Tesla Investors: Get Ready Now - What Was Missed in the Record Q3 Report ⚡️

This really was a throwback and historic Tesla quarter, one in which Tesla exceeds basically all expectations across multiple business lines and proves that Tesla is unmatched when it comes to continually reducing costs while continuing to grow the business and delivering more value to the customer. Rather than talking in platitudes though, I want to dive right into breaking down some of the data. We'll start with autonomy because that's still the most important part of this story for the next few years. Elon and Ashok made some comments about improvements to FSD when it comes to miles per intervention or MPI. Honestly, I think it's fair to not put too much weight into these forecasts because it is pretty arbitrary from our point of view and Elon is notorious for being too optimistic with FSD.

这确实是特斯拉的一个复古且具有历史意义的季度，特斯拉在多个业务线上超出了所有预期，证明了在持续降低成本的同时，不断发展业务并为客户提供更多价值方面，特斯拉是无与伦比的。然而，与其泛泛而谈，我更希望深入分析一些数据。我们将从自动驾驶开始，因为在未来几年里，这仍然是故事的最重要部分。埃隆和阿肖克对FSD在每英里干预次数（MPI）方面的改进发表了一些评论。坦率地说，我认为对这些预测不必过于看重，因为从我们的角度来看，这有些随意，而且埃隆以对FSD过于乐观而闻名。

But in this case, what Ashok said may give us some great insight. He said Tesla has already made a 100x improvement with FSD version 12.5 when comparing to where Tesla was at the beginning of this year. At that time, most of the fleet was still on version 11, but 12.1.2 was the first version to begin rolling out to limited customers at the end of January. So if you can recall back to your drives around that time, Tesla has seen with its private internal MPI data a 100x improvement. I think this is a critical distinction because people hear these comments on calls and probably think oh wow FSD is 100 times better than it was earlier this year. And I think we would all agree it's made good progress, but it's not like the experience is 100 times better, it's just that whatever data Tesla is tracking internally, that number is 100 times better. It's worth trying to grasp this because Ashok then also said that with version 13, they're expecting a 1000x improvement from January this year on the production release software.

但在这种情况下，阿肖克所说的可能给我们带来重要的启示。他提到，特斯拉在FSD（全自动驾驶）12.5版本上，已经相较于今年年初取得了100倍的提升。当时，大多数车队仍使用11版本，但12.1.2是第一个在一月底开始有限投入客户使用的版本。所以如果你能回想起那段时间的驾驶体验，特斯拉通过其内部的MPI数据，见证了100倍的进步。我认为这是一个重要的区别，因为人们在电话会议上听到这些评论时，可能会觉得，哇，全自动驾驶比今年早些时候好了100倍。而我认为我们都会同意，它确实有显著进步，但并不是体验上提高了100倍，而是特斯拉所跟踪的某项内部数据提高了100倍。理解这点是值得的，因为阿肖克还表示，版本13预计将比今年一月的生产版软件提高1000倍。

Elon confirmed the same thing saying at least three orders of magnitude improvement for this year as a whole, which is really just another way of saying a 1000x improvement. And these statements together would imply a rough 10x improvement from where Tesla is today with version 12.5. Again, that's because version 12.5 was a 100x improvement from earlier this year. But on that point, Elon said he's expecting a five or sixfold improvement in MPI relative to 12.5 with version 13. So again, not that the experience is necessarily going to feel five to 10 times better, but Tesla's MPI data, whatever it really is, will be five to 10 times better with version 13, compared to 12.5, which yes, is a notable increase, but it's less than the increase Tesla has seen year to date.

埃隆证实了这一点，他说今年整体上至少会有三个数量级的提升，也就是1000倍的提升。这些说法表明，从目前特斯拉的12.5版本开始，大约会有10倍的改进。之所以这样说，是因为12.5版本比今年早些时候有了100倍的提升。关于这一点，埃隆表示，相较于12.5版本，他预计13版本的MPI（每英里干预）将会有5到6倍的提升。所以，并不是说用户体验一定会提升5到10倍，而是特斯拉的MPI数据（不管具体是什么）在13版本中相比12.5版本会提升5到10倍。是的，这是一个显著的提升，但低于特斯拉今年迄今为止的提升幅度。

I think reading between the lines is important on these calls. This time around, Elon was eager to talk, he was quite chipper and actually interrupted his teammates multiple times. But an engaged, confident, upbeat Elon is a great indication of where things are headed. I think if there was one chart to explain the team's confidence with FSD going forward and starting driverless rides next year, it's this one. I plotted each FSD version 12 release to align with the amount of compute that Tesla had installed at the time. Keep in mind, when a version is released, that means it's already been training for a few weeks or months prior to the release. So the current level of compute for each release is less relevant than the amount of compute in the months leading up to that release.

我认为在这些电话会议中“言外之意”很重要。这次，埃隆非常愿意发言，他十分愉快，并多次打断了他的团队成员。但埃隆的这种投入、自信和乐观，是未来发展方向的一个良好信号。我认为，如果有一张图表可以解释团队对未来FSD（全自动驾驶）及明年开始无人驾驶服务的信心，那就是这一张。我绘制了每个FSD版本12发布的图表，与当时特斯拉安装的计算能力进行对比。需要注意的是，当一个版本发布时，意味着它已经在发布前经过了数周或数月的训练。因此，每次发布时的计算能力水平不如发布前几个月的计算能力更为重要。

One key takeaway is that it likely won't be until a dot release of version 13, if not even version 14, where Tesla will actually be utilizing all of its future planned capacity. I think we'd all agree, version 12.3 was a big jump from version 11, which makes sense because in the months leading up to that release, Tesla had significantly more training capacity online. Version 12.5 has been a minor improvement for some, but again, this makes sense given the training capacity was flatlined for the time between 12.3 and 12.5. So those improvements may have had more to do with training optimization and weightings than extra compute capacity. So it's really not hard to see why the team so excited about version 13, given that the timing of this release is likely to be trained on an extra 30,000 or so H100 equivalents relative to what 12.5 was trained with. So all of this combined is why the team is so confident that in quarter two or Q3 of next year, Tesla's MPI data will surpass that of the human.

一个重要的观点是，特斯拉可能要到13.x版本（dot release 是指小版本更新）甚至是14版本才会真正运用其未来计划的全部产能。我们都能同意，12.3版本与11版本相比是一次大跃进。因为在这次发布前的几个月里，特斯拉上线了更多的训练容量。12.5版本对某些人来说是一些小的改进，但这也容易理解，因为从12.3到12.5之间的训练容量没有增加。因此，这些改进可能更多是由于训练优化和加权，而不是额外的计算容量。所以，很容易理解为什么团队对13版本如此兴奋，因为预计这次发布的训练容量相比12.5版本增加了大约3万个H100等效体。所有这些因素加在一起，就是为什么团队如此自信在明年第二季度或第三季度，特斯拉的MPI数据将超过人类的原因。

Exactly what this means is still somewhat of a mystery, but if Tesla can deploy paid driverless rides in Texas and California, the perception of the company will undoubtedly change because Tesla's path to scaling to other states will only be limited by regulations given Tesla is producing many times more autonomous vehicles per week compared to Waymo's entire fleet. And sorry about Tuesday, I spent all day researching and then two hours recording, only to realize my audio recorder was off and then I closed the tabs and all of my notes in reader mode were also gone. But an important part of what a Palo Alto council member said about the possible partnership with Tesla deploying cybercabs for ride hailing was that Tesla may be able to use the license or the permit that the Palo Alto link service already has, which could fast track things for Tesla. There's still a lot of uncertainty here, but the fact that we're hearing about specific cities and specific licensing is very encouraging.

这到底意味着什么仍然有些神秘，但是如果特斯拉能够在德克萨斯州和加利福尼亚州推出收费的无人驾驶服务，公司在公众心目中的形象无疑将会改变。因为特斯拉每周生产的自动驾驶汽车数量是Waymo整个车队的好几倍，所以特斯拉在其他州扩展业务的唯一限制将是法规。而抱歉关于周二，我整天都在做研究，然后花了两个小时录音，结果发现录音机关了，随后关闭了浏览器标签，所有阅读模式下的笔记也都丢失了。不过，重要的是帕洛阿尔托市的一名议员提到，关于与特斯拉合作推出网络出租车服务的可能性，特斯拉可能能够使用帕洛阿尔托链接服务已经拥有的许可或执照，这可能会让特斯拉的进展加快。虽然仍存在很多不确定性，但我们能听到关于具体城市和许可的消息是非常令人鼓舞的。

By the way, that Palo Alto link only goes intersection to intersection, not destination to destination. And it only has about 10 vehicles in the fleet, four of which I believe are Teslas. But the costs are too high and the funding is running out for that program. We also learned Tesla is already operating the ride hail network in the bay area for employees with a safety driver and have been testing that for about one year. So Tesla's already using and refining the ride hail app. I do think a quick glimpse of this would have been great to share on 10 10 but less than 12 months from now and we may get to see the real deal with the public. The last few things on autonomy, Elon said the take rate for FSD has improved substantially after the 10 10 event. Tesla's never going to become a company focused on advertising, but hopefully this serves as an incentive for them to get the message out there more about what Tesla vehicles are capable of.

顺便说一下，那个关于帕洛阿尔托的链接只覆盖从一个路口到另一个路口，而不是从起点到目的地。而且车队大约只有 10 辆车，其中 4 辆应该是特斯拉。但该项目的成本太高，资金即将耗尽。我们还了解到，特斯拉已经在湾区为员工运营一个带安全司机的打车网络，并且进行了大约一年的测试。因此，特斯拉已经在使用和完善打车应用。我认为快速了解这一情况本应该在 10 月 10 日分享会很不错，但或许不到 12 个月的时间，公众就能看到真正的情况。关于自动驾驶的最后一些信息，埃隆表示，10 月 10 日活动后，全自动驾驶（FSD）的接纳率显著提高。特斯拉永远不会成为一个专注于广告的公司，但希望这可以激励他们更多地传播关于特斯拉车辆性能的信息。

Elon confirmed if a hardware 3 vehicle turns out not to be powerful enough for unsupervised FSD, Tesla will provide a free upgrade computer replacement. Tesla doesn't yet know if they can push version 3 far enough to get it there, but they're definitely going to try to avoid that expense. Hopefully this calms those fears for hardware 3 owners, although I'd still expect that upgrade to be years away. A big distinction, Elon said the cybercab should hit volume production in 2026, not just that it'll start production in 2026. That's significantly more bullish than what he said at the Wii robot event. The team is clearly pumped about the new manufacturing methods for the cybercab and the much faster cycle times, which could be around 5 times better than a normal auto factory. Volume production for Tesla is usually around 5,000 units per week, so if Tesla can pull that off for the cybercab, that would be a run rate of 250,000 units in 2026.

埃隆证实，如果硬件3的车辆在无人监督的全自动驾驶中表现不够强大，特斯拉将免费提供电脑升级替换。特斯拉还不确定是否能把硬件3推到那个水平，但他们肯定会尽量避免这笔开支。希望这能缓解硬件3车主的担忧，不过我预计这种升级可能还需几年时间。一个重要的区别是，埃隆表示网络出租车应该在2026年实现量产，而不仅仅是开始生产。这比他在Wii机器人活动上的说法要乐观得多。团队显然对网络出租车的新制造方法和更快的生产周期感到兴奋，这可能比普通汽车工厂快5倍。特斯拉通常的量产水平是每周大约5000辆，因此如果能把网络出租车做到这一点，2026年的年生产速度将达到25万辆。

Not that they produced that many in 2026, but they could exit 2026 having hit that run rate, and eventually they plan to do 2 million cybercabs per year, up to 4 million per year. I'm also very thankful Elon and the team made it abundantly clear, the cybercab will not ever have a wheel or pedals. They said that would be pointless, silly, and at odds with what Tesla believes. We said this on the channel after the 1010 event, because both Lars and Franz said the same thing, but plenty of people in the community did not listen, but now hopefully that argument and line of thinking should be dead. On the semi, we got official confirmation that all of the semis already produced do have the hardware for FSD, and they're currently training with that fleet. Which is a statement that's easy to gloss over with everything else going on, but this is obviously a massive deal given the economics of the Tesla Semi and transporting goods.

在翻译2026年期间生产的数量可能不是那么多，但他们可能在2026年底达到了这种生产能力，并计划最终每年生产200万辆网络出租车，最高可达400万辆。我非常感谢埃隆和他的团队明确表示，网络出租车永远不会有方向盘或踏板。他们说这样做是毫无意义的，荒谬的，并且与特斯拉的理念相悖。在1010活动之后，我们在频道上也说过，因为拉尔斯和弗朗兹也表达了相同的看法，但社区中还是有不少人没听进去，希望这种争论和想法现在可以终结。关于半挂卡车，我们得到了官方确认，所有已经生产的半挂卡车都配备了自动驾驶硬件，目前正在用这批车队进行训练。虽然在其他事情的关注下，这个声明不太被重视，但考虑到特斯拉半挂卡车在经济和货物运输中的意义，这显然是一件大事。

It really should not be hard to see how a Tesla Semi with unsupervised FSD is going to totally blow up the goods transportation game. And sure, that's likely still a few years down the line, but that's absolutely where this is all headed. And just a sidebar, if you want to know why people often talk past each other today, this is it. I found this one article that over 400 sources reported on. 40% of that reporting came from the left and only 24% came from the right, but compare the headlines. On the left, one outlet framed the indictment by arguing the Supreme Court will unjustly give Trump immunity and on the right, Republican Senator Rick Scott called the indictment election interference. So depending on which article people see first, their view of the story will be completely different. And how is seeing all of that data possible?

翻译成中文如下： 理解如果一辆配备无人监督FSD（全自动驾驶功能）的特斯拉半挂卡车，是如何在货物运输领域大展拳脚的，这真的不应该很难。虽然这种情况可能还需要几年才能实现，但未来的方向显然是如此。顺便提一句，如果你想知道为什么现在人们经常鸡同鸭讲，这就是原因所在。我找到了一篇文章，有超过400个来源对此进行了报道。其中，40%的报道来自左派，只有24%来自右派，但比较一下标题。在左派中，有个媒体将起诉描绘为认为最高法院将不公正地给予特朗普豁免，而在右派中，共和党参议员里克·斯科特称该起诉是选举干预。因此，取决于人们首先看到的文章，他们对事件的看法将完全不同。那么，如何看到所有这些数据呢？

If you guys need a gift idea with the holidays around the corner, a subscription to the ground news vantage plan would be an excellent option for anybody that likes to stay informed without media bias. Ground news does sponsor the channel, but with the election right around the corner, this has become an invaluable tool for me. Ground news does not use these manipulative algorithms and they're affiliated with zero corporate media or tech companies. The my news bias feature gives you a real time snapshot into all of the different articles that you're reading. For every article you read, you get these tags telling you who actually owns that source, the factuality level of that source, and again, the political leaning. And with one click, it'll take you direct to that source. Plus, the blind spot feed will show you what stories are being disproportionately reported by either side.

如果你们正在为即将到来的节日寻找礼物灵感，那么订阅Ground News的Vantage计划是一个非常棒的选择，特别适合那些希望获取不带偏见新闻的人。虽然Ground News是赞助了我们的频道，但随着选举的临近，这已成为我非常宝贵的工具。Ground News不使用那些操控性的算法，也不与任何企业媒体或科技公司挂钩。它的“我的新闻偏见”功能让你实时了解所阅读文章的不同视角。每篇文章都有标签，告诉你该来源的所有者、可靠性程度，以及它的政治倾向。而且，只需点击一下，就可以直接访问该来源。此外，“盲点报道”功能会显示在不同立场中哪些故事被不成比例地报道。

I know a lot of you have already signed up for the vantage plan, which gives you unlimited access. So if you'd like to support a company doing what I believe is very important work in this day and age, you can get 40% off the vantage plan at ground.news slash electrified, or you can use the QR code right on the screen. Real quick on the more affordable vehicles. Yes, plural. Elon said more affordable models and lower cost vehicles multiple times. But before everybody gets giddy here, the other comments the team made have it sounding like it may be more of a lower priced entry level model three or why a more limited variant of those vehicles more so than an entirely new vehicle.

我知道很多人已经注册了Vantage计划，这一计划可以让你享受无限访问。如果你想支持一家在当今时代做着非常重要工作的公司，你可以在ground.news/electrified上获得Vantage计划的60%折扣，或者使用屏幕上的二维码快速获取。 关于更实惠的车辆，注意是复数。埃隆多次提到更实惠的车型和低成本车辆。但在大家兴奋之前，团队的其他评论听起来这可能更像是价格较低的入门级Model 3或Model Y的一种更有限的变体，而不是一款全新的车型。

The word is at least some version in the new model lineup will be under $30,000, but that's with incentives. So these new models may still start above $35,000 and I'm not expecting anything close to a starting price of $25,000. That's the cybercab and that will use a fully unboxed process on separate manufacturing lines. These new models next year are on existing lines. One more reason to think is just new variants of the three and why plus we heard Elon rant about how heroic it is to reduce 20% of the costs from an existing vehicle, which to me is clearly what they've been trying to do for this new affordable variant reduced 20% of the costs from the three and why the cheapest model three is about $42,500 minus the $7,500 credit takes it to $35,000 and then 80% of that would be $28,000.

消息透露，新款车型阵容中，至少有一些版本的价格将在30,000美元以下，但这需要考虑激励措施。因此，这些新车型的起步价格可能仍高于35,000美元，我也不认为会有接近$25,000的起步价。那是指Cybercab，这款车将采用全新的无框工艺，并在单独的生产线上制造。而明年的新车型将在线上现有生产线上制造。还有一个原因可以认为这些只是三款和Y型的新变种，因为我们听到埃隆曾谈到在现有车辆成本上削减20%的难度。在我看来，他们明显在努力为这一新型实惠车型做到这一点，即将三款和Y型的成本降低20%。最便宜的三型车约为42,500美元，减去$7,500的信贷，再将剩下的80%计算出来，得出28,000美元。

So if Tesla reduced 20% of the costs from the base model three, you can see how with the same structure they can sell a new base variant under $30,000 after incentives and keep a similar margin profile. So I know it's not as exciting as an entirely new product, but it'll certainly increase the total addressable market for Tesla and should serve as a perfect bridge until the cybercab is ready for production the following year. Don't forget, the model why was unveiled in March 2019 and then entered production in January of 2020.

因此，如果特斯拉将基础款Model 3的成本降低20%，你可以看到，他们可以在保持相同结构的情况下，通过激励措施将新的基础款车型售价降到3万美元以下，同时保持类似的利润水平。我知道这不像推出全新产品那样令人兴奋，但这无疑会扩大特斯拉的潜在市场规模，并且应该能成为一个很好的过渡方案，直到"Cybercab"在次年准备好投入生产。别忘了，Model Y是在2019年3月发布的，并于2020年1月开始生产。

That's about a 10 month timeline and we're already only 8 months away from June of next year. So that's just one more anecdote leading me to believe this will be more of a new entry level type variant than an entirely new product. Hopping to the financials is just a few takeaways here. Of course the Cybertruck hitting gross margin profitability for the first time is a great sign. We'll see if they can sustain this offering the $80,000 variant now into quarter four, but to hit this metric after customer deliveries only began November of last year. So about 11 months ago, with so many new manufacturing techniques and so much of the electrical architecture being new, which requires a new supply chain, I think that's incredibly impressive and means great things for Cybertruck profitability long term.

这是一个大约10个月的时间安排，而距离明年6月我们只有8个月的时间。因此，这只是另一个让我相信这更像是一个新的入门级变体，而不是一个全新的产品的实例。来看一下财务方面，这里有几个要点。当然，Cybertruck首次达到了毛利率的盈利，这是一个很好的迹象。我们将看看他们是否能在第四季度维持这一水平，提供80000美元的车型。交付仅从去年11月开始，也就是大约11个月前，再加上许多新的制造技术和许多全新的电气架构，这需要一个新的供应链，我认为这非常令人印象深刻，并且对Cybertruck的长期盈利能力意味着很大的发展前景。

The auto gross margin X credit figure came in at 17.1% well ahead of expectations. And there was some talk about this being inflated thanks to Tesla recognizing $326 million in the quarter for deploying FSD to the Cybertruck and smart summon to the rest of the fleet. And let's assume all of this revenue flows through as profit. The good news is even if you take this number out entirely from Tesla's automotive gross profit, which I don't think is a rational argument, but to make it anyway, the auto gross margin X credit figure would still have been 15.4%, which would be up from 14.6% in quarter three. In the 10Q that Tesla already released this morning, they recognized $711 million of deferred revenue for the first three quarters of this year. That means $385 million was recognized across quarter one and quarter two, which is an average of $192 million per quarter.

汽车的毛利率剔除补贴后的数据为17.1%，远超预期。有些人认为，这一数字可能由于特斯拉在这个季度计算了3.26亿美元的收入而有所膨胀，这些收入来自于对Cybertruck部署全自动驾驶（FSD）功能以及智能召唤功能推广到其他车队。假设所有这些收入都转化为利润，好消息是，即使把这部分收入完全从特斯拉的汽车毛利润中去除，汽车的毛利率剔除补贴后仍将达到15.4%，高于第三季度的14.6%。特斯拉今天早上发布的10Q报告中显示，今年前三季度确认了7.11亿美元的递延收入。这意味着，第一和第二季度确认的收入为3.85亿美元，平均每季度为1.92亿美元。

So this number will certainly fluctuate and the quarter three recognition was materially higher than quarter one and quarter two. But if we just normalized this number, recognizing the average from quarter one and quarter two of $192 million instead of the 326 million actually recognized the auto gross margin X credits would have been 16.4% and the Wall Street expectation was 14.9%. So basically what I'm trying to say is no matter how you slice this, this was a great number. Going forward, Vibov did say it's going to be tough to maintain these margins in quarter four given macro challenges. So we should not be expecting a significant increase here despite the expectations for Tesla having its best delivery quarter ever this quarter four. As we highlighted, it's likely the deferred revenue recognition comes back down to normalized levels in quarter four as well.

因此，这个数字肯定会有波动，第三季度的确认收入明显高于第一季度和第二季度。但是，如果我们将这个数字标准化，以第二季度和第一季度的平均值1.92亿美元来替代实际确认的3.26亿美元，那么扣除积分后的汽车毛利率将达到16.4%，而华尔街的预期为14.9%。所以我想说的是，无论怎么分析，这个数字都很出色。展望未来，Vibov表示，由于宏观挑战，第四季度维持这些利润率将很困难。因此，尽管有期望第四季度特斯拉将迎来有史以来最佳交付季度，我们不应该期待利润率有显著提升。如我们所强调的，推迟确认的收入很可能在第四季度回落到正常水平。

One metric everyone should track because it's the one that Tesla has said they're optimizing for is operating margin. That came in at 10.8%. The highest since quarter one of 2023 and up from the 6.3% in quarter two this year. If we normalize that a bit and add back the $55 million in restructuring charges, that number jumps to 11%. So it looks like Tesla has already taken the brunt of the hit from the mass layoffs and has moved into the phase where it's beginning to realize the benefits of lower personnel operating expenses. The highest this number has ever been was quarter one of 2022 when it hit 19.2%. It'll likely take a few years to approach that figure again even with Tesla energy achieving record margins of 30.5% in quarter three.

每个人都应该关注一个指标，因为这是特斯拉表示正在优化的：营业利润率。这个季度达到10.8%，是自2023年第一季度以来的最高水平，比今年第二季度的6.3%有所上升。如果我们对这些数据进行一些调整，加回5500万美元的重组费用，营业利润率会增加到11%。这表明特斯拉已经承受了大规模裁员带来的冲击，并进入了开始享受较低人员运营开支好处的阶段。历史最高记录是在2022年第一季度，达到19.2%。即使特斯拉能源在第三季度实现了30.5%的创纪录利润率，可能还需要几年时间才能再次接近那个数字。

As we've talked about, a lot of this margin does flow directly to operating profit in part because Tesla energy has lower operating expenses relative to automotive. A fun fact, the highest gross margin Tesla ever achieved with the auto business was 32.9% in quarter one of 2022. And the highest X regulatory credits was the same quarter at 30%. So X credits, this Tesla energy gross margin is higher than Tesla ever achieved on the auto side without regulatory credits. This will fluctuate as deployments and recognition of inventory is volatile with the timing but over time Tesla called out they expect margins to continually improve for Tesla energy. The margin may dip a bit as well when the megapack factory in Shanghai begins to ramp and some of those costs are recognized but they should be able to ramp output there even faster than they did in Lathrop helping to offset that.

如我们所讨论的，许多利润率直接转化为营业利润，部分原因是特斯拉能源的运营费用相对于汽车业务较低。有趣的是，2022年第一季度，特斯拉在汽车业务上获得的最高毛利率是32.9%；扣除法规信用的毛利率也是在同一季度，达到了30%。所以，在不考虑法规信用的情况下，特斯拉能源的毛利率高于特斯拉在汽车业务上不包括法规信用时的最高毛利率。由于安装和库存确认的时间波动，这一毛利率会有所波动，但特斯拉表示，从长远来看，他们预计特斯拉能源的毛利率会不断提高。当上海的Megapack工厂开始增加产量时，毛利率可能会略有下降，因为一些成本将被确认，但他们应该能够比Lathrop工厂更快地提升产量，从而抵消这些成本。

But with record cash on hand, a new record low cost per vehicle, regulatory credit revenues continuing to come in at an increased pace already over $2 billion for this year. And looking at this chart from the lovely Alexandra, this figure is still growing annually and not shrinking like so many people predicted years ago. We have auto gross margins X credits going back up even after adjusting for the deferred revenue recognition. Operating margin is the highest it's been in 6 quarters and free cash flow was the highest it's been in two years. So clearly Tesla's in a very strong financial position given they implied even greater deliveries in quarter four. On that note, they're expecting slight volume growth over last year which was 1,808,581 units in 2023. So far this year, Tesla's delivered 1,293,656 vehicles.

但目前特斯拉的现金储备创历史新高，单车成本创历史新低，且监管积分收入继续快速增长，今年已经超过20亿美元。从Alexandra的这张图表来看，这一数字仍在逐年增长，而不是像多年前很多人预测的那样缩减。即使调整了递延收入确认，扣除积分后的汽车毛利率也在回升。运营利润率是最近6个季度以来的最高水平，自由现金流也是两年来的最高点。因此，显然特斯拉在财务上处于非常强劲的状态，因为他们暗示第四季度的交付量会更大。就此而言，他们预计与2023年相比，车辆交付会有略微增长，2023年的交付量为1,808,581辆。而到目前为止，今年特斯拉已经交付了1,293,656辆。

Which means Tesla would need to deliver 514,926 units in quarter four to sell one more vehicle than last year which could be defined as slight growth. If they pull that off, that would be the first time in Tesla's history delivering over 500,000 units in a quarter. The current high watermark is 484,507 which was in quarter four of 2023. Assuming Tesla does achieve slight volume growth this year, the top of Elon sales growth forecast for next year is 30% so the bullish delivery guess for 2025 is 2,351,155 units. Even if we use the 20% growth on the lower end of his prediction, that's still 2,170,298. Significantly ahead of the Wall Street consensus which will likely be adjusted up in the weeks ahead, pushing earnings per share expectations higher in the process. So all of this given that we're still technically in this in-between growth wave phase, this is about as good of a quarter as you can ask for as a Tesla investor at this point.

这意味着，特斯拉需要在第四季度交付514,926辆汽车，才比去年多卖一辆车，这可以被定义为轻微增长。如果他们成功做到这一点，这将是特斯拉历史上第一次在一个季度内交付超过50万辆汽车。目前的最高纪录是2023年第四季度的484,507辆。假设特斯拉今年确实实现了轻微的销量增长，埃隆·马斯克明年的销售增长预测最高为30%，因此2025年乐观的交付预测是2,351,155辆。即使采用他预测范围的低端20%的增长，仍然是2,170,298辆，明显领先于华尔街的共识，华尔街可能会在接下来的几周内上调预测，从而推高每股收益预期。因此，虽然我们仍处于增长波动的过渡期，但从目前来看，作为特斯拉投资者，这是你能期望的一个相当不错的季度成绩。

As Elon said, Tesla's basically winning at everything it's doing. The only not so great update was the sentiment around the Roadster and how the design is not finalized yet. And while that will be a really fun product and it'll generate some buzz, it was never going to move the needle for the stock or for profits. So the timeline for that is still pretty unclear. But the 4680 sale now has a clear path to being the most competitive sale from a dollars per kilowatt hour standpoint in North America in the near future, which is incredibly astounding for a car company to do in-house in just a few years. Hopefully this report will serve as a wake-up call to the loud cohort that has grown impatient and frustrated with Tesla. There really is just no way legacy auto can compete with Tesla given their decades of structural inefficiencies trying to compete in this new AI and software driven electric future.

正如埃隆所说，特斯拉在它所做的一切上基本上都在获胜。唯一不太令人满意的消息是关于Roadster车型，其设计尚未最终确定。虽然这款车将会非常有趣，并能引发一些关注，但它从来不会对股票或利润产生实质性影响。所以其时间表仍不明朗。然而，4680电池销售如今有望在不久的将来成为北美每千瓦小时成本最具竞争力的产品，这对于一家汽车公司在短短几年内自主完成这一壮举而言，实在是令人惊叹。希望这份报告能够唤醒那些对特斯拉变得不耐烦和沮丧的人。传统汽车公司由于几十年来的结构性效率低下，在这新时代的人工智能和软件驱动的电动汽车竞争中，实在无法与特斯拉匹敌。

We may have another few quarters where the results aren't groundbreaking, but it's becoming clear as day that within the next 12 to 24 months, this may start feeling like the good old days of Tesla back in 2020. Not even for the stock, per se, as that can be weighed down by macro factors and other things, but I'm talking about Tesla printing incredible gross margin numbers on the foundation of elite continual cost reduction. Tesla's approaching a trillion dollar company status and it still runs like a startup. This really just can't be touched by any auto company and even for other tech companies, Tesla is in a league of its own when it comes to solving real world AI with some of the largest supercomputer buildouts in the world.

我们可能还会有几个季度的业绩不太出色，但可以非常明显地看到，在接下来的12到24个月内，特斯拉可能会重现2020年那种辉煌的感觉。这并不仅仅是说股票，因为股票可能会受到宏观因素和其他因素的影响。我指的是特斯拉凭借持续的卓越成本降低，取得了令人惊叹的毛利率数据。特斯拉正接近成为万亿美元市值的公司，但它仍然像一家初创公司一样运作。这一点是其他任何汽车公司无法比拟的。即使是在其他科技公司中，特斯拉在解决真实世界的人工智能问题时，也属于独一无二的领域，因为它建造了世界上最大的一些超算。

There's a reason that today, Tesla stock was up over 21% up $46.83 per share, which adds around $150 billion in market cap in one day. This was a big sentiment shifter for Tesla, the company and Tesla, and this really is just heading toward the exit of the no growth in between phase. It sounds like we'll be back to delivery growth next year, Tesla energy is crushing it, Cortex is continually adding capacity, and FSD may make some very significant improvements with version 13 and dot releases thereafter, which is expected to allow those driverless rides next year. Optimists can always surprise as well, but even without anything from Optimus for a few years, as Vibob said, the future is looking incredibly bright for Tesla.

今天，特斯拉股票上涨了超过21%，每股上涨了46.83美元，使其市值在一天内增加了约1500亿美元。这对特斯拉公司是一个重大的情绪转变，也是走出“无增长阶段”的标志。看起来，明年我们将重新迎来交付增长，特斯拉能源表现出色，Cortex持续增加产能，而FSD（全自动驾驶）在推出第13版及其后续更新后，可能会有非常显著的改进，预计明年将实现无人驾驶乘车体验。乐观主义者总是能带来惊喜，即使未来几年的产品发展没有来自Optimus的贡献，正如Vibob所说，特斯拉的未来仍然显得非常光明。

I know some people out there sold shares ahead of earnings given the negative sentiment hoping to buy back at a lower price, but there's a reason we've said for years, you have to be careful because in the short term anything can happen, and Tesla has a history of outperforming expectations. The problem is our culture has a what have you done for me lately mentality nowadays, so too many people were focused on the past few quarters alone and they were lulled to sleep. But Tesla has a way of reasserting its dominance, and this is really just a sneak peak of the sentiment that will arise when Tesla does become the most valuable company in the world. And I don't know how long it's going to take, but from everything I can see, we're still squarely on track for that.

我知道有些人因为市场情绪不佳，在特斯拉财报发布前抛售股票，希望能以更低的价格买回来。但多年来我们一直提醒大家需要谨慎，因为短期内任何事情都有可能发生，而特斯拉过去常常超出预期。问题在于我们的文化如今更注重眼前的回报，所以很多人只关注过去几个季度的表现，因而掉以轻心。然而，特斯拉总能再次证明它的实力。现在，我们只不过是提前看到了一点特斯拉成为全球市值最高公司时会带来的市场情绪。我不知道这需要多久，但从我观察到的一切来看，我们还在实现这一目标的道路上。

It was a great quarter from Tesla, and truly there are plenty of things to be very excited about in the next 12 months, which yes, is part of why Tesla stock had a historic trading day. I'm still working to confirm this, but previously the biggest dollar move for Tesla stock in one day was $47.67 back on January 3rd 2022. So adjusting for the split, this may have just become Tesla's second biggest trading day ever when it comes to dollars per share growth. There's still plenty of work ahead, but days like this most certainly deserve to be celebrated.

这对特斯拉来说是一个非常出色的季度，接下来的12个月中确实有很多令人期待的事情。这也是为什么特斯拉股票最近经历了历史性交易日的部分原因。我还在确认这一点，但之前特斯拉股票单日涨幅最大的一次是在2022年1月3日，上涨了47.67美元。如果调整分拆因素，这或许刚刚成了特斯拉有史以来单日美元涨幅第二大的交易日。未来还有很多工作要做，但像这样的日子无疑值得庆祝。