特斯拉 2026 年第三季度财报电话会议：FSD 需求、Cybercab 与 Optimus

大家下午好，欢迎参加特斯拉 2026 年第三季度问答网络直播。我是投资者关系主管 Travis Aklerad，今天与会的还有 Elon Musk、Vaibhav Taneja 以及其他几位高管。我们的第三季度业绩已于中部时间下午 3 点左右，在本次网络直播同一链接发布的更新报告中公布。在本次电话会议中，我们将讨论业务前景并做出前瞻性陈述。这些评论基于我们截至今天的预测和预期；实际事件或结果可能因多种风险和不确定性而存在重大差异，包括我们最近向 SEC 提交的文件中提到的那些。在今天的问答环节，请每人限提一个问题和一个追问。请使用举手按钮加入提问队列。在进入问答之前，Elon 有一些开场白。

Good afternoon everyone and welcome to Tesla’s third quarter 2026 Q&A webcast. My name is Travis Aklerad, head of investor relations, and I’m joined today by Elon Musk, Vaibhav Taneja and a number of other executives. Our Q3 results were announced at about 3 p.m. Central Time in the update deck we publish at the same link as this webcast. During this call we will discuss our business outlook and make forward-looking statements. These comments are based on our predictions and expectations as of today; actual events or results could differ materially due to a number of risks and uncertainties, including those mentioned in our most recent filing with the SEC. During the question and answer portion of today’s call, please limit yourself to one question and one follow-up. Please use the raise hand button to join the question queue. Before we jump into Q&A, Elon has some opening remarks.

谢谢。第三季度非常强劲，我想向整个特斯拉团队致以热烈祝贺，他们努力工作才让这一切发生。对我们车辆的需求正在快速增长。需求大幅超过我们的产量，显然我们将会、而且正在有意义地爬坡增产。需求最大的驱动力是 Full Self-Driving，而今天，在你能买到的任何汽车里，FSD 其实都没有竞争对手。说清楚一点，我们正在尽我们所能生产和交付汽车，并且正在非常努力地推动 FSD 在更多市场获批。我们现在已经在欧洲许多国家获批，并希望获得欧盟范围内的批准。这很难预测，也许几个月之内，但如果拖过第一季度，我会感到意外。显然这个时间表不由我们控制，但我们确实预期会在整个欧洲和许多其他国家获得完全批准。我们也在第三季度开始生产 Tesla Semi 卡车，并做了新款 Roadster 的演示。我认为那是最令人兴奋的活动之一，可以说是有史以来任何产品里最酷的演示。这证明了特斯拉和 SpaceX 团队了不起的工作。我们也在为多得多的制造产能打基础。我认为我们的目标是到大约 2030 年达到每年 500 万辆，而且我认为我们很有机会实现这个目标。能源需求非常强。2027 年的 Megapack 我们实际上几乎已经卖光，我们正在尽可能快地爬坡生产以满足需求。我们也预期那块业务的毛利率会有意义地提高。所以，简而言之，经过很多年的苦干、事情相当平淡之后，我认为我们现在正走向很多乐趣和显著增长。关于 Robotaxi，我们看到 Robotaxi 的行驶里程，或者说完全自主的行驶里程，正在指数式增长，意思是增长非常快。这一点怎么强调都不为过。自主里程真的在以极快的周环比速度增长，而且我认为，随着我们继续爬坡，世界上没有任何公司能在自动驾驶汽车的规模上与我们相比。例如，奥斯汀的 Cybercab 车队自大约一个月前启动以来，已经增长了 10 倍。累计里程真的是——当我看每周的图表时——哇。我的意思是，它看起来基本上像一枚火箭在入轨。它增长得非常快。我建议，如果你想试一试，来奥斯汀，你会看到 Cybercab 在奥斯汀大规模使用，我想很多人在最近的 Austin City Limits 音乐节上也看到了。这真的在走向一个非常好的地方。关于 Optimus，我们已经在弗里蒙特的高产量产线上开始了我会称之为早期生产的阶段。现在我想把这一点展开。我以前说过：Optimus 的增长是一条 S 曲线，但 S 曲线的初始部分斜率平缓，而且有些颠簸，因为它是一个如此彻底的新产品。可以这样想：机器人的生产有大约 1 万件事必须做对，而它前进的速度取决于这 1 万件事里最慢、最不走运的那一件。但我们现在确实已经把一条每条线百万台产能的产线的绝大部分落到了弗里蒙特。我预计，要达到那个年化速率、达到每周 2 万台，大概至少需要一年，也许再长一点。从根本上说，这是因为设计里仍有一些部分在变。很多人问我 Optimus，所以我必须把这一点说透，因为如果人们不太理解新技术，或者不太理解复杂新硬件的量产，就会觉得，好吧，你不能立刻把这些东西大批做出来送出去吗？答案是，不行。你基本上得建起一整条巨大的全球供应链，在国内有一座非常非常大的工厂，而且一切都得能工作。设计上仍有一些地方需要调整，因为你要交付的是一台安全、可靠、寿命长的机器人，不能用一个月就磨坏。要让它真正胜任开放式任务，AI 和机器人本身还有相当多的工作要做。但正如我说的，我预期 AI、机电设计，以及可靠性和寿命，也就是机器人的使用寿命，这些问题会在明年年底左右得到解决。所以，那就是可以预期我们开始向特斯拉以外的公司交付 Optimus 的时间。我仍然确信，Optimus 会是地球上最大的产品，而人形机器人总体上会是地球上最大的产品。所以 Optimus 是一件非常大的事。但我想把预期调低：我们还有很多要闯过去，闯过所有这些大概还要一年左右。就我所知，在通用人形机器人这件事上，没有人哪怕接近特斯拉。关于 Terafab，位于奥斯汀的研究工厂进展很快。我们将在明年上半年开始调试设备，而且如果进展顺利，希望在大约 5 月或 6 月做出第一颗测试芯片，只是看看我们到底能不能做出任何芯片。所以这相当令人兴奋。我们的愿望是在大约 2028 年第三季度量产某种有用的芯片。正如我在 X 上公开说过的，我认为 Terafab 会成为世界上最有价值的建筑，其经济价值大概能排进国家的前 20 名。所以这是一个史诗级的项目，而且这将是一项极其重要的资产，甚至是必要的资产，以便规模化，尤其是 Optimus，因为我们可以应对半导体上的瓶颈，无论是存储、封装、逻辑，还是别的什么。关于与 SpaceX 的潜在合并，我认为昨天该说的我都说了，在那里是雷区里跳踢踏舞，就这样。最后，我们对自动驾驶和机器人路线图超级兴奋。有大量不可思议的好东西要来，我想再一次感谢特斯拉团队的强劲执行，也感谢所有支持者与我们同行。

Thank you. So the third quarter was very strong, and I’d like to give a big congratulations to the whole Tesla team for working hard to make it happen. Demand is growing rapidly for our vehicles. It’s exceeding our production by a significant margin, and obviously we’re going to be ramping—and we are ramping—production meaningfully. The biggest driver for demand is Full Self-Driving, and today there’s really no competitor for FSD in any car that you can buy. And to be clear, we are producing and delivering as many cars as we can, and we’re pushing very hard for FSD approval in additional markets. We’re now approved in many countries in Europe, and we’re hoping for an EU-wide approval. It’s hard to predict—maybe within a few months—but I would be surprised if it takes longer than Q1. Obviously that’s not a timeframe we control, but we do anticipate being fully approved throughout Europe and many other countries. We also began production of the Tesla Semi truck in Q3 and did the new Roadster demo, which was, I think, one of the most exciting events—arguably the coolest demo of any product ever. And that’s a testament to amazing work by the Tesla and SpaceX teams. We’ve also been laying the groundwork for a lot more manufacturing capacity. We’re, I think, aiming for 5 million vehicles a year by approximately 2030, and I think we’ve got a good shot at achieving that goal. Energy demand is very strong. We’re actually almost sold out of all Megapacks for 2027, and we’re ramping production as quickly as possible to meet demand. We also expect gross margins on that to increase meaningfully. So, in a nutshell, after many years of sort of grind and things being fairly flat, I think we’re now headed for a lot of fun and significant growth. Regarding robotaxi, we’re seeing the robotaxi miles driven—or fully autonomous miles driven—scale exponentially, meaning scaling very rapidly. And it would be hard to overstate this point. It really is an extremely fast week-over-week growth in autonomous miles, and I don’t think there’s any company in the world that can match our scale with autonomous vehicles as we continue to ramp. The Cybercab fleet in Austin, for example, has grown by a factor of 10 since launch about a month ago. And the cumulative mileage is really—when I look at the weekly chart—it’s wow. I mean, it looks like a rocket going to orbit, basically. It’s growing very fast. I recommend if you want to try it out, come to Austin and you’ll see Cybercabs in use at scale in Austin, as I think a lot of people saw with the Austin City Limits concert recently. This is really going to a very good place. And regarding Optimus, we’ve started what I’d call early production on our high-volume line in Fremont. Now I want to elaborate on this point. I’ve said this before: the growth of Optimus—it’s an S-curve, but the initial part of the S-curve has a shallow slope and is somewhat bumpy because it is such a radically new product. The way to think about it is there are like 10,000 things that have to go right for production of the robot, and it’ll move as fast as the slowest, least lucky of the 10,000. But we do have the vast majority of a million-unit-per-line production line landed in Fremont right now. I would expect that it takes us probably at least a year, maybe a little longer, to get to that annualized rate—to get to 20,000 units a week. This is just fundamentally because there are still parts of the design that are changing. A lot of people ask me about Optimus—that’s why I’ve got to belabor this point—because if people don’t understand new technology that well, or volume production of complex new hardware that well, it sort of seems like, well, can’t you just make a whole bunch of these things immediately and send them out? And it’s like, no—you’ve got to basically build up an entire gigantic global supply chain and have, domestically, a very, very big factory, and everything’s got to work. There are still some things we need to tweak about the design, because you need to deliver a robot that is safe and reliable and has long duration—so it can’t wear out after a month. And there’s still quite a lot of work to do with regard to the AI and the robot for it to be truly useful for open-ended tasks. But, like I said, I would expect that the AI, the electromechanical design, and the reliability and duration—like the usage life of the robot—I do expect these issues to be solved around the end of next year. So that’s the time to expect when we’re delivering Optimus to companies outside of Tesla. And I remain convinced that Optimus will be the biggest product on Earth, and humanoid robots in general will be the biggest product on Earth. So Optimus is a very big deal. But I want to temper expectations that we still have a lot to get through, and it’s about a year or so that it’ll take us to get through all that stuff. As to the best of my knowledge, there is no one that is even close to Tesla with respect to a general-purpose humanoid robot. Regarding the Terafab, the Austin-based research fab is progressing quickly. We’ll begin commissioning machinery in the first half of next year and, hopefully, if things go well, we’ll make our first test chip—just to see if we can make any chip at all—around May or June. So that’s pretty exciting. Our aspiration is to be making some kind of useful chip at volume around Q3 of 2028. And as I’ve said publicly on X, the Terafab, I think, will be the most valuable building in the world and will probably have an economic value that will place it in the top 20 of countries. So it’s an epic project, and this will be an incredibly important asset, if not necessary asset, in order to scale—especially Optimus—as we can address bottlenecks in semiconductors, whether it’s memory, packaging, logic, or whatever the case may be. With regard to a potential merger with SpaceX, I think I said everything that should be said yesterday—tap dancing through a minefield there—and yeah, that’s it. In conclusion, we’re super excited about our autonomy and robotics roadmap. There’s a lot of incredibly awesome stuff coming, and I’d like to once again thank the Tesla team for their strong execution and all of our supporters for joining us on this journey.

很好。非常感谢，Elon。Vaibhav 也有开场白。

Great. Thanks very much, Elon. And Vaibhav has opening remarks as well.

谢谢，Travis。第三季度我们在很多方面继续推进。Elon 已经讲了其中很大一部分。我想补充的是，我们也开始交付使用自产 LFP 电池的 Megapack，并且第一次出现连续两个季度的车辆交付超过 48 万辆。在汽车业务上，我们以四年来最高的积压订单结束这个季度，交付主要受产量约束。因此我们专注于提高所有工厂的产量，以满足这不断增长的需求。虽然我们正在积极提高产量，但从供应链到制造再到交付物流，把每一个环节都调到位确实需要一些时间。Model Y 在本季度继续表现出色，并在几个关键市场创下交付纪录。根据我们收集的初步数据，我们相信它是欧洲销量第一的纯电动车，也是韩国销量第一的乘用车。我们还在包括澳大利亚、日本和台湾在内的关键市场创下新纪录。正如 Elon 所说，FSD 继续是第三季度需求的关键驱动力。北美超过 75% 的交付在交付时就签约了付费 FSD 订阅，这大约比第二季度高出 20%。总体而言，FSD 的附带率继续改善，全球付费客户超过 170 万，其中订阅第一次占到 50%。鉴于 FSD 在美国的强度，我们正在采取额外步骤，在欧洲销售某些带有一些权益的车，试图在更广泛的 FSD 批准到来之前，把更多车队放到外面。在汽车方面，剔除积分后的毛利率环比从 16.3% 提高到 18.1%。毛利率的改善由亚太交付占比更高所推动。大宗商品价格上涨和利率变化继续增加我们的成本，并造成进一步的逆风，我们已经调整了车辆融资方案，以部分抵消这类成本。在能源业务上，我们部署了 13.7 GWh 的储能，这是我们部署量第二好的季度。能源毛利率从 20.4% 提高到 33.1%。环比毛利率的改善，主要归因于本季度确认的关税退税，以及与供应商就某些电芯相关成本将由他们报销达成一致之后的质保冲回。正如 Elon 所说，这项业务正目睹来自数据中心的极端需求，我们正在尽最大努力提高产量，因此向前走时，我们会看到 ASP 的改善，以及毛利率的进一步改善。服务及其他的毛利率环比从 14.1% 提高到 15.9%，创历史新高。这是服务及其他里多项业务走强的结果，包括超级充电和保险。Robotaxi 业务属于服务及其他，随着 Cybercab 在奥斯汀开始载客而继续增长。本季度无人监管里程累计增长了 8 倍，因为整个季度都有两位数的周增长。我们正在把奥斯汀扩张中的全部经验拿来，帮助我们为在其他市场更广泛地推出 Cybercab 做准备。我们的优先事项是，通过提供最佳体验的车辆，比如 Cybercab，推进网络的密度和扩张。我们也会随时间继续用其他车型扩充车队，以应对全体人口的地面交通需求。我想强调，我们试图做的，字面上就是接管地面交通的每一个部分，不只是运送人。我们从把人从一个地方送到另一个地方开始，但我们想基本上拥有地球上任何一种运输。我们正处于一个巨大的投资周期，为未来做准备。因此，正如先前指引的那样，运营费用继续环比增加。增加的大部分来自大量的研发相关活动，包括 AI 开发成本，比如额外计算机芯片开发带来的折旧，以及 Optimus、Megapack 3 和 Semi 卡车等新产品的试生产爬坡成本。运营费用增长的这一趋势会持续到我们把这些新产品商业化。净利润受到我们持有的 SpaceX 股权按市值计价减少 1.42 亿美元的影响，这部分被大约 2.9 亿美元的比特币按市值计价收益所抵消。我们继续努力高效、负责地管理现金。我们以超过 15 亿美元的正自由现金流，以及超过 450 亿美元的现金和投资总余额结束这个季度。由于我们高效地管理资本部署，我们正在把今年的资本开支指引修订为大约 200 亿美元，并计划主要用经营现金流来资助它。此外，本季度我们锁定了最多借入 300 亿美元的能力，以支持本公历年预期的增长。我们从关税退税收到的钱，正在被再投资，以扩大美国的产能。我想强调，我们喜欢做大量硬核的、真正的苦活，把生产设施建起来并运转。未来几年我们的资本开支会继续扩大，因为我们要增加更多 Robotaxi 车队，扩大 Optimus 的产能，安装太阳能制造产能，购买额外的 AI 算力基础设施，并建造一座半导体工厂。这一战略将把公司放到下一个增长时代的位置上，颠覆许多现有市场，同时创造新市场。这是在特斯拉最令人兴奋的时刻。围绕自动驾驶、机器人和能源正在打下的基础，正在为我们准备一个极其光明的未来。我想以感谢作结：感谢特斯拉团队、我们的客户、投资者和供应商，在这段旅程上对我们有信心。

Thanks, Travis. We continued our progress on a lot of fronts in Q3. Elon already covered a lot of them. Additionally, I’d like to add we also started deliveries of Megapacks with our in-house LFP batteries and, for the first time, two consecutive quarters of vehicle deliveries of over 480k. On the automotive business, we exited the quarter with our highest backlog in four years, with deliveries constrained primarily by production. We’re therefore focused on increasing production at all our factories to meet this growing demand. While we are actively working to increase production, it does take some time to get every aspect dialed in from supply chain to manufacturing to delivery logistics. The Model Y continued to outperform in the quarter and set delivery records in several key markets. Based on the preliminary data which we’ve gathered, we believe that it was the number one selling BEV in Europe and the number one selling passenger car in South Korea. We also set new records in key markets including Australia, Japan, and Taiwan. As Elon said, FSD continues to be the key driver for demand in Q3. Over 75% of deliveries in North America included a sign-up for paid FSD subscription at the time of delivery, which is approximately 20% higher than Q2. Overall, the FSD attach rate continued to improve, reaching over 1.7 million paid customers globally, of which, for the first time, 50% is based on subscriptions. Given the strength of FSD in the U.S., we are taking extra steps to sell certain cars in Europe with some benefits to try and get more fleet out there before a broader FSD approval is obtained. On automotive, margins excluding credits increased sequentially from 16.3% to 18.1%. The improvement in margins is driven by a higher mix of deliveries in APAC. While commodity price increases and interest rate changes continue to add to our cost and create further headwinds, we have adjusted the vehicle financing programs to partially offset such cost. On the energy business, we had our second-best quarter of deployments of 13.7 GWh of energy storage. Energy gross margins improved from 20.4% to 33.1%. The sequential margin improvement is primarily attributed to tariff refunds recognized in the quarter and warranty true-downs resulting from alignment with our vendor on certain cell-related costs to be reimbursed by them. As Elon said, the business is witnessing extreme demand from data centers and we are doing our best to increase production and, accordingly, we’ll see improvements in ASP and further improvements in margins as we move forward. Services and other margins improved sequentially from 14.1% to 15.9%, an all-time high. This was the result of strength across several of the businesses in services and other, including Supercharging and Insurance. The Robotaxi business, which is part of Services and Other, continued to grow with the launch of Cybercab rides in Austin. Cumulative unsupervised miles 8x in the quarter, as we had weekly double-digit growth throughout the quarter. We’re taking all the learnings from our Austin expansion to help us prepare for broader Cybercab launch in other markets. Our priority is progressing the density and expansion of the network through vehicles that provide the best experience, like the Cybercab. We will also continue to augment the fleet with other variants over time to address the terrestrial transportation needs of the total population. I’d like to stress that what we are trying to do is literally take over every part of terrestrial transportation, not just getting people, and we’re starting with getting people from one place to the other, but we want to basically own the transportation of any kind on Earth. We are in a huge investment cycle to set us up for the future. Therefore, as previously guided, operating expenses continue to increase sequentially. Most of the increase came from significant research and development–related activities, including AI development costs such as depreciation from additional computer chip development, and pre-production ramp costs for new products like Optimus, Megapack 3, and Semi truck. This trend of growing OPEX would continue till we commercialize these new products. Net income was impacted from a mark-to-market reduction of our SpaceX holding of 142 million, which was offset by a Bitcoin mark-to-market gain of approximately 290 million. We continued our efforts to manage cash efficiently and responsibly. We ended the quarter with positive free cash flow of over 1.5 billion and an overall cash and investment balance of over 45 billion. We’re revising our CapEx guidance to be approximately 20 billion for the current year as a result of managing our capital deployment efficiently and plan to fund it primarily from our cash flows from operations. Additionally, during the quarter we secured capacity to borrow up to 30 billion to support the expected growth this calendar year. The money that we received from the tariff refunds is being reinvested to expand production capacity in the U.S. I’d like to stress that we like to do a lot of hardcore, real grunt work to get production facilities up and running. Our CapEx for the next few years would continue to expand as we add more Robotaxi fleet, we expand our production capacity for Optimus, install manufacturing capacity for solar, purchase additional AI compute infrastructure, and build a semiconductor fab. This strategy will position the company for the next era of growth and disrupt many existing markets while creating new ones. This is by far the most exciting time to be at Tesla. The foundation being laid around autonomy, robotics, and energy is setting us up for an extremely bright future. I would like to end by thanking the Tesla team, our customers, investors, and suppliers for their confidence in us on this journey.

太好了。谢谢你，Vaibhav。现在我们转到 Say.com 的问题。第一个问题是：Cybercab 车队如何规模化到奥斯汀以外，鉴于 NHTSA 仍有一项未结束的安全审查，得克萨斯州以外的监管路径是什么？Ashok，你想回答这个吗？

Fantastic. Thank you, Vaibhav. Now we’re going to go to questions from Say.com. The first question is, how does the Cybercab fleet scale past Austin and what’s the regulatory path outside Texas given NHTSA’s open safety review? Ashok, did you want to take that?

好，我从这里开始。首先，我们本季度推出了 Cybercab，这主要是靠从其他车辆平台学习数据做到的，例如 Model Y、Model 3 等等。所有这些数据帮助我们做出了 Cybercab，而不需要离谱数量的车辆来产生数据。所以我们在 Cybercab 的 AI 能力上看到非常好的规模化。我们在奥斯汀推出，那只是因为我们有了一些车，而且我们想让 Cybercab 遍布奥斯汀。但没有根本性的理由把我们限制在奥斯汀。事实上，下个月我们将在达拉斯推出，然后同一套软件就会连同 Model Y 的 Robotaxi 一起，扩展到很多其他城市。

Yeah, I’ll start here. First of all, we launched Cybercab this quarter, which we achieved by mostly learning from data from the other vehicle platforms, e.g., Model Y, Model 3, etc. All of the data helped make the Cybercab without requiring a ridiculous amount of vehicles to produce the data. So we are seeing very good scaling in terms of Cybercab’s AI capabilities. We launched in Austin, but that’s only because we had some cabs and we want to swarm Austin with Cybercabs everywhere. But there’s no fundamental reason that we are limited to Austin. In fact, next month we’ll be launching in Dallas, and then the same software just scales to lots of other cities along with the Model Y robotaxis.

很好，非常感谢。下一个问题是：对于 Optimus，你怎么看众包训练数据，相对于你们内部的 Optimus Academy 方法？数据目前是主要瓶颈吗？我想你们已经讲了不少，但还有什么别的补充吗？

Great, thanks very much. The next question is, what are your thoughts on crowdsourcing training data for Optimus versus your internal Optimus Academy approach? Is data currently the primary bottleneck? I think you covered this quite a bit, but any other color there mostly?

我已经谈过这个，但不是。众包不是 Optimus 的问题。关于我们仍在做的 AI 架构，有很多事情必须做对。我们进展良好，但那需要理清楚。而且真的，为了让 Optimus 拿到我们需要的训练数据，我们实际上得造出很多 Optimus。所以我们需要造出，我认为，大约 1 万到 2 万台机器人这个量级，然后让它们做自我对弈。我们正在建设一座 Optimus Academy，Optimus 可以在那里练习非常广泛的任务，然后我们可以用这些数据形成一种飞轮，让它在这些不同任务上越来越好。Optimus 和汽车的一个很大不同是，我们没有一支可以从中收集训练数据的庞大车队。所以对 Optimus，我们必须靠造出机器人、让它们自我对弈来引导启动，而这其实就是人类学习在世界上行动的方式。你知道，当你有一个刚学走路的孩子，你能看到这一点：他们长大，一开始有点摇晃，不知道怎么做事，最终学会自己擦屁股，这是一个大里程碑，任何有过孩子的人都知道。然后他们学会阅读，学会做各种各样的事，并且能够做这些事。他们是靠玩学会的。我们实际上需要 Optimus 大规模地玩，才能让它真正擅长现实世界里的任务。Ashok，你也许想补充一下。

I talked about this already, but it’s not. Crowdsourcing is not the issue for Optimus. There’s a lot of things we have to get right about the AI architecture that we’re still working on. We’re making good progress, but that needs to be sorted out. And really, in order for Optimus to get the training data we need, we actually need to make a lot of Optimus. So we need to make, I think, on the order of 10 or 20,000 bots and then have them do self-play. And we’re building out an Optimus Academy where Optimus can practice doing a very wide range of tasks, and then we can use that data in sort of a flywheel to make it better and better at these various tasks. A big difference with Optimus versus the cars is we don’t have this vast fleet of cars from which we can gather training data. So with Optimus, we have to bootstrap it by making the robots and having them do self-play, which is really how a human learns to navigate the world. You know, you can see this when you have a toddler, and they grow up and they initially are sort of wobbly and they don’t know how to do things, and eventually they learn to wipe their own ass, which is a big milestone, as anyone who’s had kids knows. And then they learn to read and do all sorts of things and be able to do these things. And they learn this by playing. And we actually need Optimus to play at scale in order to have it be really good at tasks in the real world. Ashok, you may want to add to that.

是的，就像我提到的，Elon，对汽车来说，我们实际上有来自汽车本体自身的数据。但对 Optimus，没有外骨骼可以让人钻进去，把 Optimus 变成木偶之类的东西。Optimus 必须先靠看别人做任务来学习，然后开始自己练习这项任务。例如，你可以想象，如果你在看一场足球赛，仅仅因为你在看，并不意味着你自己就会踢足球。你得开始练习，把这项运动练好，并从机器人自己的经验里学习。我们正在建立这个巨大的强化学习循环：先用来自人类的基础训练做引导，然后放到 Optimus 上，做大量自我对弈，既在仿真里，也在现实世界里，在 Optimus 自己的本体上学习。我认为我们相当接近在这个大循环里得到非常好的结果。

Yeah, like I mentioned, Elon, for the car we actually have the data from the car’s embodiment itself. But for Optimus, there’s no exoskeleton for humans to get into and make Optimus into a puppet or something like that. Optimus has to learn by watching other people do tasks initially and then start practicing the task itself. For example, you can imagine if you’re watching a soccer game, just because you’re watching it doesn’t mean you’ll be good at soccer yourself. You got to start practicing and get good at the game and learn from the bot’s own experience. And we are setting up this huge reinforcement learning cycle with initial base training done from humans to bootstrap, and then put it on Optimus and do a lot of self-play, both in simulation and in the real world, to learn on the embodiment of Optimus itself. I think we’re pretty close to having very good results in this big loop.

是的，我认为我们也许再过一个月左右就可以准备发布一段 Optimus 视频。我想确保，当我们发布这段视频时，它不是一段预先编排的演示。当你看人形机器人演示时，有些看起来非常惊人，但它们并没有真正使用任何先进的 AI。我的意思是，它们是预先编程的确定性任务。如果你真的试着在现实世界里使用那台人形机器人，你会遇到很多麻烦。它基本上不会工作。所以我想发布一段视频，里面是 Optimus 在做没有预先编程的事。我认为那也许还要一两个月，才能有一个你可以相信自己所看到的那种演示。然后还有另一个我们想达到的重要里程碑，就是把特斯拉 AI5 芯片集成进 Optimus。如果事情按计划进行，那将是大约明年 4 月。那是能力上的巨大提升，每秒运算高出一个数量级。AI5 相对 AI4 是多出一个数量级的算力。这会帮到 Optimus 很多，因为人形机器人必须处理的自由度，远远大于一辆汽车必须处理的。所以，要把操作 Optimus 全部自由度、以及那么广泛的任务所需的一切都塞进去，是有一点吃力的。你可以想，一个人形机器人做的事是一辆汽车的一千倍，基本上是某个非常大的数字。所以我认为 AI5 的额外算力会带来很大差别，而且希望在明年 4 月左右放量。这就是为什么我认为所有这些拼图，机电灵巧度的打磨、确保 Optimus 能长时间运转而不磨坏、把 AI 软件做对、以及集成 AI5 芯片，这些东西会汇到一起。所以我认为大约一年之后，就是你可以交付给客户、同时产量也在达到规模的时候。所以，就 Optimus 的首批客户而言，那大概是该去想的正确时间点。我认为一开始大概会是工业应用，危险的、重复的任务，人们真的不想做、而你希望基本上 24 小时不停做的事。全天候的活，那会是 Optimus 一开始价值最高的地方。随着我们扩大产量，它当然会成为人们可以买回家的东西。但最初的高价值用途会是重工业、危险任务，或者可能造成重复性劳损、而你希望 24 小时不停完成的事。可以这样想。很好，谢谢各位。

Yeah, I think we may be ready to release an Optimus video in a month or so. I want to make sure that when we release this video that it’s not a canned demo. When you see humanoid robot demonstrations, some of them look very impressive, but they’re not really using any kind of advanced AI. I mean, they’re pre-programmed deterministic tasks. And if you actually try to use that humanoid robot in the real world, you would have a lot of trouble. It wouldn’t work, basically. So I want to release a video where it’s Optimus doing things that are not pre-programmed. And I think that maybe it’s a month or two for that, to have a demo that you can believe what you see type of thing. And then there’s another milestone that’s important we want to get to, which is to integrate the Tesla AI5 chip into Optimus, which, if things go according to plan, will be in approximately April of next year. And that’s a giant improvement in capability, with an order of magnitude more operations per second. It’s an order of magnitude more compute, AI5 versus AI4. And that will help Optimus a lot because the degrees of freedom that a humanoid robot has to deal with are far greater than what a car has to deal with. So it is a bit of a struggle to squeeze everything needed to operate with all the degrees of freedom of Optimus and the wide range of tasks. You can think of a humanoid robot doing a thousand times more things than a car would do—some very big number, basically. So I think the additional compute of AI5 will make a big difference, and that comes in hopefully in volume around April of next year. That’s why I think all the pieces—the refinements in electromechanical dexterity, making sure Optimus can operate for long periods of time without wearing out, getting the AI software right, and the AI5 chip integrated—all of these things kind of come together. That’s why I think about a year from now is when it’s something that you could ship to customers and it would also be reaching scale in production. So that’s probably the right timing to think about in terms of the initial customers for Optimus. I think it probably would be things that are industrial applications at first—things that are dangerous, repetitive tasks, things that people really don’t want to do and that you want done basically 24/7. So around-the-clock stuff—that’s going to be the highest value for Optimus in the beginning. As we scale production, of course it would be something that people can buy for their home. But the initial high-value use will be kind of heavy industry, dangerous tasks, or things that may cause repetitive stress injuries and that you want done 24/7. So that’s the way to think about it. Great, thank you, guys.

下一个问题非常相似，是关于 Optimus 的规模化。所以我们可以跳过那一题。下一个问题是：我们什么时候可以预期客户的车上出现无人监管的 FSD？Ashok，你想先讲这个吗？

The next question is very similar. It’s about Optimus scaling. So we can go past that one. And the next question is, when can we anticipate FSD unsupervised in customer cars? Ashok, did you want to kick that off?

好。就像我刚才提到的，Cybercab 和 Model Y 的软件是同一套，很快我们就能把同一套软件作为 V15 版本的一部分发布给客户。唯一的额外要求是确保，例如，客户的 Model Y 能够进入特斯拉为了让它完全自主而设的地理围栏和车队管理。我们估计，大约在明年第二季度，我们可以打开首批客户车辆，让它们成为 Robotaxi 车队的一部分。

Yeah. Like I mentioned earlier, the software for Cybercabs and Model Y is the same, and pretty soon we’ll be able to release the same software as part of the V15 release to customers. The only additional requirement is to make sure that, for example, the customer Model Ys can be part of the geofences and the fleet management that Tesla has for it to be fully autonomous. We suspect somewhere around Q2 next year is when we can turn on some of the first customer vehicles to be part of the Robotaxi fleet.

很好，谢谢你，Ashok。下一个问题是：Cybercab 什么时候可以供客户购买、拥有？Lars，你想先讲这个吗？

Great, thank you, Ashok. The next question is, when will the Cybercab be available for customer purchase ownership? Lars, did you want to kick this one off?

当然。谢谢，Travis。我认为这有一点是建立在 Ashok 刚才说的话上，但简短地说，我们正试图在今年、也就是 2026 年年底之前把 Cybercab 交到客户手里。显然我们正在得克萨斯的产线上生产它们，但目前所有这些都流向我们自己内部的 Robotaxi 车队。随着特斯拉成长，我们将主要向客户交付完全按自动驾驶设计的车辆，而 Cybercab 会是第一款。所以，我们今年从小规模开始，并学习 Ashok 提到的关于交付、所有权，以及自动驾驶车辆在客户手里的车队管理的所有事情，然后我们会继续把这件事规模化。Elon，我不知道你有没有什么想补充的。

Sure. Thanks, Travis. I think it builds a little bit on what Ashok just said, but in short, we’re trying to get Cybercabs to customers by the end of this year, 2026. Obviously we’re producing them on our line in Texas, but all of those are currently going to our own internal Robotaxi fleet. And as Tesla grows, we’ll mainly deliver fully designed autonomous vehicles to our customers, and Cybercab will be the first. And so, as we start small this year and we learn about all the things Ashok mentioned about delivery and ownership and fleet management of autonomous vehicles in customers’ hands, we’ll continue to scale that. Elon, I don’t know if there’s anything you want to add to that.

没有。我的意思是，Cybercab 主要会是一种作为运输即服务来提供的东西，而我们的其他车更会是你想拥有的车，比如 3 和 Y 之类的，当然还有 Roadster。但 Cybercab 的绝大部分，本质上会是特斯拉拥有的，并且是运输即服务。

No, I mean, Cybercab is primarily going to be something that is offered as transport-as-a-service, whereas our other cars would be more the cars that you’d want to own, like the 3 and the Y and whatnot, and the Roadster, obviously. But the vast majority of the Cybercab stuff will be essentially Tesla-owned and transport-as-a-service.

很好，谢谢各位。接下来两个问题相当相似，都围绕单位经济成本。第一个是每付费 Robotaxi 英里，第二个是每台 Optimus 的成本。首批付费客户是什么样子，我们已经讲过了。所以，VT，你想在这里进入单位经济这个问题吗？

Great, thanks, guys. The next two questions are quite similar around unit economic costs. The first one is on per paid Robotaxi mile, and the second one is cost per Optimus unit. And we already covered what the first paying customer looks like. So, VT, did you want to get into the unit economics question here?

好。在 Robotaxi 这边，我们正在运营的主要地区，已经看到利用率在改善。一周又一周，正如 Elon 在说的，我们看里程，以及有多少付费行程在发生。那么我们今天达到 30 美分的目标了吗？没有。但我们看得到一条走到那里的路径吗？非常——是的，我们看得到。而这完全是关于把运营规模化。接下来几个月你会看到的是，今天 Cybercab，也就是主要的 Robotaxi 产品，只在奥斯汀可用，但我们确实计划在今年年底之前把它带到其他一些城市。随着我们增加更多 Cybercab，我们也在从管理一支车队的运营经验里学习：接客和送达、清洁，所有这些。所以这些成本都在被调到位。这就是为什么，就像我之前说的，我们肯定看得到一条走到 30 美分目标的路径。

Yeah. So on the Robotaxi side, the main geographies where we are operating, we’re already seeing utilization improving. And week by week, as Elon was talking about, we look at the miles and how many paid rides are happening. So are we at our goal of 30 cents today? No. But do we see a path to get to that? Very—yes, we do. And it’s all about scaling operations. What you’ll see over the next few months is that today Cybercab, which is the primary Robotaxi product, is available only in Austin, but we do plan to bring it out in some other cities before the end of this year. And as we add more Cybercabs, we are also learning from the operational experience of managing a fleet—pickup and drop-off, cleaning, all those things. So all those costs are being dialed in. And that’s why, like I said before, we definitely see a path to get to our 30-cent target.

然后你想讲一下 Optimus 的成本吗？

And then did you want to cover the Optimus costs?

好，我的意思是，我认为就设计已经成熟到什么程度而言，我们内部对 Optimus 成本的跟踪相当好，但要说我们已经到达那里，还稍微太早，因为就像 Elon 说的，我们仍在做很多设计迭代。好的一面是，当你从成本的角度看，我们正趋近于我们会说 Cybercab 成本最终会落到的那个位置。

Yeah, I mean, I think Optimus costs internally we are tracking pretty well in terms of how the design has matured, but it would be a little too early to say that we have got there, because like Elon was saying, we’re still making a lot of design iterations. The good thing is when you look at it from a cost perspective, we’re trending closer to where we would say our Cybercab costs would end up being.

很好，非常感谢。这个我们其实已经回答过了：你们预期 Optimus 什么时候会产生收入？所以我们在前一个问题里讲过了。下一个是：FSD 的地图继续是一个问题。三个最大的问题是识别正确的限速、停车，以及典型路线。我理解你们使用各种第三方地图，但为什么不用你们自己的地图，授权受信任或经过验证的特斯拉车主来提供输入？Ashok，你想从这里开始吗？

Great, thank you very much. We’ve kind of already answered this—when do you anticipate Optimus will be generating revenue? So we covered that in the prior question. The next one is: mapping for FSD continues to be a problem. The three biggest issues are recognizing proper speed limits, parking, and typical routes. I understand you use various third-party maps, but why not use your own maps—authorize trusted or verified Tesla owners to provide the input? Ashok, did you want to kick off here?

对。是的，你完全正确，我们不该依赖外部的人的地图。所以我们正在建我们自己的地图，但不是再多一张地图那种方式。我们正在建一套全新的架构，本质上是基于 AI 的地图和导航。它会超级棒。它会和语音集成。你可以跟它说话，要求不同的路线。它会学习你的个性化偏好，再加上车队的偏好。所有汽车的集体智能会流进去，让地图和导航变得超级酷。你会看到，行程的起点和终点很快就会作为 V15 版本的一部分使用这项技术，接下来几个月它会接管全部导航功能，停车、目的地处理，并且就驾驶这项任务而言，它在某些方面会比你更了解你自己，并完整地处理它。

Yep. Yeah, you’re totally right that we shouldn’t be depending on external people’s maps. So we are building our own maps, but not the same way as one more maps. We are building an entirely new architecture with essentially AI-based mapping and navigation. It’s going to be super awesome. It’s going to be integrated with voice. You can talk to it, ask for different routes. It’s going to learn your personalized preferences plus also the fleet preferences. The collective intelligence of all the cars is going to flow into making the maps and navigation be super cool. You’ll see the start and end of trips already use this technology pretty soon as part of the V15 release, and over the next several months it’s just going to take over all of the navigation functions—parking, destination handling—and it will just know you better than you know yourself in some regard for the driving task, and handle it fully.

很好，谢谢你，Ashok。然后我们最后一个问题是：你们现有四座工厂目前的生产速率是多少？你们预期会建新的吗？Lars，你想先讲这个吗？

Great, thank you, Ashok. And then our last question is, what is the current production rate at your four existing factories? Do you anticipate building new ones? Lars, do you want to kick that off?

好，当然。我认为 Vaibhav 在开场白里有一点谈到了这个。我要说，我们的工厂本季度正以纪录速率运转，当然它们只能以最慢的工序或零部件的速率来造。Vaibhav 提到，我们正在积极管理生产的所有方面。你知道，正如我们在过去的电话会上说过的，过去一年多我们在全球做了战略投资，以提高现有工厂的速率，本季度我们开始看到这一点，并预期随着这些投资结果落地，接下来几个季度会继续。除此之外，我们也在爬坡 Cybercab，显然是在 Giga Texas。我们申请了适当的许可，以扩建柏林的工厂。正如 Elon 提到的，我们刚刚在里诺启动了新的 Semi 工厂。所以，简短地说，我们正在建新工厂。而且我们的目标是到 2030 年达到那 500 万辆，正如 Elon 说的。

Yeah, sure. I think Vaibhav kind of touched a little bit on this in his opening remarks. And I’ll say our factories are operating at record rates this quarter, and of course they can only build at the rate of the slowest process or component. And Vaibhav mentioned that we are actively managing all the aspects of production. You know, as we’ve spoken in our past calls, we’ve made strategic investments globally to increase rates of our current factories over the last year plus, and this quarter we are starting to see that and we expect that to continue in coming quarters as those investments come to fruition. Beyond that, we’re also working to ramp Cybercab, obviously in Giga Texas. And we applied for the proper permits to expand the factories in Berlin. And as Elon mentioned, we just started our new Semi factory in Reno. So, in short, we are building new factories. And we’re aiming to hit that 5 million goal by 2030, as Elon said.

很好。非常感谢。Say.com 的问题到这里。现在我们进入分析师提问。提醒一下，正如 Elon 所说，今天我们不会回答任何关于与 SpaceX 潜在合并的问题。那么，第一个问题来自 RBC 的 Tom。Tom，准备好了请取消静音。

Great. Thank you very much. That’s it for the Say.com questions. Now we will be going to analyst questions. As a reminder, as Elon said, we will not be addressing any questions today on a potential merger with SpaceX. With that, the first question comes from Tom from RBC. Tom, please unmute yourself when you’re ready.

好，谢谢接受这个问题。是关于 FSD 的。附带率，按我的计算，在北美付费基础上大约是 30% 到 40%。正如你们说的，它正在成为更多汽车销售的飞轮，而且这个附带率大大超过其他车企基于订阅的自动驾驶。我想，以一种奇怪的方式，你们会不会成为自己成功的受害者：其他车企注意到这种成功，试图积极做出类似的 L2 以上类型的产品，也许与 NVIDIA 或 Qualcomm 合作？如果那发生了，会不会对 FSD 的定价形成通缩？

Yeah, thanks for taking the question. It’s on FSD. So the attach rate, I think according to my math, is like 30 to 40% in North America on a paid basis. It’s becoming a flywheel, as you said, for more car sales, and the attach rate exceeds other automakers’ subscription-based autonomy by a lot. I guess in a weird way, could you guys be a victim of your own success in seeing other carmakers noting the success and trying to aggressively come up with similar L2-plus type products, maybe partnering with NVIDIA or Qualcomm? And if that happens, could that be deflationary for FSD pricing?

我认为人们把这件事想得非常简单，觉得你可以很容易地，软件可以被解决。不是这样。我让 Ashok 来展开，但基本上这是一个非常难的问题。

I think people take this thing very simply—that you can easily, the software can be solved. It’s not. I’ll let Ashok get into it, but basically it’s a very hard problem.

这是一个超级难的问题。是的，做到前 90%、做出一个看起来很酷的演示，是容易的。但当你进入这个问题的长尾，它长到人们没有意识到：任何头脑正常的人都不该试图解决这个问题。它如此困难、尾巴如此之长，尤其是要到达一个安全点，让你能安全地开几百万英里而没有伤害等等。那会极其困难。当我们把无人监管的 FSD 发布到更广泛的客户车队时，我认为几年之内没有人能与之相比。

It’s a super hard problem. Yeah, it’s easy to get to the first 90% and make a cool-looking demo. But when you get into the long tail of this problem, it’s so long that people don’t realize anyone with their sane mind should not attempt to solve this problem. It’s so difficult and so long-tailed, and especially to get to a safety point where you can drive safely for millions of miles with no injuries and so on. That’s going to be extremely hard. As we release unsupervised FSD to the broader customer fleet, I don’t think anyone can match it in several years.

还有一件事我要说，如果刚才不够清楚：尤其在我们美国这里的展厅，客户现在走进来会说，我们能体验一次 FSD 驾驶吗？他们不是在要 3 或 Y 或 Cybertruck；他们说的是，我们能体验一次 FSD 驾驶吗？而这已经成为关键驱动力，因为它实际上为人群里不同的部分打开了多得多的灵活性，这些人以前在方便地开车这件事上有困难。现在他们靠能够使用 FSD，把自由拿了回来。而且很多人就是觉得，我的意思是，这是一个改变游戏的东西。我的意思是，如果你还没试过，我每次都这么说，一旦你试了，你就没法再回去开任何别的车。

And one other thing which I’ll say, if it wasn’t clear—especially in our showrooms here in the U.S., customers are now coming in and saying, can we get an FSD drive? They’re not asking for 3 or Y or Cybertruck; they’re saying, can we get an FSD drive? And that has become the key driver, because what it does is it actually opens up a lot more flexibility for different parts of the population wherein they had issues driving conveniently before. Now they’re getting their freedom back by being able to use FSD. And a lot of folks are just like, I mean, it’s a game changer. I mean, if you haven’t tried it—I keep saying it every time—if you do it, you won’t be able to go back to driving any other car.

好，追问和那个关于中国的问题类似。我的意思是，在那里自动驾驶似乎是标配。答案是不是和第一个问题同一类：那些车企将无法开发出与你们在 FSD 上已有的类似的产品？还是你认为定价动态肯定会不同？

Yeah, follow-up is similar to that question on China. I mean, there it seems like autonomy is table stakes. Is it the same kind of answer as question one—that they’re not going to be able, those carmakers aren’t going to be able to develop similar products that you guys have had on FSD? Or do you see there being definitely a different pricing dynamic?

我的意思是，一开始，就这项技术而言我们拥有的那种跨越式领先，今天在任何其他中国车企那里都还没有。我们正在非常勤奋地工作，试图在中国获得更广泛的，我是说，FSD 批准。这很难，但对某些车，我们已经能够让 FSD 运行，而且人们在那里试过的，无论谁试过，都明白它比任何竞争对手好多少。事实上，我不知道你有没有看到这个。有一些竞争对手的首席执行官来到美国这里做了 FSD 驾驶，他们自己也被 FSD 能做到的事震撼了。所以这是一个类似的答案。再说一次，我们没有掉以轻心。我们正在积极与中国的监管机构合作，因为我不知道你有没有看到，我们以前发布过 FSD 的安全报告，但我们也发布了无人监管、也就是 Robotaxi 的安全报告。我们只是认为，这是把人们带进一个安全环境的正确方式，因为，你知道，由于驾驶员不警觉而发生的事故数量，而这是我们能确保世界上伤亡更少的唯一方式。任何看过这些数据的监管者，都很难反驳它。

I mean, at the beginning, like definitely the leapfrog which we have in terms of this technology isn’t there with any of the other Chinese OEMs today. And we’re working very diligently to try and get broader, I mean, FSD approval in China. It is hard, but for certain cars we have been able to have FSD operating and whatever people have tried, whoever has tried it there, they understand how much better it is as compared to any of the competition. And in fact, I don’t know if you’ve seen this. There have been some competitor CEOs who have come and done FSD drives here in the US and they themselves have been blown away as to what FSD is capable of doing. So it is a similar answer. And again, we’re not taking it lightly. We’re actively working with the regulators in China because I don’t know if you saw this, but we did release our, we’ve released our safety report on FSD before, but we’ve also released our unsupervised or Robotaxi safety report. We just think this is the right way to get people into a safe environment because, you know, the amount of accidents which happen because of drivers not being alert and this is the only way we can ensure that there are less casualties in the world. And any regulator who looks at the data, it would be very hard for them to refute it.

很好，非常感谢。下一个问题来自 Wolfe 的 Emmanuel。Emmanuel，请取消静音。

Great, thanks so much. Our next question comes from Emmanuel from Wolfe. Emmanuel, please unmute yourself.

很好，谢谢。你好，Tim。关于 Robotaxi，你们谈到无人监管里程维持两位数的周环比增长。我很好奇，你们有没有办法为我们把这一点三角定位到 Robotaxi 业务的收入，也许还有盈利能力。如果你们维持，比如说这些里程 15% 的周环比增长，到今年年底你们大概会越过 1000 万英里。要产生毛利，你们需要多少英里？

Great, thank you. Hi, Tim. So on Robotaxi, you’ve talked about sustaining double-digit week-over-week growth in unsupervised miles. I’m curious if there’s a way you can triangulate that for us into revenue and maybe profitability of the Robotaxi business. If you sustain, let’s say 15% week-over-week growth in these miles, you’ll probably cross 10 million miles by the end of the year. How many miles would you need to generate gross profit?

我认为 Elon 在开头有一点谈到了这个：用指数数学，这会在 2027 年成为我们收入和盈利能力的一个重要部分。

I think Elon talked about this a little bit at the beginning, that using exponential math, this would become a significant part of our revenue and profitability in ’27.

是的，我不会说它像是，它不会在本季度撬动指针，但如果你说也许，我认为它大概在第二季度开始变得引人注意，然后到明年第三季度或第四季度大概会是一件大事。

Yeah, I wouldn’t say it’s like, it’s not like going to move the needle this quarter, but if you say like maybe, I think it starts to become noticeable probably in Q2 and then probably a big deal by Q3 or Q4 of next year.

明白了。然后如果我从车队规模来想，我认为你们的 Cybercab 产线产能被列为每年超过 12.5 万台。生产刚刚开始。你们预期 Cybercab 产线的爬坡曲线会是怎样？

Got it. And then if I think about it in terms of fleet size, so I think your Cybercab line capacity is listed at more than 125,000 units a year. Production just recently started. How do you expect the ramp up curve to be on the Cybercab line?

我的意思是，我先开头，然后也让 Lars 补充。但是，你知道，我要称赞工程团队。他们设计了一个产品，因为我不只和工程师谈，我也和制造的人谈。他们认为就制造而言，这是他们最容易规模化的产品之一。所以规模化不会是那么大的挑战。我认为我们正努力确保的关键是，我们拿到所有监管批准，从而可以以我们生产 Cybercab 的速度开始部署它们。再说一次，Lars，请随意补充。

I mean, I’ll kick it off, but then I’ll also let Lars chime in. But, you know, I would give kudos to the engineering team. They have designed a product because not only do I talk to the engineers, I also talk to the manufacturing guys. They think this is one of the easiest products for them to scale in terms of manufacturing. So scaling would not be that much of a challenge. I think the key thing which we’re trying to make sure is that we have all the regulatory approvals where we can start deploying the Cybercabs as fast as we can produce them. And again, Lars, feel free to chime in.

好，谢谢。我感谢这些称赞。团队在 Cybercab 的设计和制造上确实干得漂亮。就你的观点而言，我们的零件比我们造过的任何其他车辆少 50%，所以出错的东西也少 50%。我们肯定一直在把 Cybercab 产线往上冲，看我们能跑多快。我认为列出的产能更多是我们今天能够跑的一个速率，但这条车辆产线的实际产能要高得多。眼下，我们按 Robotaxi 服务需要车辆的速率在造。而且，你知道，我要说 Ashok 和我有一点像在赛跑。所以谁会保持领先？他在里程上有指数增长到来，你知道，一个月又一个月，我们也会寻求用车辆生产去跟上。

Yeah, thanks. I appreciate the accolades. The team definitely crushed it on the Cybercab design and manufacturing. Just to your point, we have 50% fewer parts than any other vehicle we’ve ever built, so that’s 50% fewer things to go wrong. We’ve definitely been bursting the Cybercab line to see how fast we can run it. I think the listed capacity is more of a rate at which we’re capable of running today, but the actual capacity of that vehicle line is significantly higher. Right now, we’re building at the rate that the Robotaxi service needs the vehicles. And, you know, I’ll say Ashok and I are at a bit of a race. So who’s going to stay ahead? He’s got exponential growth coming on the miles, you know, month over month, and we’re going to look to meet that as well with the vehicle production.

很好，谢谢各位。下一个问题来自摩根士丹利的 Andrew。Andrew，请取消静音。

Great, thanks, guys. The next question is going to come from Andrew from Morgan Stanley. Andrew, please unmute yourself.

很好。非常感谢接受这个问题。我想从你们的能源业务开始。你们显然在储能或 Megapack 这边看到大量需求，而且你们有一些目标，要上马自己的太阳能制造。但我希望能多一点了解：你们如何思考特斯拉在同时促成分布式电力、以及分布式算力上的角色。我想上个季度你们提到过一个 Megapod 产品。所以只是希望多一点了解，你们在那里设想的是什么，以及这最终对特斯拉能有多大。

Great. Thanks so much for taking the question. So I want to start on your energy business. You’re clearly seeing a lot of demand on the energy storage or Megapack side and you’ve got some goals to bring on your own solar manufacturing. But I was just hoping to get a little bit more insight in terms of how you’re thinking about Tesla’s role in enabling both distributed power, but also distributed compute. I think last quarter you alluded to a Megapod product. So just hoping to get a little bit more insight in terms of what you’re envisioning there and ultimately how big this could be for Tesla.

我认为在那条线上还早。我们不知道它会展开到什么程度。我认为我们大概会在第一季度或第二季度对它有更好的感觉，但它确实有很大潜力。只是现在很难说潜力有多大。

I think it’s still early days on that front. We don’t know the full extent of how that’ll unfold. I think we’ll have a better sense for that probably in maybe in Q1 or Q2, but it does have a lot of potential. It’s just hard to say how much potential right now.

好，明白了。然后也许就 Robotaxi 追问一下。看起来自从你们推出 Cybercab，过去一个月左右，你们在奥斯汀确实到了一个拐点。所以，当你考虑扩展到新的地区时，你会大体上说，眼下增长的瓶颈更多是监管驱动的，而且你们在软件一侧已经跨过了大部分障碍吗？我只是好奇，对接下来几个月你们计划规模化进入的那些新增地区，你大体上如何给这些增长障碍分类？

Okay, understood. And then maybe just a follow up on Robotaxi. It does seem like you’ve kind of hit an inflection point here in Austin over the last month or so since you launched the Cybercab. So would you generally say the bottlenecks at this point to growth are more regulatory driven as you think about expanding into new geographies and that you’ve kind of gotten through most of the hurdles on the software side? Just kind of curious how you generally kind of categorize some of those growth hurdles in some of the incremental regions you’re planning on scaling into over the next few months?

我认为……我的意思是，仍有某些软件挑战我们还在解决。你知道，Ashok 谈过这个，所以也许 Ashok，也许 Simon。

I think the… I mean, there are still certain software challenges which we are still working through. You know, Ashok talked about it, so maybe Ashok, maybe Simon.

好。例如，我们一开始把 Robotaxi 服务放在早上 6 点到晚上 9 点，然后扩展到晚上 10 点、晚上 11 点；现在很快就会到凌晨 2 点，所以我们在慢慢扩大 Robotaxi 的可用时间，而这是由 AI 软件的能力驱动的。我们只是想确保没有哪怕一次安全事件。那是我们的运营原则。哪怕存在最轻微的可能风险，我们也不承担那个风险。所以我们在这个意义上非常厌恶风险，也就是我们要的是零伤害，或者任何不好的事都不发生。那是驱动原则。我们会看到，在今年剩余时间里，我们将解锁 24 小时全天候运营，走上更高速度的道路，我们认为所有这些对于规模化、并向很多用户提供一个出色的产品都是必要的。所以我们既想先做广度、做多个城市，我们现在做六个城市。我们也在做深度，在一座城市里放很多车，那就是奥斯汀。我们还会扩展到更多天气条件、白天和夜晚的条件、高速公路和高速度。所有这些希望在今年年底之前完成，到那时你会看到多得多的规模化，有多得多的城市被 Robotaxi 覆盖。所以它并不纯粹是监管之类的。它是技术驱动的。但我们确实看到，它会相当快地变成监管问题，也许在第一季度、第二季度的时间框架里。

Yeah. For example, we had, like, initially had the Robotaxi service from 6 a.m. to 9 p.m., and then we expanded 10 p.m., 11 p.m.; now it’s going to be pretty soon 2 a.m., so we are slowly expanding the availability of Robotaxi and that is driven by the capability of the AI software. And we just want to make sure that there’s not a single safety incident. That is our operating principle. And even if there’s the slightest possible risk at all, we don’t take on that risk. And so we are very risk averse in that sense, as in we want zero injuries or anything bad happening. And that is the driving principle. And we will see that over the rest of this year we’ll unlock 24×7 operation, go to higher-speed roads, and we think all of that is necessary to sort of scale and offer a great product to a lot of users. And that’s why we wanted to both do the breadth first of doing multiple cities—we do six cities now. We are also doing the depth of having a lot of vehicles in a single city, which is in Austin. And we’ll expand to more weather conditions, day and night conditions, highways and high speeds. And all of that will be done hopefully by the end of this year, at which point you will see a lot more scaling, a lot more cities being covered by Robotaxi. So it’s not purely regulatory or something like that. It is technology driven. But we do see that pretty quickly become regulatory, perhaps in Q1, Q2 timeframes.

很好，谢谢各位。下一个问题来自德意志银行的 Edison。Edison，请取消静音。

Great, thanks guys. The next question is going to come from Edison, from Deutsche Bank. Edison. Please unmute yourself.

很好。谢谢接受我们的问题。首先，想问 Terafab。我知道，Elon，你提到过一个愿望目标，是 2Q28。你对要到达那里需要投入多少资本有没有感觉，以及沿途我们应该留意什么样的里程碑？

Great. Thank you for taking our questions. First, want to ask about Terafab. I know, Elon, you mentioned that you had an aspirational goal of 2Q28. Do you have a sense on how much capital needs to be invested to get there and also what kind of milestones we should be on the lookout for on the way there?

嗯，那会是相当可观的资本。你知道，对 Terafab 来说肯定是在数百亿美元。不过我们得弄清楚，SpaceX 和特斯拉之间谁为 Terafab 的什么付钱，那种相对的分担大概是怎样。但到 2028 年初之前，资本设备不会少于数百亿美元，然后它会——但长期大概会超过 1000 亿美元。是的。

Well, it’s going to be some pretty significant capital. You know, it’s certainly in the tens of billions for Terafab. Although we got to figure out, like, what the sort of relative split is between SpaceX and Tesla in terms of who pays for what at Terafab. But it’s not going to be less than tens of billions in capital equipment up through early 2028, and then it will—but it’ll probably be north of 100 billion long term. Yeah.

我只补充一句，你知道，这座工厂的第一批设备，是特斯拉在建的那座，我们实际上已经下了大部分订单，所以当 Elon 谈到明年年中时，那是来自第一座工厂，也就是研究工厂。

And I’ll just add, you know, the first set of equipment for the fab, which Tesla is building, we’ve actually already placed most of the orders, and that’s why when Elon was talking about middle of next year, that is from the first fab, which is the research fab.

是的，那就是我们正在做的，弄清楚设备怎么运转这类事。是的。但真正的生产会是 2028 年晚些时候，按晶圆厂的标准，那是闪电般地快。

Yeah, that’s what we’re—just figuring out how to work the equipment type of thing. Yeah. But real production would be late ’28, which is lightning fast by fab standards.

明白，明白。然后就 Optimus 快速追问一下。我知道你提到我们很快能看到一段视频，进展相当多，但我想你们仍在做一些打磨。笼统地说，你认为接下来大部分设计改动还会发生在哪里？仍然是手的问题吗？有什么你可以点出来的？

Understood, understood. And then just a quick follow up on Optimus. I know you mentioned that we could see a video soon, that there’s quite a bit of progress, but you’re still making some refinements, I guess. Just generally speaking, where do you see most of the design changes still happening going forward? Is it still the hand that’s an issue? Anything you can call out there?

是的，那是大部分的困难。我的意思是，Constantinos，你来回答这个问题吧？

Yeah, it’s most of the difficulty. I mean, Constantinos, why don’t you answer this question?

所以手实际上是最具挑战的硬件领域之一。机电系统的大部分，我会说，信心相对较高，意思是它已经收敛到某种工作得还不错的东西。从机电的观点看，手比运动系统的执行器更复杂，比如。另外，还有一些定制齿轮，在消耗和可靠性上承担着很大负担，我们仍在开发。

So hand is one of the most challenging hardware areas, actually. Most of the electromechanical system, I would say, is relatively high confidence, meaning that it has converged to something that works decently well. The hand is, from an electromechanical point of view, more complicated than actuators for the locomotion system, for example. Also, there are some bespoke gears that are carrying a lot of burden for consumption and for reliability that we are developing still.

是的，这会是最主要的。是的，主要是手，就是手部设计的微调，尤其是手指。这是一个难问题，因为，你知道，人类有自我修复，像我们的，你知道，我们的手会修复，我们的身体会自我修复，但机器人不会。所以你必须，而且我们的，你知道，手指有点柔软，皮肤里有很多传感器。很难做出一只手，既能承受大量磨损，又仍然柔软到能拿起精细的物体，并且有全部传感器，在你重复动作时还不损坏传感器。有很多小事会让你意识到，归根结底，人类的手是一件多么不可思议的东西。所以从机电的立场看，主要是手。整台机器人上也有其他东西在改进，但那是主要的。然后，正如我提到的，用 AI5 升级机器人的智能、机载智能，那是智能上的一大步变化，会大大帮助 Optimus 的能力。所以我以前说过，人形机器人是地球上最大的产品，而它的有用程度，粗略近似，会是机电灵巧度、AI 软件和 AI 芯片硬件的乘积。这三件事都在指数式改善，我想我们经常用这个词，而价值大致是这三件事彼此相乘。然后 Optimus 机器人会建造未来的 Optimus 机器人。所以你有一个递归函数。当你有一个递归的、相乘的指数，事情真的会增长得极其快。它们也许从一个小的基数开始，但会增长到疯狂的水平。我认为到某个时候，地球上的人形机器人会比人类更多。

Yeah, this would be the most. Yeah, it’s mostly the hand and just kind of the fine-tuning of the hand design and the fingers especially. It’s a hard problem because, you know, humans have self repair—like our, you know, our hands repair, our body repairs itself—but robots don’t. So you have to—and our, you know, fingers are kind of squishy and have lots of sensors in our skin. It’s hard to make a hand that can handle lots of wear and tear but is still squishy enough to pick up delicate objects and has all the sensors and doesn’t damage the sensors when you do repeated actions. There’s a lot of little things that you realize—actually, what an incredible thing the human hand is at the end of the day. So it’s mostly, from an electromechanical standpoint, the hand. There are other things that are improving as well throughout the robot, but that’s the main thing. And then, as I mentioned, upgrading the intelligence, the onboard intelligence of the robot with AI5, which is a big step change in intelligence that will help a lot with the capabilities of Optimus. So I’ve said before that humanoid robots—biggest product on Earth—and its usefulness will, it’s a rough approximation, be a multiple of the electromechanical dexterity, the AI software, and the AI chip hardware. And all three of those things are improving exponentially—I think we use that word a lot—and the value is roughly those three things multiplied by each other. And then the Optimus robots will build the future Optimus robot. So you have a recursive function. So when you have a recursive multiplicative exponential, things will really grow extremely fast. They may start from a small base, but they’re going to grow to crazy levels. I think at some point there will be more humanoid robots on Earth than there are humans.

很好，非常感谢。下一个问题来自美国银行的 Alex。Alex，请取消静音。

Great, thanks so much. The next question comes from Alex from Bank of America. Alex, please unmute yourself.

你好，谢谢在这里接受我们的问题。我想先说 Robotaxi。看起来在大多数市场，你们的 Robotaxi 定价低于竞争对手的网约车，也低于其他 Robotaxi 玩家。鉴于你们的一些结构性成本优势，长期用这种方式来夺取市场份额，是计划吗？

Hi, thanks for taking our questions here. I guess just first on Robotaxi. So it looks like for Robotaxi, you’re pricing lower than competitor rideshare and other Robotaxi players in most markets. Is that the plan to do longer term to take market share given some of your structural cost advantages?

是的，我的意思是，记住当你比较时，你是在把我们和网约车比较，那里面有一个司机，而我们没有，再加上我们自己制造这辆车。所以我们的整体成本基础是不同的。我认为我在开场白里说过，我再说一次，只是不想太重复，但基本上我们的目标是接管任何一种运输。

Yeah, I mean, remember when you’re comparing, you’re comparing us with rideshare, which has a driver in there and we don’t, plus we are manufacturing the car ourselves. So our cost basis overall is different. And I think I said it in my opening remarks, and I’ll say it again, just don’t want to be too repetitive, but basically our goal is to take over any kind of transportation.

我的意思是，我认为，除非我们被政府拦住，这当然是一种可能，比如如果他们试图保护本国产业，我认为在任何允许特斯拉自由运营的国家，特斯拉最终会完成地球上全部客运英里的大多数。你可以回来看这个预测，但我对这个预测有信心。

I mean, I think we, unless we are stopped by governments, which is certainly a possibility, like if they’re trying to protect the domestic industry, I think Tesla will end up doing a majority of all passenger miles on Earth in any country where it is allowed to operate freely. And you can come back and look at this prediction, but I feel confident about this prediction.

是的，在可预见的未来我们会继续规模化。我们沿途看不到障碍。显然有工作要做。但我们确实看到一条清晰的路径，在接下来几个月里规模化到数百万辆车。这之所以发生，主要是因为我们过去几年做的技术投资。而且在我们今年早些时候提到的 V15 技术栈里，所有的基础都像是很难看见的，像一座冰山。人们只看见冰山的尖。里面投入了大量工作。而现在我们开始看到那全部工作的回报，你知道，开发端到端神经网络、向人类学习，等等。那一切都在产生回报。所以你看到的，是多年工作的结果。它并不是就在过去几个月里发生的。同一套技术应当规模化到数百万辆车，遍及所有地方，也就是我们被允许运营的任何地方。

Yes, we will continue the scaling for the foreseeable future. We don’t see any hurdles along the way. Obviously there’s work required. But we do see a clear path towards scaling to millions of vehicles over the next several months. This is happening due to primarily the technology investments we have made over the last several years. And also in V15 stack that we alluded to earlier this year, like all of the foundations is like hard to see, is like an iceberg. People only see the tip of the iceberg. There’s a ton of work that goes into it. And like now we are starting to see the payoff of all that work of like, you know, developing end-to-end neural networks, learning from humans and so on. All of that is paying off. So that’s what you’re seeing, that you’re seeing the result of many years of work. It didn’t just like happen in the last several months. And the same technology should scale to millions of vehicles across all over, like wherever we are allowed to operate.

下一个问题来自 New Street 的 Pierre。Pierre，请取消静音。

Our next question is going to come from Pierre from New Street. Pierre, please unmute yourself.

谢谢接受这个问题，各位。我只是想回到这一点，我对美国 75% 的汽车带 FSD 销售非常着迷。有趣的是，如果你拿到这些车，它们在卖出的当天大概会产生大约 7000 美元的毛利。但当然 FSD 每年产生超过 1000 美元。所以在车卖出的那个季度，它对数字的影响是最小的。但在车的生命周期里，它大概至少把每辆车的利润翻倍。所以我想走到的地方是：我这样看对吗？你们会在某个时候向我们披露 FSD 所代表的年度经常性收入吗？我自己算了算，我认为它现在大概正在接近 10 亿美元的 ARR。这个数量级对吗？

Thanks for taking the question, guys. I just wanted to come back, like I’m very really fascinated by 75% of, like, US cars selling with FSD. And the interesting thing that if you get the cars, they probably generate on sale the day they sell like $7,000 of gross profit. But then of course FSD generates more than $1,000 a year. And so it makes like a minimum impact on numbers in the quarter the car is sold. But over the lifetime of the car it probably at least doubles the profit per car. And so the place where I’m going to want to get to is, am I looking at it the right way? And are you at some point going to disclose for us, like, the annual recurring revenues that FSD represents? And I did my own math and I’m thinking it’s probably now approaching a billion dollars of ARR. Is that the right order of magnitude?

是的，我认为你分享的这些数字是对的。随着它进一步规模化，我们将开始分享更多围绕 FSD 的指标，不只是收入，还有围绕它的其他关键指标。

Yeah, I think the numbers which you shared are correct. And as it scales further, we will start sharing more metrics around FSD, not just the revenue, but also other key metrics around it.

好，作为快速追问，你们是不是已经有很好的感觉，能否告诉我们留存率？比如，三个月或六个月之后，有多少人还留在月度订阅上，以及你们是否看到一定比例的司机流失？

Okay, and as a quick follow-up, do you have like a good sense already and can you tell us about, like, the retention rate? Like, how many people, like, three or six months down the line stick to the monthly subscription, and if you see a percentage of drivers falling off?

是的，我的意思是，你知道，就像我说的，我们将分享更多围绕流失率之类的指标。但我们在这个阶段不这样做的原因是，我们正在做很多事情，尤其是新的国家在加入。然后我们提供免费试用。还有一件事我也想确保说清楚：当我说在北美、交付时超过 75% 附带了它，这是指我们交付的每一辆车都带一个月的免费 FSD 订阅，而人们即使有那个免费可用，仍然在签约。

Yeah, I mean, you know, like I said, we will share more metrics around churn rates and whatnot. But the reason we are not doing it at this stage is there are a lot of things which we are doing, especially with new countries being added. Then we offer free trials. And one thing which I’d also like to make sure it’s clear: when I say in North America that more than that 75% attached to it at the time of delivery, this is like every car which we’re delivering comes with a month of free FSD subscription and still people are signing up even though they have that free available.

下一个问题来自 Oppenheimer 的 Colin。Colin，请取消静音。

Our next question is going to come from Colin at Oppenheimer. Colin, please unmute yourself.

非常感谢，各位。你知道，Elon，你能不能谈一下，把 Optimus 产量稍微爬一点、以便开始训练的策略？即使手还没有用最终的地面设计完全烤熟，看起来即使没有完善的手，你们在训练循环上、以及让一些仿真技术成熟上，仍然可以做很多事。

Thanks so much, guys. You know, Elon, can you talk about the strategy around ramping Optimus production just a little bit to start training? Even if the hands aren’t fully baked with the floor final design, it seems like there’s a lot that you could end up doing in terms of the training cycle and maturing some of the simulation technology even without perfected hands.

嗯，我们正在做的就是这个：我们在造还没有完善的手的机器人。我们正在 Optimus 训练上取得进展，并理顺供应链和生产系统的所有方面。我们在并行做所有这些事，比如我们弗里蒙特工厂里这条巨大的产线、庞大的供应链，然后在微调设计和工程的同时做 AI 软件。所以这些事同时在发生，我们在这里以人力所能的最快速度前进。我不认为任何人会有，事实上，如果任何人有一台在我们开始向客户交付时还能与 Optimus 勉强竞争的机器人，我会感到震惊，而那很可能是在明年年底前后。

Well, that’s what we’re doing is we’re building robots that do not have perfected hands. And we are making progress with Optimus training and ironing out all aspects of the supply chain and the production system. We’re doing all these things in parallel, like this giant production line at our Fremont factory, massive supply chain, and then fine-tuning the design and engineering simultaneously with the AI software. So all these things are happening simultaneously and we’re going as fast as humanly possible here. I don’t think anyone’s going to have—and, in fact, I’d be stunned if anyone has—a robot that is remotely competitive with Optimus when we start delivering it to customers, which is likely to be towards the end of next year.

是的，你应该知道，不只是弗里蒙特工厂，我们也已经开始在奥斯汀这里建造一座工厂，它能做的机器人会比弗里蒙特那座多得多。所以，回到 Elon 说的，我们在做我们能做的一切。我们没有停，我们只是以我们所能的最快节奏前进。

Yeah, and you should know that not only the Fremont factory, we’ve also started constructing a factory here in Austin which would be able to do a lot more robots as compared to the Fremont one. So, going back to what Elon was saying, we’re doing whatever we can. We’re not stopping, we’re just going at the fastest pace as we can.

是的，我们绝对是，就像我在拉斯维加斯说的，押在还没来的牌上。

Yeah, we’re definitely, as I say in Vegas, betting on the come.

太棒了。然后，你知道，就半导体设备而言，你知道，产能的可得性。你能不能谈谈你们如何与供应商应对这件事？你知道，显然看到一些设备上放下了定金，交付在往前走，这对关税出去是净正面的。但有没有一些更积极的事，你们现在就可以做，以便开始为更大的爬坡做准备？我们应该如何考虑这流入资本开支预算？

Awesome. And then, you know, in terms of the semicaps, you know, capacity availability. Can you talk about how you’re navigating that with the suppliers? You know, obviously seeing some deposits get put down on equipment and getting deliveries moving forward is a net positive on tariff out. But are there some more aggressive things that you can do to start getting ready for the larger ramp now? And how should we think about that flowing through the CAPEX budget?

眼下影响并不巨大。我认为资本开支实际上最终会比我们预期的略低一些，但明年会开始变得更有意义，然后在 2028 年变得引人注意，比如在数百亿美元。

It’s not a massive impact right now. I think capex is actually ending up being a little lower than we anticipated, but will start to become more meaningful next year and then noticeable, like in the tens of billions, in ’28.

好的。我们这里时间快到了。我们感谢每个人的问题，并期待下个季度再和你们谈。非常感谢，再见。

All right. We are coming up on time here. We appreciate everyone’s questions and we look forward to talking to you next quarter. Thank you very much and goodbye.