Tesla: You Can't Fix Stupid ⚡️

Tesla shareholders have been wrestling with some harsh realities as of late, like the fact that for the last 3.7 years, September 1st 2020, Tesla stock has returned effectively 0%. This means if you would have just left your money in the NASDAQ composite index for the past 3 years, you would have been better off as that would have returned nearly 33%. And even worse, back to Tesla, anyone who started buying Tesla stock in this range is still down and those that bought around the peak are going to be down roughly 60%. It's periods like this why I often argue many normal investors are better off investing in low cost funds because they can't handle the swings of a high beta company like Tesla. I mean, if you have people that are getting paid to manage people's money professionally, like Ross Gerber and Gary Black selling Tesla down at these levels, not up here, do we really think your average retail investor with a 9-5 and a family is going to be able to do much better over time? So from a Tesla stock perspective, let's just call the spade a spade. The bears have been more right than wrong for a few years now, so clearly Tesla bulls are deluded fanboys, right? In all honesty, I'd respectfully argue there are a few of those out there but, as I always love to say, nothing ever goes up monotonically and long-term investing requires time horizons of at least 5 years and I'd prefer 7.

特斯拉的股东最近一直在面对一些残酷的现实，比如过去3.7年，即2020年9月1日至今，特斯拉股票实际回报为0%。这意味着如果你在过去3年把钱放在纳斯达克综合指数里，你会赚得比较多，因为那时的回报将近33%。更糟糕的是，回到特斯拉，任何在这个价格范围内开始购买特斯拉股票的人仍然处于亏损状态，而那些在高峰时购买的人可能亏损约60%。正是这种时期，我经常主张许多普通投资者最好投资于低成本基金，因为他们无法承受像特斯拉这样高贝塔公司的波动。我的意思是，如果像罗斯·格伯和加里·布拉克这样的专业人士在这些水平卖出特斯拉，而不是在高位，我们真的认为你们那些每天上班、有家庭的普通零售投资者长期来说能做得更好吗？所以从特斯拉股票的角度来看，我们就坦率地说吧。熊市已经在过去几年里更多地表现正确，因此显然特斯拉的多头粉丝都是被蒙蔽的粉丝，对吧？老实说，我要尊重地争辩，确实有一些粉丝，但我总是喜欢说，没有什么是单调上涨的，长期投资需要至少5年甚至7年的时间跨度。

With this context out of the way, not one of us can go back and change our investment strategy the past few years. What we can do is reassess whether being in Tesla stock makes sense for us individually going forward with everything we just learned. If you've been around, you've heard me say there will be a bifurcation in Tesla investors. Those that believe Tesla will solve FSD and those that do not. It's fine either way, having two sides is what makes a market, but Elon has now given everyone repeated warnings. It is only a matter of time before we exceed the reliability of humans and not much time at that. But really the way to think of Tesla is most entirely in terms of solving autonomy and being able to turn on that autonomy for a gigantic fleet. If you value Tesla as just like an order company, you would just have to fundamentally, if it's just the wrong framework, it will count to be. If you ask the wrong question, then the right answer is impossible. If somebody doesn't believe Tesla is going to solve autonomy, I think they should not be an investor in the company. I stand by what I said, Tesla has a very solid underlying business with car sales and energy products, but the valuation multiples Tesla is awarded would undoubtedly be much lower if autonomy was not part of the story.

在这种背景下，我们任何人都无法改变过去几年的投资策略。我们可以做的是重新评估是否持有特斯拉股票对我们个人而言是合理的，考虑到我们刚刚学到的一切。如果你一直在关注，你会听我说特斯拉投资者将会分为两派。一派相信特斯拉将解决FSD，另一派则持相反看法。无论哪种观点都可以，市场之所以存在，是因为有两种不同的立场，但埃隆现在多次发出警告。只是时间问题，我们很快就会超越人类的可靠性。但实际上，思考特斯拉的最主要方式是解决自动驾驶，并且能够为庞大的车队开启自动驾驶。如果你把特斯拉看作只是一个订单公司，那么你就必须从根本上重新思考，如果这种框架是错误的，那将是很重要的。如果问的是错误的问题，那么正确的答案就是不可能的。如果有人不相信特斯拉将解决自动驾驶问题，我认为他们不应该是该公司的投资者。我坚持我所说的，特斯拉有一个非常扎实的底层业务，包括汽车销售和能源产品，但如果自动驾驶不是故事的一部分，特斯拉所获得的估值倍数肯定会低得多。

But what Tesla just said about its production plans is a monumental shift from what many were expecting. Tesla's current installed capacity right now is over 2.35 million. They just said they're now planning to use these existing lines that have been operating at less than full capacity, which hurts margins to a degree to build a new line of vehicles plural. This point has to land on everybody, no new manufacturing lines, no new factories are required to take Tesla's installed capacity to over 3 million units per year and to roll out new vehicles in the process. Not only is this going to save Tesla hundreds of millions of dollars the next two years, but it's also going to save them time as Tesla is now expecting to launch these new vehicles in early 2025, if not later this year. It's certainly fair to have your doubts about Tesla's product launch timelines, but I'll repeat, the lines are already installed. This also likely means Giga Mexico, Giga India, Giga wherever will be put to the back burner at least for a bit while Tesla transitions to this next phase. This also means Wall Street is going to be forced to start modeling for the cybercab, what Tesla is calling the Robotaxy, and whatever other vehicles Tesla launches. The stock market prices in for activity 6-12 months into the future, and this new lineup now falls into that timeline. There will be a delay though because at this point we still have plenty of uncertainty on the specifics, but that's what August 8th is for. It's looking like Tesla may launch this new ride-hailing app with supervised FSD first. This screenshot showed 4 passengers max and the Model Y seats 5.

但特斯拉刚刚提到的生产计划与许多人的预期有着巨大的转变。特斯拉目前的装机容量已经超过235万台。他们刚刚表示，现在计划利用那些一直运营不足全产能的现有生产线，虽然会对利润率造成一定程度的伤害，来建立一个新的车辆生产线。这一点应该让每个人都明白，不需要新的生产线，也不需要新的工厂，就可以将特斯拉的装机容量提升到每年超过300万台，并在这个过程中推出新车型。这不仅将为特斯拉在接下来的两年内节省数亿美元，而且还将节省他们的时间，因为特斯拉现在预计将在2025年初甚至是今年晚些时候推出这些新车型。对于特斯拉的产品发布时间表有所怀疑是合理的，但我要重申，这些生产线已经安装好了。这也很可能意味着特斯拉的墨西哥工厂，印度工厂，无论在哪里的工厂，至少在短期内将被搁置一段时间，而特斯拉将转入下一个阶段。这也意味着华尔街将不得不开始为特斯拉命名的“抢手出租车”以及特斯拉推出的其他车型进行建模。股市会对未来6-12个月的活动进行定价，而这个新的产品线现在正好进入这个时间范围。不过，可能会有一些延迟，因为目前我们还有许多不确定因素，但这就是8月8日的目的所在。看起来特斯拉可能会首先推出这款新的网约车应用程序，其中包括监督自动驾驶。这个截图显示最多可容纳4名乘客，而Model Y座位可以容纳5人。

Elon also clearly outlined Tesla owners will be able to deploy their cars whenever they want like Airbnb. Tesla will give customers the option to filter who actually gets in their car, friends, family, 5 star users, etc. And the owner can have their car back anytime they need it. But before that point, why couldn't Tesla let owners do this with them in the driver's seat? Maybe on the weekends or evenings they want to make some extra money and educate passengers about Tesla and FSD in the process. And how about this one from Bernstein? They said we struggle with why Tesla needs a discrete Robotaxy offering. Let me help you out here. Given the AI scaling laws that Ashok mentioned, model scaling, data scaling, compute scaling, and architecture scaling, as the fleet grows, so does the training flywheel and this ever-growing feedback loop where Tesla has real-time dashboards of critical interventions to then feed into the system.

Elon还清楚地概述了特斯拉车主将能够随时像Airbnb一样使用他们的汽车。特斯拉将为客户提供选择筛选谁实际搭乘他们的汽车，朋友、家人、五星级用户等等。车主可以随时取回他们的车。但在那之前，特斯拉为什么不让车主坐在驾驶座上进行这样的操作呢？也许在周末或晚上，他们想赚点额外的钱，并在过程中教育乘客有关特斯拉和FSD的知识。还有来自伯恩斯坦的这个观点呢？他们表示我们不明白为什么特斯拉需要提供离散的Robotaxy服务。让我帮你解释一下。鉴于阿肖克提到的人工智能扩展规律，模型扩展、数据扩展、计算扩展和架构扩展，随着车队的增长，训练的飞轮也在增长，产生一个不断增长的反馈循环，特斯拉实时掌握关键干预事件的仪表板，然后将其反馈到系统中。

So after the next training cycle, the system improves. The scaling compute law implies Tesla can now train the neural nets for much longer with more compute which yields better results. And Tesla is working on better architectures which for the same level of compute also yields better results. And on the customer side, the sooner Tesla rolls out these cybercabs or even the existing fleet with a human supervisor and the ride-hailing app, the sooner the public gets comfortable with the app, the service, they can have good experiences and this will start to build that ever-important trust. That's going to be incredibly difficult to compete with. Just like Google search, plenty of competitors have tried to break that monopoly and have failed. We've heard people like Gary Black say that even if Tesla solves autonomy first, someone's going to copy them shortly after.

因此，在下一个训练周期之后，系统会得到改进。按照计算定律，特斯拉现在可以使用更多计算资源对神经网络进行更长时间的训练，从而获得更好的结果。而特斯拉也正在研究更好的架构，同样的计算资源下也能产生更好的结果。而在客户端，特斯拉越早推出这些无人驾驶出租车，甚至是现有的车队配备人类监督员和叫车应用程序，公众就会越早习惯使用这款应用程序和服务，他们可以获得良好的体验，从而建立起这种至关重要的信任。这将是非常难以匹敌的。就像谷歌搜索一样，很多竞争对手尝试打破这种垄断地位，但都未能成功。我们听到像Gary Black这样的人说，即使特斯拉率先解决了自动驾驶问题，其他公司很快也会效仿。

In my opinion, this is flat out wrong. The only other companies making EVs at scale other than Tesla are Chinese companies and their cars are not coming to the US anytime soon. Tesla is going to corner the US Robotecti market because as we've said for years, crews in Waymo cannot scale. And because Tesla investors have lacked the feeling of stability as of late, since I'm always thinking of you guys, here's that tangible stability many of us have been missing. We know a big part of Tesla's edges in data collection and sorting. As a society, we also have plenty of really good new data coming out that most people are not paying attention to. There are now many studies showing that prolonged sitting throughout the day is quite terrible for our health. What this means is that we should probably spend about half of our work time standing and about half of it sitting. If you had to do all one or the other, standing is going to be better than sitting. Ashley has now been using her adjustable desk from FlexiSpot for over one year, and they are the sponsor of this video, but we figured it was time for a long term review.

在我看来，这是完全错误的。除了特斯拉以外，唯一在规模上生产电动汽车的其他公司是中国公司，而且他们的车很快可能不会进入美国市场。特斯拉将主导美国的机器人技术市场，因为正如我们多年来所说的，Waymo的团队无法扩大规模。由于最近特斯拉的投资者一直缺乏稳定感，因为我一直在为大家着想，这就是我们许多人一直缺乏的那种实在的稳定感。我们知道特斯拉在数据收集和整理方面有很大优势。作为一个社会，我们也有许多非常出色的新数据出现，但大多数人并没有注意到。现在有很多研究表明，整天长时间坐着对我们的健康非常不利。这意味着我们可能应该把一半的工作时间花在站立上，另一半时间坐着。如果你必须选择站立或坐着，站着比坐着要好一些。Ashley现在已经使用FlexiSpot的可调节桌子超过一年了，他们是这个视频的赞助商，但我们认为现在是写一篇长期评测的时候了。

I have an adjustable desk as well, but if you want something that's affordable, very solid quality and easy to set up, FlexiSpot is a website to bookmark. Ashley assembled her FlexiSpot desk by herself in about 15 minutes. Now, let me just warn you that not all standing desks are that easy to assemble. One year later, the motors on this FlexiSpot desk are still going very strong. Now, I do not recommend doing this, but in the name of science, I figured it was worth showing you this is not some flimsy desk. You may also remember this unnecessary test that it passed with flying colors. Ashley's version has built-in storage, a memory function so you can choose your own preset heights, or you can use the arrows to set it where you want it each time. There are three USB charge ports for all of your devices, plus on the FlexiSpot website, there are a ton of different desk styles and desktop choices like bamboo, glass, chipboard, etc. FlexiSpot also offers a five year warranty and free returns for 30 days, so again, you really have nothing to lose. For stability, I'm a fan of their E7 desk series, plus they have some ergonomic chairs as well. I just think that for the money, it's a really great value. You can use my promo code electrified30 for $30 off on their com-har and E7 series. The link will be below.

我也有一张可以调节高度的桌子，但是如果你想要一些价格实惠、质量非常稳固且易于安装的东西，FlexiSpot是一个值得收藏的网站。Ashley独自用大约15分钟装好了她的FlexiSpot桌子。现在，我只是想提醒你，并不是所有的升降桌都那么容易安装。一年后，这张FlexiSpot桌子的电动机依然非常稳健。现在，我并不建议这么做，但为了科学的名义，我觉得值得向大家展示这不是一张易损坏的桌子。你可能还记得它通过了这个不必要的测试。Ashley的版本有内置的存储空间，一个记忆功能，这样你可以选择自己的预设高度，或者你可以使用箭头每次按设置它在你想要的位置。还有三个USB充电口供给你所有的设备用，此外在FlexiSpot网站上，有很多不同的桌子款式和桌面选择，比如竹制、玻璃、刨花板等。FlexiSpot还提供5年保修和30天免费退货，所以再次强调，你真的什么也不会失去。对于稳定性来说，我喜欢他们的E7系列桌子，还有一些符合人体工程学的椅子。我只是觉得，以这个价格，它真的是一个很不错的价值。您可以在他们的com-har和E7系列上使用我的优惠码electrified30，获得30美元的折扣。链接将在下面提供。

We also need to recognize how despite all of the disruptions in quarter one and the inventory build-up, thanks to production timing mismatch and cars stuck on ships, Tesla still managed to keep gross margins fairly stable coming in ahead of almost everyone's expectations. Tesla's proving to everyone they're more than capable of navigating unexpected challenges in tough environments. Despite all of Tesla's price cuts, Tesla is managing to cut vehicle production costs at a similar pace, giving those savings back to the customer. It's clear that Tesla is in the midst of an all-out assault on cost. Tesla expects to save $1.1 billion per year from these recent layoffs. Tesla is hyper-focused on using its CapEx efficiently, and Elon said they're not sacrificing any meaningful projects from the layoffs. And to debunk the stupidity that is still out there from people like this guy who are arguing Tesla faces the prospect of bankruptcy in three years because they burned $2.5 billion in cash in quarter one.

尽管在第一季度发生了诸多不确定因素和库存增加，包括由于生产时间不匹配和汽车被困船上，但我们也需要认识到，特斯拉仍设法保持了较为稳定的毛利率，超出了几乎所有人的预期。特斯拉正在向大家证明他们有能力在困境中处理意外挑战。尽管特斯拉进行了众多降价，但他们设法以类似的速度削减车辆生产成本，并将这些储蓄传递给客户。很明显，特斯拉正在全面攻击成本。特斯拉预计每年将通过最近的裁员节省11亿美元。特斯拉正在高度专注于有效利用其资本支出，埃隆表示他们并没有因裁员而放弃任何有意义的项目。此外，要辟谣那些仍然声称特斯拉可能在三年内面临破产的愚蠢说法，因为他们在第一季度燃烧了25亿美元现金的人。

This is just one of those, tell me you have no idea what you're talking about without telling me moments. Clearly, this guy did not listen to the call because the negative free cash flow was a one-off situation and is nothing to worry about. It was due to that inventory increase we just talked about thanks to the Red Sea disruptions, the Berlin arson attack, and maybe more importantly, it was because Tesla increased its spend on AI compute to the tune of $1 billion in the quarter. Vibov said Tesla expects this inventory build to reverse in quarter two and to return to positive free cash flow in quarter two. So, case closed. I know there will always be stupidity like this in the world that Tesla cannot fix, but we can all step up and provide rationales and facts to disprove disinformation like this. But while we're here, share buybacks will have their day at Tesla. But for now, I would much rather see Tesla buying compute than its own shares back. The faster Tesla solves FSD, the faster Tesla stock can take off.

这仅仅是那种告诉我你完全不知道在说什么的时刻之一。显然，这家伙没有听电话，因为负的自由现金流是一次性情况，不值得担心。这是由于我们刚刚谈到的库存增加，多亏了红海干扰、柏林纵火袭击，也可能更重要的是，特斯拉在这个季度增加了10亿美元的人工智能支出。维博夫表示，特斯拉预计这种库存增长将在第二季度逆转，并在第二季度恢复为正的自由现金流。所以，案件结案。我知道世界上总会有像这样的愚蠢事情，特斯拉无法解决，但我们都可以站出来提供理性和事实来驳斥这类虚假信息。但在这里，特斯拉的股票回购将有它们的时代。但眼下，我更愿意看到特斯拉购买计算机而不是自己的股票。特斯拉尽快解决完全自动驾驶，股票就会更快起飞。

Even Elon himself said share buybacks are dependent on generating significant positive free cash flow. Free cash flow is just operating cash flow minus CapEx. If we just look at Tesla's operating cash flow of $200 million with that inventory build and all of the disruptions from ramping the Model 3 Plus and the Cybertruck, if you needed proof that Tesla is elite at navigating storms, here you go. And from an earnings perspective, it's not major, but Tesla's tax rate was higher this quarter at 26% than it's been the past few quarters. But this could be a new normal since the local Shanghai government granted Tesla a corporate income tax rate of 15% from 2019 to 2023, lower than the usual 25%. But that has just expired.

甚至埃隆本人也表示，股份回购取决于产生显著正面的自由现金流。自由现金流只是营运现金流减去资本支出。如果我们只看特斯拉的营运现金流为2亿美元，再加上库存增长以及由于推出Model 3 Plus和Cybertruck而导致的所有中断，如果你需要证明特斯拉在应对风暴方面表现出色，这里就是一个例子。从盈利角度来看，这并不是什么重大问题，但特斯拉本季度的税率为26%，比过去几个季度要高。但这可能是一个新常态，因为上海当地政府从2019年到2023年给予特斯拉15%的企业所得税率，低于通常的25%。但这个优惠刚刚到期。

We said earlier this year it's wise to wait until quarter two to get a clear picture of the real demand for Tesla. So it was great to hear Elon and the team say they expect Q2 to be much better than Q1. And they expect to sell more cars this year than they did last year, despite the very slow start in quarter one. So this will clearly backload that sales growth for Tesla into the next three quarters. This right here should serve as strong evidence why we should never overreact to quarterly delivery numbers. The logistical challenge Tesla faces matching excess demand for certain models in certain markets is massive and this will not always work out well every quarter. But over a few quarters, these imbalances work themselves out and play out in the numbers. Tesla has also already streamlined the ordering process for customers so they can buy a Tesla in less than one minute.

我们在今年年初曾说过，最好等到第二季度才能清晰地了解特斯拉的真正需求。所以听到埃隆和团队说他们预计第二季度比第一季度要好得多真是太好了。他们预计今年销售的汽车数量将比去年多，尽管第一季度起步非常缓慢。因此，这将明显使特斯拉的销售增长后移至接下来的三个季度。 这正是表明我们永远不应该对季度交付数量过度反应的强有力证据。特斯拉面临的物流挑战是巨大的，在某些市场中部分车型存在供不应求的情况，这不会每个季度都顺利进行。但在几个季度内，这些不平衡会自行调整并在数字中体现出来。特斯拉已经为客户简化了订购流程，他们能在不到一分钟的时间内购买一辆特斯拉。

And just real quick, speaking of that excess demand for certain vehicles with the launch of the Model 3 performance yesterday, we already have people like Jason Camisa raving about the car. No, no, NVH, great ride. Superb seats amazing. Screen and all the infotainment. Great. I didn't even turn the stereo on. I've heard great stuff about don't care. Wait, what about body controls since we were talking momentarily recently about suspension? Like so comfort and body control both. And yes, the performance variant does qualify for the tax credit making this car a ridiculous value for what you get. Tesla knows they have to continually offer better products at better prices and that is precisely what they're focused on doing.

而且真的很快，说到昨天推出的Model 3性能版引起的某些车辆需求过剩，我们已经有像Jason Camisa这样的人对这辆车大加赞赏。没有噪音振动和耸人听闻的乘坐感受。座椅极好、屏幕和所有的信息娱乐设备都很棒。太棒了。我甚至都没有打开音响。我听说了些很棒的东西，但不在乎。等等，我们最近在谈论悬挂系统，那车身控制又如何呢？因此，舒适性和车身控制两者都很棒。是的，性能型号符合税收抵免条件，使得这辆车的性价比成为一个令人难以置信的价值。特斯拉知道他们必须不断提供更好的产品，以更优惠的价格，这正是他们专注于做的事情。

Shifting to Tesla energy just to run some simple numbers, we had record margins for Tesla energy in quarter one 24.6%. But that was not the highlight. Vibov said they expect Tesla energy storage deployments to grow at least 75% this year over 2023 and that this will begin to contribute significantly to overall profitability. Tesla's energy storage revenue was $6.035 billion in 2023. So the expected growth implies $10.6 billion of energy storage revenue for 2024. At least then if you toss a 26% gross margin on that, sure it could be higher. That's $2.8 billion in profits for the year. Tesla energy profits were only $1.1 billion in 2023 and we have the Shanghai mega-pactory set to enter production in early 2025.

为了简单计算，我们将重心转移到特斯拉能源。在第一季度，特斯拉能源的利润率达到创纪录的24.6%。但那并不是重点。维博夫表示，他们预计今年特斯拉能源存储部署将比2023年增长至少75%，这将开始对整体盈利贡献显著。特斯拉的能源存储收入在2023年达到60.35亿美元。因此，预期增长意味着2024年能源存储收入将达到106亿美元。至少，如果在这上面假设26%的毛利率，当然可能更高。这意味着该年的利润为28亿美元。特斯拉能源在2023年的利润只有11亿美元，而我们还有位于上海的巨型工厂计划于2025年初投产。

But at this point, if we're not laser focused on Tesla's autonomous efforts, we really are failing to see the company for what it plans to be here pretty soon. The team clearly said they do not think regulatory approval for robotaxes is going to be that challenging and that it should be straightforward thanks to companies like Cruz and Waymo paving the way. Tesla is expecting those efforts to pave the way for a broad-based regulatory approval in the United States and getting approval in at least a few states to start should be relatively easy. Then when you layer in Tesla's quality testing process and all of the data that Ashok laid out, the regulators are going to have a hard time denying approval when the data clearly shows the system makes roads safer. On top of that, we learn Tesla plans to roll out FSD globally, effectively wherever they get approval and that includes China. Tesla sounded confident about FSD's ability to hit the ground moving in other geographies as it stands now and then of course they can begin country-specific tweaks from there but many of the rules of the road do carry over. So not only can Cruz and Waymo not really scale in the United States but Tesla's system should receive broad US approval relatively easily and it also has a clear path to scaling globally. So these systems really are just not the same at all. And don't forget what Ashok said. So we have this like constant feedback loop of issues, fixes, evaluations and then it runs on repeat. And especially with the new beautiful architecture, all of this is automatically improving without requiring much engineering interventions. Tesla has a good sense of what the cars will do in three to four months ahead of us and Elon said versions 12.4 and 12.5 could be versions 13 and 14 as their almost complete retrains of the neural nets. Now yeah, Elon has certainly hyped up releases in the past but it's fair to be optimistic here given the progress of V12. Another data point that allows for optimism is Tesla growing its hardware compute by 130% quarter over quarter and they plan to more than double it again by the end of this year, expecting to exit 2024 with around the equivalent of 85,000 H100s. But don't forget it's not just about raw compute, it's about the efficiency of set a compute. And on the architecture side, that's where Tesla continues to refine.

但是目前，如果我们没有专注于特斯拉的自动驾驶努力，我们真的看不清公司近期的计划。团队明确表示，他们认为机器人出租车的监管批准并不会那么困难，并且由于克鲁兹和Waymo等公司的开拓，这应该是直接的。特斯拉期望这些努力为在美国获得广泛的监管批准铺平道路，并且至少在几个州获得批准开始应该相对容易。然后，当你考虑到特斯拉的质量测试流程以及阿肖克提出的所有数据时，监管机构将很难拒绝批准，因为数据清楚地显示该系统使道路更安全。除此之外，我们了解到特斯拉计划在全球范围内推出FSD，有效地无论他们在哪里获得批准，包括中国。特斯拉对FSD在其他地区的能力现在感到自信，当然他们也可以从那里开始进行特定国家的调整，但许多交通规则是通用的。所以克鲁兹和Waymo不仅在美国很难扩展，而且特斯拉的系统应该相对容易地获得广泛的美国批准，而且它还有一个明确的全球扩张路径。因此，这些系统真的完全不同。不要忘记阿肖克说的话。所以我们有这种持续反馈循环的问题，修复，评估，然后重复运行。特别是有了新的精美架构，所有这些都能够在不需要太多工程干预的情况下自动改进。特斯拉对车辆在我们前三到四个月内会做什么有很好的把握，埃隆说版本12.4和12.5可能会成为版本13和14，因为它们几乎完全重新训练了神经网络。是的，埃隆过去确实炒作过发布，但考虑到V12的进展，此处保持乐观是公平的。支持乐观主义的另一数据点是特斯拉硬件计算每季度按季度增长130％，他们计划在今年年底再次将其增加一倍以上，预计到2024年底将达到相当于85000个H100的水平。但不要忘记，这不仅仅是关于原始计算，还涉及计算效率。在架构方面，特斯拉继续完善。

Yeah, well, you know, let's talk about earth shattering. I can't think of anything more earth shattering than a car that drives itself. And I think that's exactly what people should be focused on. So yes, our, you know, previously published price target was $2,000 per share by 2027. We're updating that now. So look out for that research. But we think that two thirds of the enterprise value in that year could be attributable to a robo-taxi network to ignore this is crazy in my mind. I think that this Tesla will be one of the greatest AI companies of our generation. And you know, for a while they were compute constrained and that's not the case anymore. So basically, you know, new problems arise and they just need the compute to solve it. And now they they're, you know, have that or at least have line of sight to having that. I think that's huge. So I'm super excited for the robo-taxi event later this year.

是的，你知道的，让我们谈谈惊天动地的事情。我想不出比自动驾驶汽车更令人震惊的事情了。我认为这正是人们应该关注的。所以是的，我们之前发布的股价目标是2027年每股2000美元。我们现在正在更新这个目标。所以请留意我们的研究报告。但我们认为到了那一年，企业价值的三分之二可能要归功于一个无人出租车网络，忽视这一点在我看来是疯狂的。我认为特斯拉将成为我们这个时代最伟大的人工智能公司之一。你知道，有一段时间他们受到计算能力的限制，但现在已经不再是这样了。基本上，新问题出现了，他们只需要计算能力来解决。而现在他们已经有了，或者至少有望拥有。我认为这是巨大的。所以我非常期待今年晚些时候的无人出租车活动。

Elon did also say that even if he was abducted or left Tesla now, there's enough momentum that Tesla would solve autonomy for vehicles, maybe just a bit slower. He did, however, add that optimists and future products would become question marks. So if you want Tesla to have goofy levels of upside and a next generation of products that are not even on the map yet HVAC, boat, van, we really do need Elon at the helm. Given the absurd upside of Optimus long term, greater than everything else combined at Tesla. Yeah, I'm good with saying voting no on Elon's pay package is a dumb idea if you want a higher Tesla stock price in the future.

埃隆还说过，即使他被绑架或现在离开特斯拉，特斯拉也有足够的动力来解决车辆自主性问题，只是可能会慢一点。然而，他也补充说，乐观主义者和未来产品将成为疑问。所以，如果你希望特斯拉有着疯狂的增长潜力和下一代产品（比如暖通空调、船、货车）可能会超出预期，我们真的需要埃隆掌舵。考虑到Optimus的潜在增长，长期来看，这将超过特斯拉其他一切。是的，我觉得说拒绝埃隆的薪酬方案是一个愚蠢的想法，如果你想在未来获得更高的特斯拉股价。

I will say though, for some reason the comments Elon made about Optimus I'm having a hard time taking it face value. He said Optimus could be unlimited production in the factory doing useful tasks before the end of this year and they may be able to sell it externally by the end of next year. However, it is true that Tesla should be best positioned to reach volume production of a humanoid bot. Many of these other humanoid startups don't have high volume manufacturing experience. You can also argue that solving for a few use cases for Optimus in a closed factory setting will be much easier than solving for autonomy. So we'll see. We did get some clarity on a few loose ends as well and projects that aren't going to boost Tesla stock in the next year or so, but they're definitely things to look forward to.

然而，我必须说，埃隆关于Optimus的评论让我很难完全相信。他说Optimus可能在今年年底前可以在工厂内进行无限生产等有用任务，并且他们可能在明年年底前能够在外部销售。然而，事实上，特斯拉应该是达到人形机器人大规模生产的最佳位置。许多其他人形机器人初创公司没有大规模制造经验。你也可以争辩说，在封闭的工厂环境中解决Optimus的几个用例将比解决自动化问题容易得多。所以我们拭目以待。我们还对一些未解之谜和未来一两年内不会提振特斯拉股价的项目有了一些澄清，但这些肯定是值得期待的事情。

Tesla's finalizing the engineering of the semi, but as we've said all year, volume production was most likely not happening this year since the foundation has not been laid for the new factory yet. We did get a new realistic timeline for external customer deliveries not named Pepsi and it's now 2026 for volume semi production. Tesla will marginally increase production before then at Gig and Nevada on the current prototype-ish lines, but nothing meaningful before 2026. I also still find it comical how everyone acts like Tesla's regulatory credits are monopoly money.

特斯拉正在完成半挂车的工程，但正如我们全年都在说的那样，由于新工厂的基础尚未奠定，今年最有可能不会进行大规模生产。除了百事可乐之外，我们得到了一个新的现实时间表，现在2026年开始进行半挂车的大规模生产。在那之前，特斯拉将在吉格和内华达的当前原型线上略微增加生产，但在2026年之前不会有任何有意义的增长。我还觉得有趣的是，每个人都把特斯拉的监管信用当作是垄断货币。

Vibob clearly said they're not going anywhere for the next few years since almost all other major automakers are pulling back from selling BEVs. Critics have been saying these credits are going away since 2018. Tesla's competitors have funded billions of dollars of Tesla's growth in new vehicles, new factories and new tech, and this is going to continue for another few years at least. But perhaps the most satisfying benefit for Tesla of other automakers pulling back, Tesla is now getting batteries for significantly cheaper. And don't forget, the reason many of the world's largest battery makers now have excess supply is because legacy OEMs were placing massive orders last year and prior when they were charging ahead with EVs, but now those orders are way down.

Vibob明确表示，由于几乎所有其他主要汽车制造商都在减少销售电动车，他们在未来几年内不会走。评论家们自2018年以来一直在说这些税收优惠会消失。特斯拉的竞争对手已经为特斯拉在新车辆、新工厂和新技术的增长投入了数十亿美元，这种情况至少会持续几年。但也许对于特斯拉来说，其他汽车制造商撤离的最令人满意的好处是，特斯拉现在可以以更低的价格获得电池。不要忘记，世界上许多最大的电池制造商现在有过剩的供应，是因为传统汽车制造商在去年和之前大量订购电池，当时他们在推进电动车，但现在这些订单明显减少。

Meanwhile, Tesla has been going upstream to make deals for its own 4680s and they've passed on those learnings and those savings to their battery cell suppliers. So yeah, Tesla is supplying its battery suppliers with better deals on components while continually growing its orders from its battery cell partners. Who do you think these battery companies are going to favor going forward? The companies that basically screwed them over going back on their plans or the company that's actually saving them money on the supply side and continually placing bigger orders reliably. So even though 4680s are now more of a hedge and not as important as we once thought, to write this project off as a failure is ridiculous. Lars did say the goal was to beat supplier costs of nickel based cells by the end of this year. That's a huge milestone and yes, it'll take years to actually turn all of that R&D cost into cumulative profit.

与此同时，特斯拉一直在向上游寻找自己的4680电池，并将这些经验和成本节约传递给他们的电池供应商。所以是的，特斯拉给他们的电池供应商提供更好的零部件交易，同时继续增加从电池供应商那里的订单。你认为这些电池公司以后会偏向谁？那些基本上欺骗了他们的计划，还是那家实际上为他们节省供应成本并始终可靠地提出更大订单的公司。因此，尽管4680电池现在更多地起到对冲作用，不像我们之前认为的那么重要，但将这个项目视为失败是荒谬的。拉斯确实说过，目标是在今年年底打败基于镍的电池供应商的成本。这是一个巨大的里程碑，是的，将需要几年时间将所有的研发成本转化为累计利润。

But this is the first step that's required and keep in mind Tesla's battery suppliers cell costs have come way down making it a tougher milestone to hit for Tesla. Given the tone around the 4680 conversation, it sounds like Tesla doesn't plan to use 4680s for the next gen vehicles, which isn't a major surprise since it's likely those vehicles use LFP cells. But given Tesla said they don't have anywhere else to put 4680s other than the Cybertruck, we don't need Tesla to rapidly scale 4680s for now. The ramp will just stay ahead of Cybertruck production, keep improving yields and costs and in a few years or sooner. When battery prices go back up again when legacy OEMs try again on EVs, Tesla will be even better positioned by having 4680s. Ford is most likely. The automaker Tesla isn't talks with about licensing FSD. Elon said they may sign a dealer to this year, but it's going to take at least three to four years before any OEM actually installs Tesla's cameras and inference compute in its vehicles.

但这是必须的第一步，要记住特斯拉的电池供应商的电芯成本已经大幅下降，这使得特斯拉更难达到里程碑。考虑到关于4680电池的讨论，特斯拉似乎并不打算在下一代车辆中使用4680电池，这并不是什么大惊喜，因为这些车辆很可能使用LFP电池。但鉴于特斯拉表示他们除了Cybertruck之外没有其他地方可以放置4680电池，我们暂时不需要特斯拉迅速扩大4680电池的规模。生产速度将仅赶超Cybertruck的需求，不断提高产量和降低成本，在几年或更短的时间内。当传统汽车制造商再次投身电动汽车领域时，电池价格将再次上涨，而特斯拉将拥有4680电池，将更加有利。福特最有可能。特斯拉与其谈判授权FSD的汽车制造商。埃隆说他们可能今年签署协议，但至少需要三到四年时间，任何汽车制造商才能真正在其车辆中安装特斯拉的摄像头和推理计算。

So this is yet another catalyst for Tesla stock that will just be waiting in the wings for a few years. And yeah, there may be some short term pops on the announcement and it'll absolutely change the sentiment around FSD and Tesla's competitive position as it signals Ford or whoever else is throwing in the towel on trying to solve autonomy themselves. But when it comes to improving Tesla's financials and profits and being baked into models, we're going to have to wait. It also should be noted incorporating Tesla's hardware and software will be way easier than most of the competition if they ever even get to that point. Since a company like Ford would not have to integrate LIDAR, radar, ultrasonics, another benefit of Tesla's system being simple and much more cost effective. That translates into much less work for Ford and others and given how long it takes these companies to make model changes, that's a big deal. But again, it's probably 2028 or 2029 before a Ford vehicle has Tesla's FSD.

因此，这又是特斯拉股票的另一个催化剂，将在未来几年内等待。是的，公布消息时可能会有一些短期波动，这绝对会改变关于FSD和特斯拉竞争地位的情绪，因为这表明福特或其他公司正在放弃自行解决自动驾驶的努力。但要改善特斯拉的财务状况和利润，并纳入模型中，我们还需要等待。需要注意的是，如果他们真的达到那个阶段，那么整合特斯拉的硬件和软件将比大多数竞争对手要容易得多。由于像福特这样的公司不必整合激光雷达、雷达、超声波等，这是特斯拉系统简单且成本更低的另一个好处。这意味着对福特和其他公司的工作量要少得多，考虑到这些公司花多长时间进行模型更改，这是一大优势。但再次强调，福特的车辆很可能要到2028年或2029年才会搭载特斯拉的FSD技术。

The conversation about Tesla using its fleet of millions of cars with paid for distributed inference compute and cooling is certainly neat. And another call option that no other company will have at scale. But to me, this is a 2030s thing. Elon did use the illustration of 100 million Teslas so yeah, that's going to take some time. But if each car has 1kW of inference compute power running on a 60kW pack or so, is not a huge deal as you could run it for 10 hours and only use 10kW. With that built in liquid cooled thermal management which is a huge expense for data centers. When Tesla launches the ride hailing app, the utilization of some of the Tesla fleet will undoubtedly go up, meaning idle time and thus opportunities for revenue streams like this go down. But this is certainly something to not forget about and to look forward to as we head toward 2030.

关于特斯拉利用其数百万辆汽车进行分布式推理计算和冷却付费的对话确实很棒。这是其他公司在规模上都不具备的另一个期权。但对我来说，这是2030年的事情。埃隆确实提到有1亿辆特斯拉，所以这需要一些时间。但如果每辆汽车都有1千瓦的推理计算能力，在60千瓦的包装上运行，使用10小时只消耗10千瓦并不是什么大问题。还有内置的液冷热管理系统，这是数据中心的一大开销。当特斯拉推出搭乘应用程序时，部分特斯拉车队的利用率无疑会提高，意味着闲置时间和因此像这样的收入流的机会会减少。但这绝对是一件不应忘记的事情，也是我们期待2030年的前进方向。

As an investor, I'd much rather have the option than not and Elon highlighted Tesla's AI inference efficiency is vastly better than any other company. I think it's important we don't forget the reason Tesla has this edge all stemmed from overcoming a challenge. Tesla was constrained by the inference hardware on the car so they were forced to adapt and create a solution. Overcoming that is already paying dividends but has even greater capability to continue doing so years from now. I've been saying it's often darkest just before the dawn and Tesla's CFO reiterated just that. This is the pruning exercise which we went through and at the end of it will be much stronger and much more resilient to deal with the future because the future is really bright like I said in my wing remarks. We just have to get through this period and get there. And I'll just say it, there's a chance the darkest days are now behind us.

作为一个投资者，我宁愿有选择的机会，而不是没有。埃隆强调了特斯拉的人工智能推理效率远远优于其他任何公司。我认为重要的是我们不要忘记特斯拉拥有这一优势的根源是克服了一个挑战。特斯拉受限于车上的推理硬件，因此他们被迫调整并创建解决方案。克服这个挑战已经开始付出回报，而且在未来几年甚至有更大的潜力继续这样做。我一直说，在黎明之前通常是最黑暗的时刻，特斯拉的财务主管重申了这一点。这是我们经历的修剪锻炼，在结束时我们会变得更加坚强、更具弹性，以应对未来，因为正如我在前言中所说，未来真的是光明的。我们只需渡过这段时期，就可以到达目的地。而我要说的是，最黑暗的日子可能已经过去了。

Unexpected things can certainly happen and I'm not at all arguing the stock price goes straight up from here but peak fear should be off the table with FSD making great progress and a lower cost we think vehicle backs squarely in the picture and sooner than anyone was expecting. The way I see it the pieces really are now squarely in place. Elon is focused the team has been pruned compute is scaling rapidly regulatory approvals for robotaxes should be straightforward in the US. V12 has been pushed to 1.8 million vehicles and 50% are using it so far and that number continues to grow with Tesla quickly racking up hundreds of millions of miles on V12 in just a matter of weeks. Tesla is now going to maximize its current production lines blending those next gen principles with the current lines bringing those new vehicles sales should be much better the remaining 3 quarters this year. Battery costs are really low for Tesla and helping Tesla to continue to drive price cuts and thus affordability. We don't have to speculate about permanent FSD transfers anymore because Elon aggressively shot that down.

意外的事情肯定会发生，我并不是在辩论股价是否会直线上涨，但由于FSD取得了巨大进展并且成本更低，我们认为车辆在未来会迅速回升，比任何人预期的都要快。从我的角度来看，现在一切已经就位。埃隆正在专注于此，团队已经被修剪过了，计算速度正在迅速增长，机器出租车在美国的监管批准应该是比较简单的。V12已经推广到180万辆车，目前有50%的车辆正在使用，而这个数字还在不断增长，特斯拉在短短几周内在V12上已经累积了数亿英里的行驶里程。特斯拉现在将最大程度地利用当前的生产线，将下一代原则与当前的生产线相结合，这将大大提高今年剩下的3个季度的新车销售量。特斯拉的电池成本非常低，这有助于特斯拉继续降价，从而使其更加实惠。我们不再需要猜测关于永久FSD转让的问题，因为埃隆已经坚决否认了这一说法。

The Cybertruck ramp already hit 1000 per week and yes there are supplier limitations but as you'd expect at this point they're focusing on quality and things are progressing. In just 105 days we're going to have the cybercab unveil and hopefully learn much more about these other new products and how Tesla will handle the ride hailing app. Energy storage is on track for a major growth year which will provide a material boost to Tesla's profits. More deferred revenue recognition is likely around the corner with reverse and smart summon coming to the FSD stack and from the 10Q Tesla expects to recognize over 800 million dollars in the next 12 months. Tesla has a handful of encouraging call options that are already produced and paid for to serve as catalysts years into the future. We've already endured a major drawdown in the Tesla stock price. Even though I don't fully believe Elon's timing on the optimist's comments there are certainly arguments to be made that scaling optimist will be much easier than solving FSD.

Cybertruck已经达到每周1000辆的产量，尽管供应商存在限制，但目前他们正专注于质量，事情正在向好的方向发展。在短短105天内，我们将看到Cybercab的揭幕，希望能了解更多有关这些新产品以及特斯拉将如何处理搭乘应用程序的信息。能源储存的增长势头良好，将为特斯拉的利润提供实质性提升。特斯拉预计在未来12个月内将认定逾800百万美元的递延收入，其中包括了反向驾驶和智能召唤功能。特斯拉有一系列鼓舞人心的看涨期权，已经生产和付费，将成为未来几年的催化剂。虽然我不完全相信埃隆的乐观评论的时机，但可以肯定的是，有观点认为扩大乐观主义者将比解决全自动驾驶问题容易得多。最近特斯拉的股价已经经历了一次重大回调。

And I know this all may sound overwhelmingly positive but given the expectations everybody had going into the call I felt like I needed to portray this reversal and sentiment that I think these earnings and the call provided. So yeah while risks remain like higher rates for longer, a recession, Elon losing his conplane vote, unknown unknowns in the March of the Nines, potential delays with Project Juniper and these new vehicle delays slowing the current lines, escalating global wars, mainstream being programmed to be anti-EV etc. When it comes to the things Tesla can control, I think they're doing a great job of making tough choices, reorganizing, refocusing on hyper efficiency and communicating reasons for poor performance. It's also somewhat ironic that people have been exclaiming some Tesla shareholders should sell their shares. And now Elon has literally said as much on a conference call if you don't believe in the future Tesla is building. How many CEOs do you know with this level of honesty? So while I don't make personal recommendations about selling shares, you need to decide on your own if Tesla is actually transforming into a company that's not what you had hoped for.

我知道这一切可能听起来过于积极，但考虑到每个人在通话之前的期望，我觉得我需要展示这种转变和情绪，我认为这些收益和通话提供了。虽然风险仍然存在，比如长期的更高利率，经济衰退，埃隆失去他的监管投票，三月九日的未知未知数，项目董麻栎的潜在延迟以及这些新车辆延迟放慢目前的产线，全球战争升级，主流被编程为反电动汽车等等。当涉及到特斯拉可以控制的事情时，我认为他们在做出艰难抉择、重组、专注于超高效率以及传达业绩不佳原因方面做得很好。一些特斯拉股东应该出售股份的呼声也有些讽刺。现在埃隆在电话会议上确实说了，如果你不相信特斯拉正在建设的未来，就卖掉。你知道多少CEO会如此坦诚？所以，虽然我不做有关出售股份的个人建议，但你需要自己决定特斯拉是否真正正在转变成你所期望的样子的公司。

If you decide to stick around from here, you really should be on board with a refined company focus. The one thing I wish we got more clarity on was Tesla's new plan to advertise or not after cutting the content team in the US. I stand by the notion that Tesla desperately needs to run an educational campaign as ads are not going to move the needle for people who have been fed fud about EVs. Now only you can weigh all of these factors on your own, but for me, I got most of the clarity I needed from the call for now. We'll have to wait a few months for the Robotaxi unveil, but the next 3-5 years at Tesla look incredibly bright.

如果你决定从这里继续留下来，你真的应该全力支持公司的精细化业务重点。我希望能更清晰地了解特斯拉在美国裁减内容团队后是要推出新广告计划还是保持原状。我认为特斯拉迫切需要开展教育宣传活动，因为广告并不能改变那些对电动汽车充满疑虑的人们的观念。现在只有你自己能够权衡所有这些因素，但对我来说，通过电话我已经得到了大部分所需的清晰度。我们将不得不等待几个月以了解Robotaxi的揭秘，但特斯拉在接下来的3-5年看起来非常光明。

Circling back to the video intro, Tesla shareholders that have not sold have not lost money. You only lock in losses when you sell at a loss. For those that have actually been accumulating this entire time, the story is in the midst of being written. You can't judge a painting when it's half done, that's just foolish. You can't start a movie an hour in, watch 10 minutes, and then have an accurate understanding of the full thing. As we've said all along, the Tesla shareholders with patience and the proper time horizon for long-term investing will likely be handsomely rewarded in the end.

回到视频的开头，特斯拉的股东如果没有卖出股票，就没有亏钱。只有当你在亏损时出售股票时才会锁定亏损。对于那些实际上一直在累积股票的人来说，这个故事仍在撰写中。当一幅画只完成一半时，你无法判断它，这是愚蠢的。你不能在电影已经播放一个小时后开始观看，只看了10分钟，然后就全盘了解整个故事。正如我们一直在说的，对于长期投资而言，具备耐心和合适的投资时间视野的特斯拉股东最终可能会得到丰厚的回报。

Sure, there could have been moves made over the past few years with Tesla stock, but decades of data have proven no one times the market properly consistently. And if you just invest in quality companies via a dollar cost average method over years and decades, not weeks and months, you will likely generate wealth. The next 2-3 years are going to be some of the most exciting we've ever seen at Tesla, and we'll see who has the last laugh.

当然，过去几年中可能会有许多关于特斯拉股票的操作，但几十年的数据证明没有人能够稳定地正确时间市场。如果你只是通过分散投资的方法在数十年内投资于优质公司，而不是数周和数月，你很可能会创造财富。未来的2-3年将是特斯拉有史以来最令人兴奋的时期之一，我们将看到谁笑到最后。

And to everybody who's been calling Elon a part-time CEO, saying he doesn't care about Tesla anymore, all he cares about is X, I'd have you listen to this. Do you expect to lessen your involvement with Tesla at any point over the next 3 years? Well, Tesla constitutes a majority of my work time, and I work every day of the week. It's for me to take a study off an afternoon. So, I make sure Tesla is very prosperous, and it is like it is prosperous, and it will be very much so in the future.

对于那些称埃隆是兼职CEO，说他不再关心特斯拉，只关心X的人，我想让你们听听这个。你们是否指望在接下来的三年里减少对特斯拉的投入？特斯拉占据了我大部分工作时间，我每天工作。我只是偶尔会花点时间做些研究。所以，我确保特斯拉非常繁荣，它现在正如此，并且将来也会如此。