首页  >>  来自播客: The Limiting Factor 更新   反馈

Grid Series Video #2 // How Many Megapack Factories Will Tesla Build?

发布时间 2024-07-10 13:58:54    来源

摘要

In this video I'll walk you through my estimate for how many megapack factories Tesla will need to build. In short, about 6 to 13 in ...

GPT-4正在为你翻译摘要中......

中英文字稿  

Welcome back everyone, I'm Jordan Geisigee and this is The Limiting Factor. Tesla's Mega Pack business for grid scale battery energy storage more than doubled in 2023 and much more growth is expected in the coming years thanks to expansions already underway in Lathrop, California and Shanghai. This raises the question, how big could the Mega Pack business get and by when? To answer that question, today I'll walk you through my assumptions on how the adoption of battery energy storage will evolve between now and 2050, the potential market size of each phase and how much of that storage market Tesla could ultimately take. Before we begin, a special thanks to my Patreon supporters, YouTube members and Twitter subscribers as well as RebellionAir.com. They specialize in helping investors manage concentrated positions. RebellionAir can help with covered calls, risk management and creating a money master plan from your financial first principles.
欢迎大家回来,我是乔丹·盖西吉,这是限制因素。特斯拉的电网级电池储能业务——Mega Pack——在2023年增长了两倍多,而且由于加州拉斯罗普和上海的扩建工程已经在进行中,预计未来几年还会有更多增长。这就引出了一个问题:Mega Pack业务究竟能发展到多大规模,时间节点又是怎样的?为了解答这个问题,今天我将带你们了解我对至2050年电池储能采用趋势的假设、每个阶段的潜在市场规模以及特斯拉可能最终占有的市场份额。在开始之前,特别感谢我的Patreon支持者、YouTube会员和Twitter订阅者,以及RebellionAir.com。他们专注于帮助投资者管理集中持仓。RebellionAir可以帮助您制定备兑买权、风险管理计划并从您的财务第一原则创建一个资金管理方案。

First, note that the forecasts I created in this video should be considered back of the napkin estimates. There have been thousands of pages of content written on the topic of grid storage and how it's expected to evolve over the coming years. What I'm providing today is a rough interpretation of how I expect the grid to evolve with a focus on Tesla. As usual, I'll include my assumptions and thought process so that you can form your own view. Let's start with Tesla's estimate for the size of the grid storage market for batteries. Last year, after Investor Day, Tesla released a white paper titled Master Plan Part 3 Sustainable Energy for All of Earth. In that paper, their estimate was that for the US alone, the total amount of 8-hour lithium ion storage required to make the transition to renewable energy was 6.5 terawatt hours. That's helpful because it gives us a fair estimate for the maximum amount of batteries required.
首先,请注意,我在这个视频中创建的预测只是粗略的估算。关于电网储能及其未来发展,有成千上万页的内容被撰写。我今天提供的是我对电网未来发展的一个大致解读,重点关注特斯拉。和往常一样,我会包括我的假设和思考过程,方便你形成自己的观点。 让我们先从特斯拉对电网储能市场规模的估计开始。去年,在投资者日之后,特斯拉发布了一份名为《总体规划第三部分:地球的可持续能源》的白皮书。在那份文件中,他们估计,仅在美国,实现向可再生能源过渡所需的8小时锂离子储能总量为6.5太瓦时。这一点很有帮助,因为它为我们提供了对所需最大电池量的合理估算。

However, it's not a forecast, so it doesn't include information on what's likely and by when. As a side note, the paper doesn't define what's meant by 8-hour lithium ion storage. As far as I can tell, it just seems to be using 8-hour duration as a somewhat arbitrary unit of measurement to show how much energy is stored and released. In this case, that's 8 hours of duration multiplied by 815 gigawatts of power for a total of 6.5 terawatt hours of energy. That means the 8-hour duration doesn't seem to be implying anything about the duration rating that Tesla expects that most grid storage batteries will actually have.
然而,这并不是一个预测,所以它不包括哪些事项可能发生及其时间节点。顺带提一下,这篇论文并没有定义所谓的8小时锂离子储能具体指什么。据我所知,似乎只是将8小时的时长作为一个相对任意的衡量单位,来展示储存和释放的能量。在这种情况下,就是8小时的持续时间乘以815吉瓦的功率,总共是6.5太瓦时的能量。这也就是说,8小时的时长并不是指特斯拉预期大多数电网储能电池的实际时长能力。

Moving along to add some granularity on duration and timeframes, let's bring in forecasting from the National Renewable Energy Laboratory, or NREL. NREL expects that, for the US, the deployment of grid storage to support renewable energy will progress through four phases between now and 2050. You may notice that as the four phases progress that the durations get longer, why is that? Two reasons. First, as grids increasingly shift to intermittent energy sources like wind and solar, they'll require greater reserves of grid storage to provide power when there's demand spikes or shortages in supply. Second, it's more difficult to make long-duration energy storage cost effective. That's for several reasons, but let's cover just two.
接下来,为了更细致地讨论持续时间和时间框架,让我们引入国家可再生能源实验室(NREL)的预测。NREL预计,在美国,为了支持可再生能源的电网储能部署将会在现在到2050年之间经历四个阶段。你可能会注意到,随着这四个阶段的推进,持续时间会变得更长,为什么会这样呢?有两个原因。首先,随着电网越来越多地转向风能和太阳能等间歇性能源,它们将需要更多的电网储备来在需求激增或供应不足时提供电力。其次,使长期能源储存具有成本效益更加困难。这有几个原因,但我们只讨论两个。

First, longer duration storage has fewer revenue opportunities over the course of its life because it charges and discharges more slowly. Yes, grid storage that takes longer to discharge will typically last longer, but there's a limit to that. And in the meantime, the storage will be incurring expenses like maintenance, property taxes, and financing costs. Second, typically lower price volatility is seen on longer timeframes, which makes the revenue lower when energy is needed at the peaks. The fact that longer duration storage has fewer revenue opportunities and earns less for them means that cost is one of the primary limiting factors for the deployment of longer duration storage.
首先,较长时间的储能在其生命周期中会有更少的创收机会,因为它充电和放电的速度较慢。是的,放电时间更长的电网储能系统通常会持续更长时间,但这是有极限的。同时,储能设施还会产生维护、财产税和融资成本等费用。其次,通常在较长时间范围内价格波动较小,这意味着在电力需求高峰期时,收益会更低。由于较长时间的储能有更少的创收机会且收益更低,因此成本是限制较长时间储能部署的主要因素之一。

However, the cost of energy storage is continually getting cheaper thanks to economies of scale and technology improvements, so we'll gradually see more and longer duration energy storage projects deployed over time. Although NREL's four phases do provide more granularity than Tesla's monolithic 6.5 terawatt hour estimate, they still don't provide specific years for when they expect the phases to occur. They don't specify the types of technologies that will be used for grid storage, and they measure the capacity in terms of gigawatts of power and don't provide a figure for gigawatt hours of energy, all of which is important if we want to combine NREL and Tesla's estimate to create a forecast.
然而,随着规模经济和技术进步的推动,能源存储的成本正在持续下降,因此我们将逐渐看到更多更长期的能源存储项目实施。尽管NREL的四个阶段比特斯拉的整体6.5太瓦时估算更为详细,但它们仍未提供具体年份来预期这些阶段的发生时间。他们也没有明确说明将用于电网存储的技术类型,并且他们是以千兆瓦的功率来衡量容量的,没有提供千兆瓦时的能量数据。如果我们想结合NREL和特斯拉的估算来创建一个预测,这些信息都是至关重要的。

With that in mind, let's trim the information out of the NREL table that we don't need and then fill in the blanks. The first line of the NREL table includes capacity from before 2010 from pumped hydro. I'll delete that line because it's capacity that's already installed. The last line of the NREL table focuses on multi-day to seasonal capacity. Tesla's model assumes that seasonal energy storage will be handled by hydrogen, and I agree with that assessment. In the book, Monetizing Energy Storage by Ian Stoffel and Oliver Schmidt, they show that by 2040, grid storage durations of 16 hours or less will be dominated by lithium-ion batteries, and storage durations of 16 hours or more will be dominated by hydrogen. If you'd like to read that book, it's available for free online and I'll link it in the description. Since today's focus is on battery energy storage rather than hydrogen storage, we can eliminate the last line. But I'm still going to build in four phases for today's analysis rather than three. Why? Because there's significant overlap in the duration ranges in the NREL table for phases two and three. NREL overlap them for good reason, which is that the primary grid services they list can use multiple energy storage durations. However, my goal today is to look at how much of each duration is needed and by when, which means we need clear differentiation between the phases. So, I'm going to break the second and third phases into three phases, which cover durations up to four hours, eight hours, and twelve hours, and I'll leave phase one at up to one hour. As a side note, up to four hours means between one and four hours, and so on. But I'm listing specific durations, such as one, four, eight, and twelve hours rather than duration ranges, because it makes the calculations all do in a moment easier.
考虑到这一点,我们先删除NREL表格中不需要的信息,然后填补空白。NREL表格的第一行包括2010年之前抽水蓄能的数据,我会删除这一行,因为它是已经安装的容量。表格的最后一行关注的是多天到季节性容量。特斯拉的模型假设季节性能源储存将由氢气负责,这一点我也同意。在Ian Stoffel和Oliver Schmidt撰写的《货币化储能》一书中,他们表明到2040年,16小时以下的电网储能将由锂离子电池主导,而16小时以上的储能将由氢气主导。如果你有兴趣阅读这本书,可以在网上免费找到,我会在描述中提供链接。由于今天的重点是电池储能而非氢气储存,所以可以把最后一行移除。不过,我仍然打算把今天的分析分为四个阶段,而不是三个。为什么呢?因为在NREL表格中的第二和第三阶段的持续时间范围有很大重叠。NREL之所以这样设置,是因为他们列出的主要电网服务可以使用多种储能持续时间。然而,我今天的目标是查看每种持续时间需求的具体数量和时间点,需要明确区分各个阶段。因此,我将把第二和第三阶段分成三个阶段,分别覆盖最长四小时,八小时和十二小时的储能,而第一阶段依旧保持在最长一小时。顺带一提,“最长四小时”指的是一到四小时之间,以此类推。但我列出的是具体持续时间,如一小时、四小时、八小时和十二小时,而不是持续时间范围,因为这样更方便后面的计算。

Next, let's add some labels to each phase to provide a better understanding of the type of grid services each phase will focus on. Note that I'm just listing a primary grid service as a point of reference. The storage deployed in each phase could all be used in around a dozen different grid services. Phase one is labeled operating reserves for services like frequency regulation to stabilize the grid. Phase two is labeled peaking capacity that can be used to replace gas powered peaker plants. Phase three is labeled time shifting that can be used to smooth the load over a 24 hour period. Lastly, I've labeled phase four as renewables firming, which refers to the deeper reserves that I expect to be required to transition the grid to renewable energy.
接下来,让我们给每个阶段添加一些标签,以更好地理解每个阶段所关注的电网服务类型。请注意,我这里只列出主要的电网服务作为参考点。每个阶段部署的储能设备实际上可以用于大约十二种不同的电网服务。 第一阶段标记为“运行备用”,用于如频率调节等服务,以稳定电网。 第二阶段标记为“峰值容量”,用于取代燃气发电的峰值电厂。 第三阶段标记为“时间平移”,可以在24小时内平滑负载。 最后,我将第四阶段标记为“可再生能源调节”,这是指深层备用储能,我预计这将是转向可再生能源所需要的。

As for the development years, in my view, longer duration storage options will start being adopted as soon as the price is low enough to spur demand. And the lower the price, the quicker the adoption and the greater the demand. So beside the year column, I've added a column for dollars per kilowatt hour. A rough guess for when the price of battery cells reached the point where four hour grid storage with batteries became commercially viable was when the industry average cell prices reached $120 per kilowatt hour in 2020. Note that doesn't include the total installation cost. That is, I'm assuming that the total cost of storage roughly scales with the cost of the battery cells, which are the dominant cost in a battery energy storage project. From that $120 per kilowatt hour figure for four hour duration. Interestingly, we can get a rough idea of when eight and 12 hour duration will start becoming commercially viable. Eight hour storage will have half the revenue opportunities in a given timeframe and 12 hour duration will have roughly one third the revenue opportunities. That means the cells for eight and 12 hour storage will need to cost half to one third $120 per kilowatt hour or $60 and $40 per kilowatt hour. Energy storage of less than one hour actually has different economics and is often sold by power capacity rather than energy capacity. But to keep things simple, I quadrupled the cell price for four hour storage to arrive at the $480 per kilowatt hour for the phase one operating reserves. That threshold was crossed somewhere between 2013 and 2015. Again, these are very rough estimates, but they provide a quick and easy way to guess when a specific battery energy storage duration became viable or will become viable. Cell prices for LFP battery cells are widely expected to hit $60 per kilowatt hour in the next year in China, which will enable phase three and they're expected to hit less than $40 per kilowatt hour by 2030, which will enable phase four. On that note, it's unclear at this point whether lithium ion batteries will be able to hit $40 per kilowatt hour. But even if they don't, it's likely that sodium ion batteries will.
关于开发时间,在我看来,只要价格足够低以激发需求,长期存储选项就会开始被采用。而价格越低,采用速度越快,需求越大。因此,在年份一栏旁边,我添加了一栏显示每千瓦小时的价格。 大致估算一下,当电池单元的价格在2020年达到每千瓦小时120美元时,四小时电网存储用电池变得具有商业可行性。需要注意的是,这不包括总体安装成本。也就是说,我假设存储的总成本大致与电池单元的成本成比例,因为电池单元是电池能源存储项目中的主要成本。 从每千瓦时120美元的四小时时长价格来看,我们可以粗略估算出八小时和十二小时时长何时会开始具有商业可行性。八小时存储在给定时间框架内的收入机会将是四小时存储的一半,而十二小时存储的收入机会大约是四小时存储的三分之一。这意味着八小时和十二小时存储的电池单元必须分别达到每千瓦时60美元和40美元左右。 小于一小时的能源存储实际上有不同的经济学原理,通常按功率容量而不是能源容量出售。但为了简化问题,我将四小时存储的电池单价乘以四,得出阶段一操作储备的每千瓦时480美元。这一门槛大约在2013年至2015年之间被跨越。再次强调,这些只是非常粗略的估算,但它们提供了一种快速简便的方法来推测特定电池能源存储时长何时变得可行或将变得可行。 LFP电池单元的价格预计在明年将在中国达到每千瓦小时60美元,这将实现阶段三。此外,预计到2030年价格将降至每千瓦小时40美元以下,实现阶段四。在这方面,目前尚不清楚锂离子电池是否能够达到每千瓦时40美元。但是,即使锂离子电池不能达到这一价格目标,钠离子电池很可能会实现。

With those starting years in mind, I created a 20-year window for each phase of storage deployment to give a general idea of the overlap that will occur between the phases and the time scales involved to transition to renewable energy. But bear in mind, these aren't meant to be definitive timeframes. With the basic framework of the table in place, let's start filling out the figures related to how much power and energy need to be deployed. Starting with power, Enroll again provides broad ranges here, which makes it hard to pin down a specific amount of storage capacity deployed for each duration.
考虑到这些初始年份,我为每个存储部署阶段设定了一个20年的时间窗口,以大致了解各阶段之间的重叠情况以及向可再生能源过渡所需的时间。然而,请记住,这些并不是绝对的时间范围。有了表格的基本框架,让我们开始填充有关需要部署多少功率和能源的数据。首先谈功率,Enroll这里提供了一些大致的范围,这使得很难确定每个时间段具体需要部署多少存储容量。

However, hopefully in one of their other slides, for 2050, they provide a more detailed split of roughly one to one to point three for storage durations of up to four hours, eight hours, and 12 hours. Note that's not exact and I incorporated some of my own assumptions. For example, first, Enroll forecast that 12 hour storage in 2050 will all be from legacy pumped hydro and they don't show any growth in that market from 2020. As I said earlier, I expect that by 2040, the 12 hour storage market will be dominated by ion based battery technologies.
希望他们在另一张关于2050年的幻灯片中,更详细地划分储能时长,比如将储能时长分别为四小时、八小时和十二小时的大致比例设定为1:1:0.3。请注意,这不是精确的数据,我引用了一些自己的假设。例如,Enroll预测2050年用于12小时储能的技术都将是现有的抽水蓄能系统,他们没有预见到这一市场从2020年起有任何增长。然而,我之前提到过,我预计到2040年,十二小时储能市场将主要被基于离子技术的电池所占据。

So by 2050, I expect the prices to be so low that they actually encourage growth in the nine to 12 hour storage market to help buffer deployments of renewable energy. The second assumption comes about because Enroll didn't provide data on the timeframe for one hour or below for their storage modeling slide. The four phases slide does show less than 30 gigawatts, but that's based on a middle of the road estimate. That means we'll have to take an educated guess for the power deployment for one hour storage that scales roughly along with the other durations.
所以到2050年,我预计价格会低到实际上能够推动9到12小时储能市场的增长,以帮助缓冲可再生能源的部署。第二个假设是因为Enroll没有提供一小时或以下的储能建模时间框数据。四个阶段的幻灯片确实显示了不到30吉瓦的情况,但那是基于中间估计。这意味着我们需要对一小时储能的电力部署做出合理的猜测,其规模大致与其他时长的储能相匹配。

With all those assumptions in mind, I've estimated 64 gigawatts for phase one, 320 gigawatts for phase two and three and 106 gigawatts for phase four. Outside of the ratio derived from the Enroll table, why did I pick those specific numbers? Two reasons. First, Enroll's forecast was a middle of the road estimate, and I expect that in the long term battery prices will be lower than they expect, which will drive a faster transition to renewable energy.
考虑到所有这些假设,我估算了第一阶段需要64千兆瓦,第二和第三阶段需要320千兆瓦,第四阶段需要106千兆瓦。除了从Enroll的表格得出的比例,为什么我选择了这些具体的数字?有两个原因。首先,Enroll的预测是一个中等估计,而我预计长期来看,电池价格会比他们预期的更低,这将推动更快地向可再生能源过渡。

Second, because as we'll see in a moment, adjusting for duration, these numbers line up with Tesla's estimate of a maximum of 6.5 terawatt hours of lithium ion storage in the US. That estimate is of course overly bullish by 2050 because it's a maximum, but I'll make some adjustments later in the video that'll moderate that forecast. With the gigawatts of capacity and the average duration in place, it's straightforward from here to calculate the number of gigawatt hours required. By multiplying the gigawatts of capacity by the duration of the energy storage,
第二,因为我们很快会看到,如果考虑储能时间,这些数据与特斯拉对美国锂离子电池存储量的最大估算——6.5太瓦时——是吻合的。这个估算到2050年显然过于乐观,因为它是最大值,但我会在视频后面做一些调整,使这个预测更合理。在知道了容量(以千兆瓦计)和平均储能时间后,从这里开始计算所需的千兆瓦时就很简单了。只需将容量(千兆瓦)乘以储能时间,就能得到所需的千兆瓦时。

the result is 64 gigawatt hours for phase one, 1.28 terawatt hours for phase two, 2.56 terawatt hours for phase three, and 1.28 terawatt hours for phase four. That adds up to a total of 5.18 terawatt hours, which of course doesn't add up to the 6.5 terawatt hours in Tesla's estimate. That's because we're using a range of durations rather than the eight-hour duration that Tesla is using across the board. To account for the difference in units, I've converted all the energy figures for each duration into their eight-hour equivalents.
结果是:第一阶段产生64吉瓦时,第二阶段产生1.28太瓦时,第三阶段产生2.56太瓦时,第四阶段产生1.28太瓦时。总计为5.18太瓦时,这显然与特斯拉估算的6.5太瓦时并不一致。这是因为我们使用了不同的持续时间范围,而特斯拉的估算是基于统一的八小时持续时间。为了弥合单位差异,我已经将每个持续时间的能量数据都转换为八小时等效值。

Using four-hour duration as an example, a battery pack that has a four-hour discharge duration can cycle twice as quickly as a battery pack rated for eight hours. That means for the same number of gigawatt hours, it can move twice as much energy. Therefore, 1.28 terawatt hours of grid storage. With a duration rating of four hours is equivalent to 2.56 terawatt hours of grid storage with a duration rating of eight hours. After running that calculation for each duration,
以四小时的时间为例,一个拥有四小时放电时长的电池组,其循环速度可以比额定为八小时的电池组快两倍。这意味着在相同的千兆瓦时情况下,它能够移动两倍的能量。因此,1.28太瓦时的电网储能,如果其放电时长为四小时,相当于放电时长为八小时的2.56太瓦时的电网储能。经过对每个时长的计算后,

the total amount of eight-hour equivalent energy storage comes to 6.472 terawatt hours, which is close enough to Tesla's 6.5 terawatt hour figure. With the core of the forecast in place, the next question is, how much annual production will be required to supply these multiple terawatt hours of batteries? To answer that, it's just a matter of dividing the amount of energy deployment for each battery energy storage duration by how long we can expect each to last. That is their service life.
总计相当于八小时存储量的能源储存达到6.472太瓦时,这与特斯拉的6.5太瓦时数据非常接近。在这个预测核心已经确定的情况下,下一个问题是,需要多少年生产量才能供应这些多个太瓦时的电池?要回答这个问题,只需将每种电池储能时间的能源部署量除以预计每个电池的使用寿命即可。

I'm going to assume 10 years for one-hour storage, 20 years for four-hour storage, 25 years for eight-hour storage, and 30 years for 12-hour storage. I picked 20 years for four-hour storage because Tesla currently offers up to a 20-year warranty on their megapacks. But beyond that, the numbers are just educated guesses. That's because estimating the life of battery energy storage is difficult. Most people would assume that the life of a battery would be strictly determined by its cycle life,
我将假设一小时存储的寿命为10年,四小时存储为20年,八小时存储为25年,十二小时存储为30年。我选择四小时存储为20年,是因为特斯拉目前对其大型电池组提供长达20年的保修。但除此之外,这些数字只是基于经验的推测。因为估算电池能源存储的寿命是很难的。大多数人会认为电池的寿命完全取决于其循环寿命。

But reality is a bit more complex. That's because it's not just charging and discharging that causes degradation to megapacks. It's also the amount of time that the packs spent exposed to extreme temperatures and adverse weather conditions. That means it's not worth going too far into the weeds with trying to predict the service life of different grid storage durations for this video. After dividing the amount of storage required for each duration by expected service life, the amount of annual production capacity required is 6.4 gigawatt hours per year for Phase 1, 64 gigawatt hours per year for Phase 2, 102 gigawatt hours per year for Phase 3, and 43 gigawatt hours per year for Phase 4. But this is just for the US. What about at a global level? Let's assume that the rest of the world has roughly the same split of phases, timeframes, and durations that the US does. As I said earlier, Tesla estimated 6.5 terawatt hours of lithium ion batteries needed for the US, but they also estimated 46.2 terawatt hours needed for the world.
但是现实情况要复杂一些。这是因为使保存时间变短的不仅仅是充电和放电。电池包暴露在极端温度和恶劣天气条件下的时间也会导致电池包的老化。这意味着,在这个视频里,没必要深入预测不同电网存储时长的使用寿命。在将每种存储时长的需求量除以预期使用寿命后,每年的生产容量需求分别是:阶段1每年6.4吉瓦时,阶段2每年64吉瓦时,阶段3每年102吉瓦时,阶段4每年43吉瓦时。但是这些只是美国的情况。那么全球呢?我们假设世界其他地区的阶段、时间框架和时长分布大致与美国相同。如前所述,特斯拉估计美国需要6.5太瓦时的锂离子电池,但他们也估计全球需要46.2太瓦时。

That means to convert from the US production estimate to the global production estimate, we need to multiply by 7.1. The result is 45 gigawatt hours of production per year for Phase 1, 454 gigawatt hours per year for Phase 2, 724 gigawatt hours per year for Phase 3, and 305 gigawatt hours per year for Phase 4, for a total of 1.528 terawatt hours per year. How does that compare to Tesla's estimate for global battery energy storage production? Tesla divided 46.2 terawatt hours by 20 years of service life to arrive at 2.3 terawatt hours per year of production required to transition the world to sustainable energy. So my estimate for global annual production comes in at about 2 thirds of Tesla's estimate. That works out well, because Tesla's estimate is for full decarbonization at a global level, and I don't expect that to occur in the next 25 years.
这意味着,要将美国的生产估算转换为全球生产估算,我们需要乘以7.1。结果是,第一阶段每年生产45千兆瓦时,第二阶段每年生产454千兆瓦时,第三阶段每年生产724千兆瓦时,第四阶段每年生产305千兆瓦时,总计每年生产1.528太瓦时。那这和特斯拉对全球电池能源存储生产的估算相比如何呢?特斯拉将46.2太瓦时除以20年的使用寿命,得出每年需要生产2.3太瓦时的电量,以实现世界向可持续能源的转型。所以,我对全球年生产的估算大约是特斯拉估算的三分之二。这个结果还不错,因为特斯拉的估算是针对全球全面脱碳,而我认为这在未来25年内不太可能实现。

Given that most countries have set targets to decarbonize by 2050, 2 thirds makes for a good base case forecast. With the global forecast in place, let's look at how much of the global market share for battery energy storage Tesla can absorb with their megapack products. According to trendforce, for 2023, the global battery energy storage market was about 117 gigawatt hours, and according to Tesla's Q4-2023 earnings call, they deployed 14.72 gigawatt hours. That means Tesla deployed about 12.5% of global battery energy storage production in 2023. In 2022, the number was 12.8%, and in 2021, it was 6.6%. However, last year in 2023 was the first year that Tesla began scaling their first dedicated megapack factory, which they'll duplicate and scale over the coming years, and they're already planning to quadruple production capacity by the end of the year. With that in mind, I expect Tesla to take at least 15% of a global battery energy storage market over the coming 10 to 20 years.
鉴于大多数国家已经设定了到2050年实现脱碳的目标,三分之二的预测可以作为一个可靠的基准。现在,让我们看看特斯拉的Megapack产品能在全球电池储能市场中占据多少份额。 据trendforce统计,2023年全球电池储能市场约为117吉瓦时,而根据特斯拉2023年第四季度的财报,他们部署了14.72吉瓦时的储能产品。这意味着特斯拉在2023年占据了全球电池储能产量的约12.5%。2022年,这一比例是12.8%,而在2021年则是6.6%。 然而,2023年是特斯拉首次开始大规模生产其专用的Megapack工厂的第一年,他们计划在未来几年内复制并扩大生产,已经计划在年底前将产能扩大四倍。考虑到这一点,我预计特斯拉在未来10到20年内至少占据全球电池储能市场的15%。

What about a more bullish case? To work that out, it becomes more speculative. Let's walk through the methodology I used. I started by looking at current global energy demand from all sources, and divided up the global energy market into six regions and two countries, India and China. In the second column, I entered the share of the global energy market that each region or country represents. In the third column, I took a guess at what I think Tesla's maximum market share could look like in each region or country. Let's look at the assumptions that went into each. In the US, for the first three quarters of 2023, 13.142 gigawatt hours of grid storage was installed. Tesla deployed 11.522 gigawatt hours of grid storage in those same three quarters. Although we don't know where that storage was deployed, most of it was likely deployed in the US, and that's where Tesla's built their first megapack factory.
考虑一个更加乐观的情况?要弄清这一点,需要一些推测。让我来解释一下我使用的方法。我首先查看了来自所有来源的当前全球能源需求,并把全球能源市场划分为六个地区和两个国家,即印度和中国。在第二栏中,我填入了每个地区或国家在全球能源市场上所占的份额。在第三栏中,我猜测了特斯拉在每个地区或国家可能的最大市场份额。我们来看一下这些假设。在美国,2023年前三季度安装了13.142吉瓦时的电网储能设备。特斯拉在同样的三个季度中部署了11.522吉瓦时的电网储能设备。虽然我们不知道这些储能设备具体部署在哪里,但大部分很可能部署在美国,而且特斯拉在美国建造了他们的第一个Megapack工厂。

That is, Tesla currently dominates the battery energy storage market in the US, and that looks unlikely to change. So my assumption is that Tesla's maximum potential market share in North America over time could be as high as 75%. As for South America and Europe, they're the regions that are the nearest to the US, both geographically and geopolitically. But there will likely be some competition from China and local production, so I assumed a maximum of 50% market share. China has a strong and highly competitive local manufacturing base, but Tesla has a good relationship with China and they've already started working on their first megapack factory there, so I assumed a 25% market share. The Asia Pacific, in which I've included countries like Japan, Korea, Indonesia, and Australia, have a good relationship with the US, but they'll likely also have local battery production along with imports from China. So I'd be surprised if Tesla gets more than 25% of the market there as well. India's a big question mark because it's doing its best to build a domestic manufacturing base, so Tesla's growth in India depends on what kind of agreement Tesla and India can come to for imports and manufacturing. But even if they do come to an agreement, I expect competition from local companies. So again, I expect 25% maximum market share for Tesla. Lastly, Central Asia, Africa, and the Middle East are loaded with geopolitical question marks and they might get most of their battery storage from China. Furthermore, I don't know how fast they'll make the transition to sustainable energy, which could mean that by 2050, those countries see a smaller share of their total energy market transition to battery energy storage. So I've given Tesla at best a 10% market share in each.
也就是说,特斯拉目前主导着美国的电池储能市场,而且这种情况看起来不会改变。因此,我假设特斯拉在北美的最大潜在市场份额可能高达75%。至于南美和欧洲,它们在地理和地缘政治上都离美国最近。但可能会面临来自中国和当地生产的竞争,所以我假设50%的最高市场份额。中国有强大而高度竞争的本地制造基础,但特斯拉与中国关系良好,他们已经开始在那里建造第一座Megapack工厂,因此我假设25%的市场份额。亚太地区,包括日本、韩国、印度尼西亚和澳大利亚,与美国关系良好,但也可能会有本地电池生产以及从中国的进口。所以我认为特斯拉在那里的市场份额也不会超过25%。印度是一个不确定因素,因为该国正在尽力建立本地制造基础,所以特斯拉在印度的增长取决于特斯拉与印度达成的进口和制造协议。但即使达成协议,我也预计会有本地公司的竞争。因此我预计特斯拉的最大市场份额为25%。最后,中亚、非洲和中东充满了地缘政治的不确定性,他们可能大部分的电池储能会来自中国。此外,我不知道他们向可持续能源过渡的速度,这可能意味着到2050年,这些国家的总能源市场中转向电池储能的比例会较小。所以我认为特斯拉在这些地区的市场份额最多为10%。

After multiplying each region's or country's share of the global market by Tesla's share of each of those markets, then tallying up the results. It gives Tesla a maximum potential global market share of 35% for battery energy storage. Then if we plug the high case and the low case into the four phases from earlier, we can see that in time Tesla could be deploying at minimum about 229 gigawatt hours per year of grid storage and at most 535 gigawatt hours per year. Before we move on to the summary, what are the major risks to Tesla's grid storage business? I see too. The first is delays with connecting the megapacks to the grid and transformer shortages. According to Wood Mackenzie, those issues have resulted in up to 80% of the projects coming online being delayed. I'm not sure how solvable those two issues are and whether they're trending towards better or worse. It's potentially a good topic for the next video of the grid storage series. The second risk is battery supply. Currently the best cell chemistry and form factor for grid storage is prismatic LFP battery cells. Tesla's wholly dependent on that chemistry and form factor from companies like CATL. That's a strategic risk that may hold back their growth potential. Yes, there are rumors that Tesla's building their own prismatic LFP battery cell production in Nevada with spare equipment from CATL. But even if that's the case, it's going to take years to hit meaningful scale. The same is true for the 4680, which so far is only using a nickel chemistry and a form factor which may or may not be suited to grid storage and is still struggling to ramp.
在将每个地区或国家在全球市场中的份额乘以特斯拉在这些市场中的份额后,再将结果汇总,我们得出了特斯拉在电池储能领域的最大潜在全球市场份额为35%。然后,如果我们将最高情况和最低情况代入之前提到的四个阶段中,可以看出,随着时间的推移,特斯拉每年可以部署至少约229吉瓦时的电网储能,最多可以达到535吉瓦时。 在进入总结之前,特斯拉的电网储能业务有哪些主要风险?我看到两个。第一个是将Megapack连接到电网的延迟和变压器短缺问题。根据伍德麦肯兹的说法,这些问题导致多达80%的项目上线被推迟。我不确定这些问题能否解决,以及它们的趋势是向好还是向坏。这可能是电网储能系列下一个视频的一个好话题。 第二个风险是电池供应。目前,电网储能最好的电池化学成分和形式是棱柱LFP电池。特斯拉完全依赖于像CATL这样的公司提供这种化学成分和形式的电池。这是一个可能限制其增长潜力的战略风险。是的,有传闻称特斯拉在内华达州用CATL的备用设备建立了自己的棱柱LFP电池生产线。但即使真是如此,达到有意义的规模也需要几年时间。同样的情况也适用于4680电池,该电池目前仅使用镍化学成分和一种可能或不可能适用于电网储能的形式,并且仍在努力提升产量。

In summary, let's answer the question in the title of the video. How many megapack factories will Tesla build? Based on all the assumptions I shared today, if each megapack factory continues to be 40 gigawatt hours like the template factory and life trip, and Tesla needs to deploy 229 to 535 gigawatt hours per year, in the next 5 to 10 years, we'll see a minimum of about 6 and a maximum of 13 megapack factories. For some perspective, Tesla's currently using about 150 gigawatt hours of cells per year for all their products. So as long as the grid storage products have a similar profit margin to what Tesla's earning on average for their products today, over the next 5 to 10 years, there's a potential for Tesla to increase their market cap by 350% from just grid storage hardware. Note that doesn't include the profit Tesla could make by using their megapack hardware to build a business around selling power into the grid, which is another profit center altogether.
总结一下,让我们回答视频标题中的问题。特斯拉会建造多少个Megapack工厂?根据我今天分享的所有假设,如果每个Megapack工厂都像模板工厂和未来设想的那样,每年生产40吉瓦时,而特斯拉每年需要部署229到535吉瓦时,那么在未来5到10年内,我们将会看到至少6个,最多13个Megapack工厂。为了提供一些背景信息,特斯拉目前每年使用大约150吉瓦时的电池用于所有产品。因此,只要电网储能产品的利润率与目前特斯拉产品的平均利润率相似,那么在未来5到10年内,特斯拉通过电网储能硬件有可能将其市值增加350%。请注意,这还不包括特斯拉使用其Megapack硬件打造通过向电网售电来盈利的业务,这又是一个完全独立的利润来源。

As a final note, I've made a few bonus slides based on the four phases of energy deployment and global regional demand to take guesses as to where and when Tesla will deploy megapack factories for the 15 or 35% market share scenarios. Bear in mind, this is highly speculative because one of the most difficult things to predict is exactly when and where Tesla will build a new factory or new product. This first slide shows all the information on one page with the number of factories and the years. The next slide shows the 15% case on a year by year basis. The last slide shows the 35% case on a year by year basis.
作为最后一点,我根据能源部署的四个阶段和全球区域需求制作了一些额外的幻灯片,来预测在市场份额达到15%或35%的情况下,特斯拉可能会在何时何地部署Megapack工厂。请注意,这些预测是高度推测性的,因为最难预测的事情之一就是特斯拉具体会在什么时候和哪里建设新工厂或新产品。第一张幻灯片在一页上展示了所有相关信息,包括工厂数量和年份。接下来的幻灯片按年展示了市场份额达到15%的情况。最后一张幻灯片按年展示了市场份额达到35%的情况。

The first impression here is that this looks like a lot of work and a stretch, but megapacks are far easier to manufacture than a vehicle and that's reflected in the cost of the factories. A megapack factory costs about $400 million, whereas a Gigafactory for vehicles costs $5 to $10 billion. That means that for even the high case of 35% of the global market share, where Tesla would need to build 13 megapack factories in the next decade, the total investment is only equal to one Gigafactory. And Tesla's built three Gigafactories over the past five years alone, and Berlin, Shanghai, and Texas. So 13 megapactories in the next 10 years isn't a stretch. If you enjoyed this video, please consider supporting the channel by using the links in the description. Also, consider following me on X. I often use X as a testbed for sharing ideas, and X subscribers like my Patreon supporters generally get access to my videos a week early.
最初的印象是这看起来像是很多工作,而且有些困难,但与制造车辆相比,生产“大型电池组”要容易得多,这也反映在工厂的成本上。一座“大型电池组”工厂的成本大约为4亿美元,而一座生产车辆的超级工厂成本则在50亿到100亿美元之间。这意味着,即使在全球市场份额高达35%的情况下,特斯拉需要在未来十年内建造13座“大型电池组”工厂,总投资也只相当于一座超级工厂。而特斯拉在过去五年内已经在柏林、上海和德克萨斯建造了三座超级工厂。所以在未来十年内建造13座“大型电池组”工厂并不是个难事。如果你喜欢这个视频,请考虑通过描述中的链接支持该频道。此外,也请考虑关注我的X。我常常用X作为分享想法的试验平台,X的订阅者和我的Patreon支持者一样,一般都会提前一周看到我的视频。

On that note, a special thanks to my YouTube members, X subscribers, and all the other patrons listed in the credits. I appreciate all of your support, and thanks for tuning in.
对此,我要特别感谢我的YouTube会员、X平台的订阅者以及列在片尾字幕中的所有赞助者。非常感谢你们的支持,也感谢大家的收看。