这不仅指实体建筑，更是一种理念体系。

Aaron Price Wright argues that we're entering a renaissance of the American factory, not just as a building, but as a set of principles.

其理念是将流水线逻辑应用于能源、采矿、建筑和制造等领域，通过模块化、自动化和熟练劳动力，将复杂工作转化为可重复的系统。

The idea is to apply assembly line logic to problems like energy, mining, construction, and manufacturing, using modularity, autonomy, and skilled labor to turn complex work into repeatable systems.

以下是埃琳的分享。

Here's Erin.

工厂。

Factory.

我认为明年，企业将采用以工厂为先的思维来应对从能源、采矿到建筑和制造的各种挑战。

I think next year we'll see companies approach challenges from energy to mining to construction to manufacturing with a factory first mindset.

通过将人工智能和自主技术与熟练劳动力模块化部署，复杂的定制化流程将像流水线一样运行。

The modular deployment of AI and autonomy alongside skilled labor will make complex, bespoke processes operate like an assembly line.

美国的第一个伟大世纪建立在工业实力之上，但众所周知，我们已经失去了大量这种能力。

America's first great century was built on industrial strength, but it's no secret that we've lost a lot of that muscle.

其中一部分原因是产业外迁，以及八十年代以来everything的金融化，导致九十年代和二月大规模将工业制造转移到海外。

Some of that has been from offshoring, from the financialization of everything in the eighties leading to the large scale offshoring of industrial manufacturing in the nineties and February.

其中一些可以追溯到监管问题。

Some of it dates back to regulation.

因此，那些当初为非常良好且具体的原因而设立的规则、机构和流程，随着时间推移逐渐累积成一层厚厚的外壳，使得在美国开展新事物和建造新东西变得异常困难。

So rules and agencies and processes that were put in place usually for very good and specific reasons at the time have built up over time into a crust that makes it, you know, very hard to do new things and to build new things in America.

但我们现在就在这里，必须想办法重新培养这个国家的建设文化。

But here we are, and we have to figure out how to reinstill a culture of building in this country.

我所说的工厂并不是字面意义上的工厂。

I'm not just talking about a factory in a literal sense.

比如，你有一个仓库，里面有一条装配线，由人类和机器混合操作，最终在生产线末端会产出一个产品。

Like, you have a warehouse with an assembly line where you have some mix of humans and machines, and at the end of the factory line, there's a widget that pops out.

我真正想说的是装配线的原理本身，以及这些原理是如何被应用到那些传统上不会让人联想到工厂的行业的？

I'm really thinking about the principles of an assembly line full stop, and how are those principles getting applied to industries that aren't traditionally industries you'd think of when you think of a factory?

比如住房、数据中心的建设、矿山的建设，以及大规模能源基础设施和能源项目的建设。

So housing, the construction of data centers, the construction of mines, the construction of large scale energy infrastructure and energy projects.

我们看到创业者们正试图将这些问题分解为一系列可模块化的部件，从而将装配线的原理应用于社会规模的问题。

We're seeing founders try to reduce these problems into kind of a decomposable set of modular parts such that you can apply the principles of an assembly line to society scale problems.

人工智能是一种非常出色的方式，因为它能以高度公式化和自主化的方式理解和映射监管中的各种复杂性，而无需每次都从头开始彻底 redesign 整个流程。

And AI is a really amazing way to do that because you can understand and map out different complexities in a regulation in a very formulaic and agentic way without having to completely redesign your entire processes from scratch every single time.

我们如何将技术带出工厂，应用到现实世界中？

How do we take technology and bring the factory out into the world?

今天我们正在以前所未有的速度建设数据中心，创建标准化的知识产权和标准化设计，并以创纪录的速度完成部署。

We're building data centers at an unprecedented rate today, and we're creating standard IP and standard designs and putting them up in record time.

对于我们来说，这是一个绝佳的机会，可以测试自动驾驶、人工智能、机器人以及其他正在走向成熟的技术，如何应用于这些大规模的物理资产，因为这些建设项目正在飞速推进。

It's a great opportunity for us to test where autonomy, AI, robotics, other technologies that are coming to maturity right now can be deployed on these sort of large scale physical assets because these building projects are moving so fast.

随着数据中心市场的发展，这些技术将衍生出来，并广泛应用于各类工业项目，无论是新建高速公路、机场和跑道，还是矿山及采矿和精炼设施的建设，这些都亟需被建设。

As the data center market develops, these technologies spin out and become useful across a broad cross section of industrial projects, whether that's the construction of new freeways and airports and landing strips or the construction of mines and mining and refining facilities, are so desperately needed.

我们如何将数据中心建设中获得的快速推进经验，应用到建造新的工厂、晶圆厂和制造设施上，以生产各类商品，无论是用于国防、消费还是美国商业领域？

How do we take some of the learnings about how quickly we're able to move in data centers and apply them to building new factories, new fabs, new facilities to manufacture goods, whether it's for the defense sector, for the consumer sector, or the commercial sector in The United States?

我们如何实现大规模建造？

How do we build things at scale?

我们如何建立工业产能，并将规模化能力转化为优势？

How do we create industrial capacity and use our ability to scale as an advantage?

如果你是一位创业者或建设者，对重新定义美国工厂的建造方式充满热情，请来找我们聊聊。

If you're a founder or a builder and you were excited about reinventing what it means to build a factory in The United States, come talk to us.

亚伦明确了目标：通过将工厂原理应用于工业问题，实现更快、更可靠且规模化地建造。

Aaron framed the goal: Build faster, more reputably, and at scale by applying factory principles to industrial problems.

现在，让我们转向使这一切在机器内部成为可能的技术。

Now we make the move to what makes that possible inside the machines themselves.

布莱恩·麦金托什阐述了电工业栈的崛起，即为电动汽车、无人机、数据中心和现代制造业提供动力的电气化实体组件。

Brian McIntosh lays out the rise of the electroindustrial stack, the electrified embodied components that power EVs, drones, data centers, and modern manufacturing.

他还强调了一个关键观点：难点不仅仅在于技术本身。

He also makes the key point that the hard part is not just technology.

而在于构建能够生产、供应和规模化这一技术的生态系统。

It's building the ecosystem that can produce, supply, and scale it.

以下是瑞安。

Here's Ryan.

我叫瑞安·麦金托什。

My name is Ryan McIntosh.

我是美国动力团队的投资合伙人。

I'm an investing partner on the American Dynamis and team.

我对2026年的主要观点是：电工业栈将改变世界。

My big idea for 2026 is that the electro industrial stack will move the world.

下一次工业革命不仅会在工厂中发生，更会发生在驱动工厂的机器内部。

The next industrial evolution won't just happen in factories, but inside the machines that power them.

这就是电工业栈的崛起。

This is the rise of the electro industrial stack.

这是一种结合技术，为电动汽车、无人机、数据中心和现代制造业提供动力。

Combined tech that powers electric vehicles, drones, data centers, and all of modern manufacturing.

我认为人们常提到一些常见的刻板印象。

I think there are common tropes people report on.

人们说中国已经遥遥领先，我们追不上了。

People talk about China's so far ahead, we can't catch up.

但实际上，几年前人们还说中国远远落后，而美国速度极快。

And actually, you know, you go back a couple years ago and people were saying, you know, China's very far behind and America's incredibly fast.

所以我们看到了一种剧烈的反转，现在情况正好相反。

So we've seen sort of like a whiplash, now it's the opposite.

我认为现实是，中国拥有的技术，美国也能做到。

I think the reality is that, you know, the technology that China has, America can do.

我们在工程方面非常出色。

We're very good at engineering.

我们非常擅长做特定的事情。

We're very good at doing specific things.

事实上，就连最近关于稀土的问题，比如稀土的分离和加工，我们也懂。

And in fact, even like the recent stuff around rare earths, for example, rare earth separation and processing.

我们知道怎么做。

We know how to do this.

我们能做到。

We can do this.

我们可以非常快地做到。

We can do it incredibly fast.

真正的挑战在于构建一个能够大规模、低成本工业化生产这种能力的生态系统。

The real challenge is building the ecosystem to do this industrially at scale and doing it at a low cost.

另一个例子，人们通常会提到像SpaceX或Andoril这样的公司，这些大型企业需要极速运转，因此必须垂直整合。

Another example, you know, people typically talk about is companies like SpaceX or Andoril, these large businesses that need to move incredibly fast and thus vertically integrate.

在很多方面，它们是出于必要而非战略才进行垂直整合的。

In many ways, they're vertically integrating by necessity, not strategy.

根本就没有一个能够与他们一起规模化的公司生态系统。

There just isn't an ecosystem of companies that can scale with them.

但中国的情况并非如此。

That is not the case in China.

这些生态系统中存在着一级、二级、三级供应商、零部件和原材料，以及能够让它们极速运转的机构和政治实体。

There are tier one, two, three suppliers, components, raw materials that exist in those ecosystems, as well as the institutions and political bodies that allow them to move incredibly fast.

这些方面可能需要我们花费数年甚至数十年才能赶上中国。

Those are the things that might take years or decades for us to catch up to China.

我们可以做到技术层面，但其他所有方面都必须同步发展，否则我们只是在转移瓶颈。

We can do the technology, but everything else seems to grow with it or else we're just moving the bottleneck.

因此，如果你想在美国构建电气工业栈或支撑这些技术的核心组件，你需要将硅谷的软件人才与文化同工业领域的资深人士结合起来。

So if you wanna build the electro industrial stack or the core components that feed into these technologies in The United States, you need to blend Silicon Valley software talent and culture with industrial veterans.

即使是像SpaceX这样的公司，也在从曾参与航天飞机项目和各种传统承包商的人员中招募推进技术人才。

Even companies like SpaceX, they're pulling propulsion talent from people who worked on, you know, shuttle program and various old school contractors.

奎恩·肖特韦尔来自航空航天公司。

Quinn Shotwell came from Aerospace Corporation.

在这个世界里，你需要这种实际的专业知识。

There is a world where you need this actual expertise.

你需要了解以前尝试过什么。

You need to know what's been tried before.

这些其他公司里有很多聪明人，但你需要能够行动得快得多。

There are smart people out there in these other companies, but you need to be able to move a lot faster.

如今软件有很多优势，因此你需要能够吸引那些以前这些公司可能没有的软件人才。

There's a lot of advantages of software today, so you need to be able to get the software talent that may not exist in these companies previously.

你还希望将工程和制造集中在一起。

You also wanna co locate engineering and manufacturing.

像面向制造的设计这样的概念，当你在同一个物理空间或同一个生态系统中紧密整合时，可以大大加快进度。

Concepts like designed for manufacturing are something that, you know, when you're tightly integrated on the same footprint or in the same ecosystem, you can move a lot faster.

我认为你还需要为这项使命建立声望。

And I think also you need to build prestige around the mission.

对于许多传统的硅谷人才来说，最聪明的人可以投身于多个问题，而值得投入的问题也有很多。

For a lot of sort of traditional Silicon Valley talent, the smartest people can work on a number of problems and there are lot of problems that are worthy of working on.

有些人支付的报酬更高。

Some of them pay more than others.

因此，你需要为你的工作赋予某种声望或使命感，以此吸引顶尖人才。

So you need to attach sort of a prestige or a purpose to what you're working on and use that to attract the top talent.

软件影响物理世界的方式是通过这些具象的电气化组件。

The way that software will affect the physical world is through these sort of embodied electrified components.

这不仅仅是人形机器人或电动汽车，还包括电池、电力电子、计算设备和电机。

And it's not just a humanoid robot or an electric vehicle, but it's the batteries, it's power electronics, it's the compute, it's it's the motors.

所有这些我们都必须重新本土化或在制造最终产品的公司内部实现垂直整合。

All these things we're going to need to either reshor or vertically integrate within the companies who are building the end product.

这些都非常技术性。

These are, you know, very technical.

它们需要大量的专业知识。

These require a lot of expertise.

这些都是非常难以解决的问题。

These are very difficult problems to solve.

但能够解决这些问题、并拥有相应人才基础的公司和国家，将在二十一世纪胜出。

But the companies who solve it and the countries who have the talent base to order to support it are the ones who are gonna win in the twenty first century.

随着软件和人工智能变得越来越强大，并在自动化、工业和军事领域发挥更大作用，掌控这些供应链将变得愈发重要。

And as software and artificial intelligence get stronger and they start having, you know, more of a presence in automation, industrial, military, owning these supply chains is gonna become even more important.

我认为，当我们展望未来五十年、一百年时，今天对这些供应链的掌控，将深刻影响未来经济和军事权力的归属。

And I think as we, you know, look forward fifty, one hundred years, owning these supply chains today are gonna have a lot of effects of who controls both the sort of economic and military powers in the future.

瑞安指出，物理人工智能的规模化是一个生态系统问题，而非单一突破。

Ryan shows that scaling physical AI is an ecosystem problem, not a single breakthrough.

但即使你制造出了这些机器，你仍然需要具备实时感知和理解现实世界的能力。

But even if you build the machines, you still need the ability to see and understand what's happening in the real world in real time.

萨比·埃尔格拉姆提出了物理可观测性，通过摄像头、传感器和人工智能，为物理环境带来类似软件的可见性。

Sabi Elmgram introduces physical observability, bringing software style visibility to physical environments using cameras, sensors, and AI.

她解释了为何这对于安全部署自主系统至关重要，以及公众信任、隐私和互操作性不仅是附加功能，更是必要条件。

She explains why this is necessary for deploying autonomy safely and why public trust, privacy, and interoperability are not just add ons but requirements.

接下来是扎比。

Here's Zabby.

大家好，我是扎比·埃尔姆格伦。

Hey, I'm Zabby Elmgren.

我是a16z的投资合伙人，隶属于美国活力团队。

I'm an investing partner here at A16z on the American Dynamism team.

我对2026年的重大构想是：下一代可观测性将是物理层面的，而非数字层面的。

My big idea for 2026 is that the next wave of observability will be physical, not digital.

我认为，在过去十年中，软件可观测性彻底改变了我们监控数字系统的方式，通过日志和指标等手段使代码库和服务器变得透明。

I think over the last decade, software observability transformed how we monitor digital systems, making code bases and servers transparent through things like logs and metrics.

同样的革命也将发生在物理世界中。

And the same revolution is gonna come to the physical world as well.

随着美国境内部署了超过十亿个联网摄像头和传感器，我认为，物理可观测性——即实时理解城市或基础设施中发生的事情——如今已变得既紧迫又可行。

With more than a billion networked cameras and sensors deployed across The US, I think physical observability, is really understanding what happens in these cities or across infrastructure in real time, is becoming both urgent and possible now.

这一新的感知层，使得机器人和自主系统的下一阶段真正成功部署成为可能，因为你拥有了一种共同的基础设施，能够像软件中的代码一样，使物理世界变得可观察。

This new layer of perception, enables the next frontier of really, I think, robotics and autonomy also to really be successfully deployed is becoming possible because you have this common fabric that really renders the physical world as observable as code has become in software.

这一转变伴随着真正的风险。

This shift carries genuine risks.

同样这些可以检测野火或预防工地事故的工具，也可能引发反乌托邦式的噩梦。

The same tools that can detect wildfires or prevent job site accidents could actually enable dystopian nightmares as well.

在下一波浪潮中，获胜者将是那些真正赢得公众信任的人，他们构建了保护隐私、可互操作的原生AI系统，使社会更加透明，同时不减少自由。

The winners in this next wave will be those that really earn public trust, building privacy preserving, interoperable AI native systems that make society both more legible without making it less free.

而 whoever 构建了这种可信赖的基础设施，将定义物理世界未来十年的可观测性。

And whoever builds that trusted fabric will define the next decade of observability in the physical world.

当我谈到物理可观测性时，我的意思是将我们在软件领域已实现的实时可见性，应用到物理世界中。

When I talk about physical observability, I mean bringing the same kind of real time visibility we've had in software to the physical world.

在软件中，如果出现问题，你通常会在用户察觉之前，就从仪表板上看到异常。

In software, if something breaks, you usually see it on a dashboard before a user notices.

但在物理世界中，你往往是在某处冒烟、物品被盗，或机器发出绝对不该出现的声音时，才意识到问题。

But in the physical world, you tend to find out when something sparks or is already stolen or maybe a machine makes a sound that it should absolutely never make.

这至关重要，因为现在人们越来越关注如何真正思考保护和自动化关键基础设施，尤其是如此。

This matters because there's far more attention falling on how we actually think about securing and automating critical infrastructure, especially.

无论是远程矿山还是数据中心，许多场所如今都重要到不能再盲目运营了。

Whether it's remote mines or data centers, many sites are becoming honestly too important to just operate blind.

如果你是一家矿场，你正在24小时不间断运行，而人类的监管能力有限；同时，数据中心也已真正成为国家安全资产。

If you're a mine, you're running around the clock in places where humans have limited oversight, and data centers have effectively become national security assets as well.

而保护它们不仅仅是锁好服务器机房那么简单。

And securing them is not just about locking the server room.

还需要了解围绕设施周边发生的一切情况。

It's about understanding what's happening around the perimeter as well.

因此，物理可观测性本质上关乎两个方面：成功地将自主系统部署到现实世界中，并在烟雾变成火灾之前，无论是比喻意义上还是字面意义上，都能提前察觉。

And so physical observability essentially is about both successfully deploying autonomous systems into the real world and seeing the smoke both figuratively and literally before it turns to fire.

我认为现在正在发生变化的是，摄像头不再孤立运作。

I think what's changing now is that cameras are no longer working alone.

多年来，摄像头一直在记录我们周围大量的信息，但真正理解的内容却远少于它们所记录的。

For years, cameras have been recording a lot around us, but have understood honestly a lot less than what they've recorded.

这有点像一位好心的实习生，虽然做了详尽的笔记，但你仍无法判断哪些才是真正重要的信息。

It's slightly like a well meaning intern who takes great notes, but you can't really tell what really matters.

现在我们拥有了热成像传感器、射频传感器、声学传感器等，这些设备都能捕捉现实的不同维度。

Now we have thermal sensors, RF sensors, acoustic sensors, all of these things that kind of capture a different slice of reality as well.

当你将它与现代人工智能融合时，系统实际上能够理解你周围发生的事情，为你提供比单纯的图片或视频更丰富的上下文信息。

When you fuse it together with modern AI, the system actually ends up interpreting what's going on around you and gives you really more context than just a picture or a video of what's happening.

我认为像Androl这样的公司已经在国防领域证明了这一点，但令人震惊的是，许多其他行业仍远远落后。

I And think companies like Androl have actually proved this out in defense, but it's shocking how many other industries are pretty far behind.

建筑工地是当前物理感知能力严重不足的一个绝佳例子。

Construction sites are a great example of where physical observability is really lacking right now.

许多工地都位于偏远地区。

A lot of them are remote.

所有工地都十分混乱。

All of them are chaotic.

在那里获得稳定的电力都很困难，更不用说保障工地安全或真正了解日常发生了什么了。

And just getting stable power out there is hard, let alone securing a site or having a real understanding of what's happening on a day to day.

此外，还有大量可能非常昂贵的材料不断被搬运。

On top of that, there's a constant flow of likely very expensive materials being moved around.

这几乎变成了一场高价值的音乐椅游戏，物品就像游戏中椅子一样，就在你眼皮底下直接消失。

And it's basically becoming a high value game of music chairs where things really do go missing straight out from under you in the same way that a chair does in the game.

当你考虑在这些环境中部署机器人时，由于这些环境每小时都在变化，而你又处于盲飞状态，这变得极其困难。

And you think about deploying robotics into these settings as well, And it becomes incredibly difficult when you're flying blind in these different environments that are changing hourly.

大量钢材被运到新位置，临时墙壁被搭建，设备也在不断移动。

There are loads of steel that land in new spots or temporary walls that go up and equipment that shifts.

因此，我认为，确保机器人的心理模型能准确反映事物的变化，才是将自主性引入这些环境的关键。

And so I think just thinking about how you make sure that a robot's mental model is really accurate with how things are changing is how you're going to be able to bring autonomy to those settings.

我认为监控与隐私之间的张力是真实存在的。

I think the tension between surveillance and privacy is real.

我更倾向于说，在城市场景中，物理可观测性所面临的隐私问题比在采矿现场更为突出。

More true, I would say, in the city use cases for physical observability than on a mining site.

但现实是，同一套系统既能预防野火，也能保障数据中心的安全，但也可能被滥用于我们绝对不希望发生的事情。

But the reality is that the same system that can prevent a wildfire or keep a data center secure can also be misused for things that we absolutely don't want.

问题已经不再是能否构建这些系统，而是我们是否以符合民主价值观的方式来构建它们。

And the question has really become not whether we can build these systems, but whether we build them in a way that really aligns with democratic values.

我认为，这一领域的赢家将是那些将隐私和信任视为根本设计要求，而非附加功能的公司。

I think the winners in this category will really be the companies that treat privacy and trust as fundamental design requirements and not just bolt on features.

我认为在这个领域，赢得信任绝不仅仅是一个锦上添花的选项。

And I think in this space, earning trust is just like it really is not just a nice to have.

它是一种运营许可。

It's a license to operate.

我认为在物理感知领域取得成功，意味着要成为每个人依赖的感知层，无论是机器人、基础设施运营商、应急响应人员，还是工业流程，全都依赖它。

I think winning in physical observability really means becoming perception layer that everyone relies on, whether it's robots, infrastructure operators, emergency responders, industrial workflows, all of it.

关键是构建物理世界的实时地图，让其他系统能够接入，从而让这些极其复杂的环境变得易于理解。

It's about building the real time map of the physical world that other systems can plug into and make these really complex environments simple to understand.

我们在货运行业已经见过这种模式，当Samsara出现时就是这样。

We've seen this pattern before in the freight industry when Samsara showed up.

仅仅在地图上看到一个移动的点，就显得革命性了。

Just having a single dot, honestly, moving on a map seemed revolutionary.

而现在，这种微小的可见性所带来的巨大运营收益，同样适用于其他那些至关重要的行业。

And now that tiny bit of visibility that unlocked huge operational gains is also true in these other, like, very critical industries.

想象一下，如果能实现这种跨越式进步。

And imagine having that sort of step change.

但你拥有的不是地图上移动的一个点，而是对整个环境的实时多模态理解，包括资产的位置、正在发生的变化、潜在风险以及需要采取行动的地方。

But instead of a dot moving on a map, you have a live multimodal understanding of an entire environment, where assets are, what's changing, what's risky, what needs action.

whoever 构建这一层，将成为无数行业的支柱。

Whoever builds that layer becomes the backbone of countless industries.

在这些场景中，你最明显看到的是深度多模态感知功能。

The most obvious things that you'll see in these settings are deep multimodal sensing functionalities.

同样，这种信任水平至关重要——无论是政府为保护对国家安全至关重要的资产，还是公众在公共安全方面，对系统准确性的信心都极为重要。

And again, that level of trust where whether it's the government for securing an asset that's important to national security or the public when it comes to public safety, that confidence in systems being accurate really matters.

Savi 认为，物理人工智能依赖于可观测性、传感器和能够实时让现实世界变得可理解的系统。

Savi makes the case that physical AI depends on observability, sensors and systems that make the real world legible in real time.

但即使有了可观测性，决定胜负的仍有一个关键限制：数据。

But even with observability, there's one constraint that determines who wins: data.

不仅仅是干净的基准数据，而是来自真实运营的杂乱、多模态的工业数据。

Not just clean benchmark data, messy multimodal industrial data that comes from real operations.

Will Bitsky 认为，焦点正从计算转向数据限制，最具防御性的优势不仅在于清洗数据，更在于从现有设备、劳动力和工业规模的运营中直接采集数据，而这些是初创公司难以复制的。

Will Bitsky argues that the pendulum is swinging from one compute back toward data constraints, and that the most defensible advantage is not just cleaning data, but collecting it at the source, from installed bases, labor forces, and industrial scale operations that startups can't easily replicate.

这是威尔。

Here's Will.

我叫威尔·比茨基。

My name is Will Bitsky.

我是十六Z公司美国动态团队的投资合伙人。

I'm an investing partner on the American Dynamism team at a sixteen z.

我对2026年的主要观点是，关键产业将是AI数据竞赛的下一个前沿。

My big idea for 2026 is that critical industry is the next frontier for the crusade in AI data.

在2025年，数据中心、算力和能源主导了公众讨论。

In 2025, data centers, compute, and energy dominated the public discourse.

在2026年，我认为焦点将从算力转向数据瓶颈。

In 2026, I think the pendulum swings back from compute towards data constraints.

我认为关键产业是下一个前沿。

I think critical industry is the next frontier.

脏数据的问题并非新问题，它正是这一更广泛运动的核心。

The problem with messy data is not a new one, it's at the heart of this broader movement.

我们该如何处理来自不同模态的大量杂乱数据？

It's how do we take a bunch of messy data from different modalities?

这是一个大问题，每个人关注它都是正确的。

It's the big problem, and everybody is correct to be focused on it.

但我认为，这里许多根本性问题并不一定是新的，也不一定是AI独有的。

But I think a lot of the underlying problems here are not necessarily new or unique to AI itself.

一个懒惰的初步答案是，规模和数量随着时间推移往往会解决问题。

A lazy first order answer on how we get past it is that scale and quantity tend to fix things over time.

这就是我认为我们的基础设施团队更擅长应对的‘苦涩教训’问题。

This is the whole bitter lesson problem that I think our infra team is better equipped to apply here.

因此，如果你有工业规模的企业，它们拥有工业规模的数据供应，规模和数量就会有帮助，对吧？

So scale and quantity help if you have industrial scale businesses, they have industrial scale data supply, right?

这就是简短的答案。

So that's the short answer.

也许是个懒惰的答案。

Maybe the lazy answer.

你可以借鉴构建现代数据栈过程中积累的经验教训。

You can borrow on lessons learned just building out the modern data stack.

其中一些问题并非最近三年才出现。

And some of those issues are not new to the last three years.

关键是利用模型来推断结构，并让它们协同工作。

It's a matter of leveraging models to infer a structure and allowing them to work together.

遵循一致的数据本体、词汇表，并在可能的地方从源头进行数据标注，这些问题都不是新的。

It's complying with consistent data ontologies and vocabularies and labeling data at the source where you can, these are problems that aren't new.

随着工业领域现有企业越来越擅长利用现代数据栈和现有的通用模型，我们将能够解决整个混乱数据的问题。

As industrial incumbents get better at leveraging the modern data stack, at leveraging existing generalized models, we'll be able to get through the whole messy data problem.

因此，拥有宝贵数据护城河的行业主要集中在工业供应链领域。

So the industries that have valuable data moats are anything within industrial supply chain.

在制造业、国防、航空航天、商业航空、能源、采矿等领域，都有大量机会获取极其庞大的数据量。

So across manufacturing disciplines, across defense, aerospace, commercial aviation, energy, mining, etcetera, there's so many opportunities to pull extremely large quantities of data.

因此，前沿模型将能够利用这些工业领先企业提供的各种不同类型的数据。

And so the frontier models will be able to use all sorts of different data types from these industrial incumbents.

如果你最近听了World Labs团队在A16Z播客中的发言，就会知道这些问题本质上都是多模态的。

If you listen to the World Labs team on one of the A16Z podcasts recently, these are fundamentally multimodal problems.

这些模型必须协同工作。

These models are going to have to work together.

因此，对于工业领域的现有企业来说，他们可以利用现有软件平台的语言能力，结合空间输入以及任何基于传感器的数据。

So for the industrial incumbents, they can leverage language from existing software platforms, they can leverage spatial inputs, then anything sensor based.

比如本体感知，也就是来自触觉夹爪的任何反馈。

So proprioceptive, you know, any feedback from tactile grippers.

在构建这一基础设施的过程中，他们将能够提升对数据的获取能力。

As they build this infrastructure, they're going to be able to increase their access to data.

数据的供给将会增加。

The supply is going increase.

有些数据质量会更高，有些则会更低。

Some will be higher, some will be lower quality.

但这些基础模型需要一种比较优势，即以更低的边际成本获取这些数据，而这些数据将越来越多地来自这些工业领域的现有企业。

But these foundation models are going to need a comparative advantage, a lower marginal cost way to acquire this data, and they'll increasingly come from these industrial incumbents.

当我思考最难构建的层级、最具专有性、最具有防御性的部分以及价值最集中的地方时，我认为今天的答案与十二个月后我给出的答案截然不同。

As I think about the hardest layer to build and what's most proprietary, what's most defensible, where the most value is, I think the answer today is very different from the one that I'd give in, say, twelve months.

如今，所有人都在匆忙构建基础设施。

Today, everybody's scrambling to build infrastructure.

因此，他们专注于清理杂乱的数据，思考强化学习环境，以及设计数据管道。

So they're focusing on cleaning messy data, they're thinking about RL environments, they're thinking about data pipelines.

但从长远来看，我坚信数据收集——即思考数据输入在漏斗顶端的位置——才是价值积累最多的地方。

But longer term, I truly think collection, thinking about where the data inputs are at the top of the funnel, that's where the most value accrues.

我把这比作消费领域的封闭生态系统。

I analogize this to walled gardens in the consumer world.

任何拥有安装基础、现有劳动力和工业规模运营的工业公司，都能以更低的边际成本收集数据，因为他们可以从已有的运营中获取数据。

Any of these industrial companies that have installed bases, have existing labor forces, that have industrial scale operations, they have a lower marginal cost to collection because they can pull from their operations that already exist.

这些公司无法被从其现有运营中剥离出来。

It's impossible to disintermediate these companies from their existing operations.

初创公司正试图拼凑自己的数据收集系统，但为此付出了高昂的边际成本。

Startups are trying to hack together their own data collection operations, but they're paying a steep marginal cost.

他们正在建设机械臂农场。

They're building robotic arm farms.

他们销售远程操控的消费产品。

They're selling consumer products that are teleoperated.

如果他们能通过利用这些工业企业获得数据收集的替代途径，就能降低边际成本。

If they have an alternative path to collecting data by leveraging these industrial companies, then they'll be able to lower their marginal costs.

我只是不认为这些初创公司为每个收集到的数据单位支付费用是可持续或可扩展的。

I just don't see it as sustainable or scalable to see these startups paying for each marginal unit of data being collected.

这四个大想法之间有一条贯穿始终的主线。

Here's the through line across these four big ideas.

亚伦认为，扩展实体经济需要具备工厂优先的思维，将定制化的工业工作转化为可重复的系统。

Aaron argues that scaling the physical economy requires a factory first mindset, turning bespoke industrial work into repeatable systems.

瑞安解释了机电工业栈以及构建和供应现代机器所需组件所面临的生态系统挑战。

Ryan explains the electro industrial stack and the ecosystem challenge behind building and supplying the components that power modern machines.

扎比表明，物理AI只有在具备可观测性的情况下才能发挥作用，即一个可信的感知层，能够实时让真实环境变得可理解。

Zabi shows that physical AI only works with observability, a trusted sensing layer that makes real environments legible in real time.

威尔以一个限制条件作结。

Will closes with a constraint.

瓶颈正从计算转向数据，持久优势将属于那些能够大规模从实际运营中收集杂乱、多模态工业数据的公司。

The bottleneck is swinging from compute back to data, and the durable advantage will belong to the companies that can collect messy, multimodal industrial data from real operations at scale.

综合来看，这就是物理人工智能的真正含义。

Put together, this is what physical AI actually means.

不是更智能的聊天，而是可以在现实世界中部署的系统，建立在新的运营模式、新的工业基础设施和可防御的数据收集之上。

Not smarter chat, but systems you can deploy in the real world, built on new operating models, new industrial infrastructure, and defensible data collection.

感谢收听本集的 a16 z 播客。

Thanks for listening to this episode of the a 16 z podcast.

如果你喜欢这集，请务必点赞、评论、订阅、给我们打分或留下评价，并分享给你的朋友和家人。

If you like this episode, be sure to like, comment, subscribe, leave us a rating or a review, and share it with your friends and family.

如需收听更多集数，请前往 YouTube、Apple Podcasts 和 Spotify。

For more episodes, go to YouTube, Apple Podcasts, and Spotify.

在 X 上关注我们 @a16z，并订阅我们的 Substack：a16z.substack.com。

Follow us on x at a sixteen z, and subscribe to our Substack at a16z.substack.com.

再次感谢收听，我们下一期再见。

Thanks again for listening, and I'll see you in the next episode.

提醒一下，此处的内容仅用于信息参考，不应被视为法律、商业、税务或投资建议，也不应用于评估任何投资或证券，且并非针对任何a16z基金的投资者或潜在投资者。

As a reminder, the content here is for informational purposes only, should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a sixteen z fund.

请注意，a16z及其关联方可能仍持有本播客中讨论的公司的投资。

Please note that a sixteen z and its affiliates may also maintain investments in the companies discussed in this podcast.

如需更多详情，包括我们的投资链接，请访问a16z.com/discoveries。

For more details, including a link to our investments, please see a 16z.com forward slash disclosures.