Nuclear Energy // Deconstructing the Bear Case

Welcome back everyone, I'm Jordan Geisigee and this is The Limiting Factor. For anyone who follows energy, geopolitics, or technology, it's clear that there's momentum building around nuclear energy. It's seeing growing public support, technological advances that could make it cheaper and safer, and it's the absolute champion when it comes to low carbon energy production. The question is, will that momentum result in a renaissance for nuclear power? That's what I'll attempt to answer with this new video series. Today we'll start by looking at the bear case for nuclear, and the best example of that is from Rethink X, which has forecasted that nuclear power production will collapse in the 2020s. In my view, that forecast is overly bearish because it doesn't take into account the role that nuclear plays in the overall grid mix doesn't seem to factor in the cost of existing nuclear generation compared to new nuclear generation. Doesn't factor in that nuclear power is the ideal power source for AI data centers, and the exponential growth in electricity demand that's expected from the broad-based push for greater electrification. Once again, bear in mind that today's video will just address the bear case. In upcoming videos, I'll get deeper into topics like whether nuclear power will be needed when solar, wind, and batteries are a fraction of what they cost today, and the economics of next-generation and small modular reactors.

欢迎回来，大家好，我是Jordan Geisigee，这是"The Limiting Factor"节目。对于关注能源、地缘政治或科技的人来说，显而易见的是，核能正在获得越来越多的关注。核能正受到越来越多的公众支持，并且技术进步可能会令其更便宜、更安全。同时，核能在低碳能源生产中表现出色。那么，这种势头是否会引发核能的复兴呢？这就是我将在这个新系列视频中尝试回答的问题。今天我们将从核能的悲观前景开始，最佳例子来自Rethink X，他们预测核能发电将在2020年代崩溃。依我看，这个预测过于悲观，因为它没有考虑核能在整体电网中所扮演的角色，没有比较现有核电与新建核电的成本，也没有考虑核电是人工智能数据中心的理想电力来源，以及预计全面推动电气化会导致的电力需求指数增长。再次提醒，今天的视频只会讨论悲观的观点。在接下来的视频中，我将深入探讨主题，比如在太阳能、风能和电池成本大幅降低的情况下，核能是否仍然需要，以及下一代小型模块化反应堆的经济性。

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. For this video, I'd also like to thank Nathan Eyre of the Rocky Mountain Institute, Jesse Pelton, founder of Type 1, Matt Smith of RebellionAir, Stephen Begalski of the Georgia Institute of Technology, and Lars Lee's doll of Ristad Energy. This video couldn't have been made without their insights, but as always, any errors are a reflection of my attempt to simplify and communicate a complex topic. To set the tone for this video series, the first point worth covering is that I'm technology agnostic when it comes to energy generation technologies. I only care about using the right tool for the job. That is, I don't believe there's a holy grail of electricity generation that'll solve every use case. So for me, the question is not, which energy technology is best? Because that's the wrong way of looking at the electricity grid, but rather, what are the use cases and what are the best solutions to each of those use cases? And from there, we can try to determine how much each technology will scale over time.

在我们开始之前，特别感谢我的Patreon支持者、YouTube会员、Twitter订阅者以及RebellionAir.com。他们专注于帮助投资者管理集中头寸。RebellionAir可以协助处理备兑期权、风险管理，并从您的财务原则出发制定资金规划。在本视频中，我还要感谢来自Rocky Mountain Institute的Nathan Eyre、Type 1的创始人Jesse Pelton、RebellionAir的Matt Smith、乔治亚理工学院的Stephen Begalski，以及Ristad Energy的Lars Lee's doll。如果没有他们的见解，这个视频是无法完成的，但如往常一样，任何错误都反映了我试图简化和传达复杂主题时的不足。在这个视频系列中，值得首先指出的是，我在能源发电技术上保持中立态度。我只关心使用最合适的工具来完成任务。也就是说，我不认为有一种万能的发电方法可以解决所有使用场景。所以，对我来说，问题不在于哪种能源技术最好，因为从电力网的角度来看，这种思路是不对的，而是要思考有哪些使用场景，以及每种场景的最佳解决方案是什么？从那里，我们可以尝试确定每种技术随着时间推移的发展规模。

In the case of this video series, I'll be focusing on nuclear energy. So let's start by taking a 10,000 foot view. There are three types of electricity generation. The first is inflexible, and the best example of that is nuclear power. The advantage of nuclear power is that it has up to around 90% uptime, which provides a solid foundation for the grid. However, the disadvantage of nuclear is that it has high capital costs. So in order to be profitable, a nuclear plant needs to run as much as possible. Additionally, a nuclear reactor takes days to spool up and days to cool down. That means it's easier for grid operators to run nuclear continuously rather than trying to chase the daily ebbs and flows in power demand by throttling the reactor.

在这个视频系列中，我将专注于核能。首先，我们从一个宏观视角来了解电力的生成方式，总共有三种。第一种类型是不灵活的，其中最好的例子就是核电。核电的优势在于其运行时间可以达到大约90%，为电网提供了稳定的基础。然而，核电的缺点是其资本成本高。因此，要实现盈利，核电站需要尽可能多地运行。此外，核反应堆需要几天的时间才能启动和冷却。这意味着，对于电网运营商来说，连续运行核电更容易，而不是通过调节反应堆的功率来适应每日的电力需求波动。

The next type of electricity generation is flexible generation, and the best example of that is natural gas-powered plants that can adapt to demand by increasing and decreasing output. There are two types of gas generation. Capital cycle gas generation has low capital cost because it uses relatively simple technology in the form of a gas turbine. But the operational costs are high because they have low fuel efficiency. That combination of qualities makes them perfect for peak power demand because they can sit idle until they're needed, then capitalize on high power costs during peak demand. Combined cycle gas generation has moderate capital costs because they're more complex. They not only have a gas turbine, but also a heat recovery system and boiler that runs a steam turbine. That results in greater fuel efficiency. That means combined cycle gas turbines are low cost enough to operate as base load, but they can also ramp up and down quickly with demand to serve as flexible generation.

下一种发电方式是灵活发电，其中最典型的例子就是天然气发电厂，它可以通过增加或减少产量来适应需求。天然气发电有两种类型。简单循环发电具有较低的资本成本，因为它使用的是相对简单的技术——燃气轮机。但其运营成本较高，因为燃料效率较低。这种特点使其完美用于高峰电力需求，因为它们可以在不需要时保持闲置，一旦需要就可以利用高峰期的高电价。联合循环发电的资本成本适中，因为其技术比较复杂，不仅包括燃气轮机，还有余热回收系统和运行蒸汽轮机的锅炉，这使得燃料效率更高。这意味着联合循环燃气轮机的运营成本较低，可以作为基础负荷发电，同时它们也能随着需求快速调整，从而作为灵活发电。

With those two types of electricity generation, inflexible and flexible, grid operators can meet any demand while keeping costs low by utilizing the strengths of each technology. There's still one more type of generation, variable generation, and the best examples of that are solar and wind power. They have moderate capital costs and no fuel costs, but they only produce power when the sun is shining and the wind is blowing. That means they're low cost, but in order for them to provide reliable power, they have to be combined with inflexible or flexible generation to fill in the supply gaps, which is called firming. What all this means is that each electricity generation technology occupies a niche that it's best suited to, and the optimal solution for a reliable grid requires a variety of technologies. That technology mix can vary quite a bit depending on a number of factors like geography, interest rates, fuel costs, and latitude.

利用两种电力发电类型——不可灵活和灵活，电网运营商可以在保持低成本的同时满足任何电力需求，通过充分利用每种技术的优势。此外，还有一种发电类型是可变发电，最典型的例子就是太阳能和风能。这些发电方式的固定成本适中且没有燃料成本，但只能在阳光充足或风力充足时发电。这意味着它们成本低，但为了提供可靠的电力，必须与不可灵活或灵活的发电系统结合，来填补供应空缺，这个过程被称为稳定化。这一切说明，每种发电技术都有其最适合的应用领域，而要实现可靠的电网则需要各种技术的组合。这个技术组合会因地理位置、利率、燃料成本和纬度等因素的不同而有所变化。

In a meaningful circle, this is why I don't believe in a holy grail of electricity generation, and I try to be technology agnostic. It doesn't make sense to become a fanboy of one technology because it's counterproductive. For example, some people assume that somehow nuclear power is at odds with solar or wind power, or the inverse, which is that nuclear will make up for the shortcomings of solar and wind power. But in energy systems analysis, they often get lumped together. Why? Because they each have a tendency to produce excess energy. Solar power, for example, will sometimes produce more power than the grid can use on a sunny day, and nuclear power, for example, will continue to produce power at full tilt even when demand is low, more on that in a moment. What this means is that there's a positive correlation for solar, wind, and nuclear versus energy storage technologies like lithium ion batteries, which allow the excess energy to be captured when demand and prices are low, and sold into the grid when demand and prices are high.

在一个有意义的情况下，这就是为什么我不相信会有一种万能的电力生产技术，我努力保持对技术的中立。成为某种技术的狂热支持者并不明智，因为这可能适得其反。比如，有些人认为核能与太阳能或风能存在冲突，或者相反的看法是核能可以弥补太阳能和风能的不足。但在能源系统分析中，它们常常被一起考虑。为什么呢？因为它们都有可能产生过剩的能量。比如，太阳能在晴天有时会产生超过电网需求的电量，而核能则会在需求低时继续全负荷生产电力，稍后会详细说明这一点。这意味着太阳能、风能和核能与像锂离子电池这种能源储存技术之间存在正相关性，这些储存技术可以在需求和价格低时捕获过剩的能量，并在需求和价格高时将其输送到电网。

So, the solar and wind provide ultra-cheap power because they don't need fuel. The nuclear provides a reliable base load for when there's no sun and wind, and the battery storage makes better use of all of them. It's an ecosystem of complementary technologies. With that context in place, let's zero in on nuclear energy. As I said in the introduction, the most bearish forecast I've seen for nuclear is by Rethink X. They expect nuclear power to collapse by 2030. The core of their argument is that newly installed solar, wind, and batteries are soon going to be cheaper than nuclear power. That means the utilization rates of nuclear plants will drop, causing the cost of nuclear power to skyrocket, which in turn will result in those nuclear plants being shut down. However, there's four reasons why that logic may be incorrect. Let's take a look.

太阳能和风能提供超低成本的电力，因为它们不需要燃料。核能则在没有阳光和风的时候，提供可靠的基础负荷，而电池储能让所有这些能源的利用更加有效。这是一种互补技术的生态系统。在这个背景下，我们来关注核能。正如我在介绍中所说的，我见过的对核能最悲观的预测来自Rethink X。他们认为，到2030年，核能将崩溃。他们的核心论点是，新的太阳能、风能和电池即将比核能便宜。这意味着核电站的使用率将下降，导致核能成本飙升，从而导致这些核电站关闭。然而，有四个理由说明这个逻辑可能不正确。让我们来看看。

First, in most power grids, nuclear is treated as a must-run generation service, which means it operates at all times. Let's look at how that works in practice. As I said earlier, nuclear can't throttle quickly enough to match power demand. That night or other periods of low demand, that means the nuclear plant has to sell energy cheaply and underbid other generation technologies to continue running. That can often mean running at a loss during those periods, but it still improves the overall economics of the nuclear plant compared to throttling or shutting down. That's because the low fuel costs and high capital costs of nuclear plants mean they lose even more money if they're just sitting idle. With that in mind, this image from Rethink X doesn't add up for me. As far as I can tell, they expect nuclear plants to shut down for parts of the week or year, which might actually end up costing the nuclear plant operators more money overall than running them at a loss during periods of weak demand. That is, in my view, if the economics of a nuclear plant are no longer profitable in aggregate across periods of high and low demand, they'll just be shut down completely and mothballed rather than run intermittently. That means the utilization rates of the plants that are still running would remain above 90% and result in a flat line for capacity factor.

首先，在大多数电网中，核能被视为一种必须运行的发电服务，这意味着其必须全天候运作。让我们来看一下这在实际中的运作方式。正如我之前提到的，核能无法快速调整来配合电力需求的变化。在夜间或其他需求低的时段，这意味着核电站必须以较低的价格出售电能，甚至压低其他发电技术的出价以维持运行。在这些时段，核电站可能会亏损运行，但与减速或停机相比，这仍然能提升核电站的总体经济效益。这是因为核电站的低燃料成本和高资本成本意味着如果它们闲置，损失会更大。考虑到这些，我对Rethink X的这张图不太认同。在我看来，他们似乎认为核电站会在一周或一年的某些时段停运，实际上这可能会让核电站运营商比在低需求时期亏损运行更花钱。也就是说，如果核电站在高需求和低需求时期的整体经济效益已经无利可图，它们可能就会彻底关闭和封存，而不是间歇性运行。这样一来，仍在运行的电站的利用率会保持在90%以上，并且表现为一个平直的容量因子。

As a side note, even if nuclear power is more expensive, the continuous and stable power generated by nuclear plants may be worth a certain price premium that could keep them around on the grid longer than expected. How much that premium is, I'm not sure, and how long it could save nuclear plants from being mothballed, I don't know. But it is worth considering not just the price of electricity, but also the value it provides for grid stability. The second issue with Rethink X's analysis is that they appear to have focused on new nuclear power plants and ignored existing nuclear power plants, which have vastly different economics. They estimated that the cost of new nuclear generation in 2025 would be about $0.43 per kilowatt hour or $430 per megawatt hour. The most recent data we have from Lazard, published in June of this year, indicates that it's turned out to be $142 to $222 per megawatt hour or $14 to $0.22 per kilowatt hour, so about half the cost that Rethink X had predicted. What about fully depreciated nuclear generation from existing nuclear plants, which makes up almost all of the nuclear generation in the United States? According to Lazard, on average, the energy from those nuclear plants has a levelized cost of $32 per megawatt hour, which is about $0.03 per kilowatt hour.

顺便提一下，即使核能成本较高，核电站产生的持续和稳定的电力可能值得一定的价格溢价，因而使核电站在电网中的存在时间比预期长。这个溢价具体是多少，我不确定，而它能延长核电站多久的运作时间，我也不知道。但我们应该考虑的不仅仅是电价，还要考虑到它为电网稳定性提供的价值。Rethink X的分析有第二个问题，即他们似乎主要关注新建核电站，而忽略了经济性截然不同的现有核电站。他们估计，2025年新建核电站的发电成本大约为每千瓦时0.43美元或每兆瓦时430美元。而根据Lazard今年6月发布的最新数据，这一成本实际上为每兆瓦时142美元到222美元或每千瓦时0.14美元到0.22美元，大约是Rethink X预测的一半。那么，来自现有核电站的已完全折旧的核能发电呢，这几乎构成了美国大部分的核能发电？根据Lazard的数据，这些核电站产生的电力平均平准化成本为每兆瓦时32美元，大约是每千瓦时0.03美元。

For comparison, at the very low end with generous assumptions, onshore wind and utility-scale solar can be as cheap as $27 or $29 per megawatt hour. But again, that's with generous assumptions. As I said earlier, due to the fact that solar and wind are variable, they require other generation technologies to buffer and support them, which adds to their cost. The dark blue and light blue bars on screen show the cost of building new generation from solar and wind, with and without tax credits, across different energy markets in the United States. ERCOT is, for example, the Electric Reliability Council of Texas. The tan-colored bars above the dark blue and light blue bars represent the additional firming cost. As you can see, with the firming costs factored in, solar, wind, and batteries cost between $49 to $177 per megawatt hour, or about $5 to $0.18 per kilowatt hour. It's considerably more expensive than the $32 per megawatt hour, or $0.03 per kilowatt hour for fully-depreciated nuclear. That is, the competition from solar and wind generation isn't going to be shutting down nuclear plants at scale any time soon. Yes, there are some nuclear plants that have been shut down in recent years due to price competition, but for the most part, it wasn't due to price competition from solar and wind generation.

为了比较，按很慷慨的假设计算，陆上风能和大型太阳能发电的成本可能低至每兆瓦时27或29美元。但这只是基于慷慨的假设。正如我之前所说，由于太阳能和风能具有不稳定性，它们需要其他发电技术来缓冲和支持，这会增加它们的成本。屏幕上的深蓝色和浅蓝色条形图显示了美国不同能源市场中，建设新的太阳能和风能发电设施的成本，带有和不带税收优惠的情况下。比如，ERCOT是德克萨斯州电力可靠性委员会。深蓝色和浅蓝色条形上方的棕色条形代表额外的稳固成本。正如您所见，计入稳固成本后，太阳能、风能和电池的成本在每兆瓦时49至177美元之间，或者大约每千瓦时0.05至0.18美元。这显然比完全折旧的核电的每兆瓦时32美元或每千瓦时0.03美元要昂贵得多。也就是说，太阳能和风能发电的竞争暂时不会大规模关闭核电厂。是的，近几年有些核电站因价格竞争关闭，但大多数情况下，并不是因为太阳能和风能发电的价格竞争。

Instead, it was due to competition from natural gas generation, which, thanks to the fracking boom for fully-depreciated equipment, now costs $30 per megawatt hour all in because it's inherently firm. Furthermore, gas generation has the benefit of being able to spool up or down to match demand. That is, although Rethink X is correct that installing new solar, wind, and battery generation is cheaper than installing new nuclear, it's going to be years before the same is true for fully-depreciated nuclear plants. In fact, rather than a collapse in nuclear generation, it's looking increasingly likely that we're going to see nuclear production increase in the next 5-10 years. Why? Let's take a look at the third issue with the analysis from Rethink X. Their paper was written in 2020, before the AI boom.

相反，这主要是因为来自天然气发电的竞争。由于水力压裂技术的兴起，让设备的成本已经完全摊销，天然气发电的全包成本现在仅为每兆瓦时30美元，因为它本身就具备稳定的特性。此外，天然气发电还具有根据需求灵活调整发电量的优点。也就是说，虽然Rethink X认为安装新的太阳能、风能和电池发电设备比新建核电站要便宜，但对已经完全摊销的核电站来说，实现这一点还需要几年时间。事实上，与其说核电要崩溃，不如说在未来5到10年内我们可能会看到核能生产的增加。为什么会这样呢？我们来看看Rethink X分析中的第三个问题。他们的报告是在2020年编写的，那时人工智能的浪潮尚未兴起。

Goldman Sachs estimates that the power demand from all data centers globally was 500 terawatt hours in 2024, and that will more than double to about 1,050 terawatts by 2030. That number is probably conservative, but let's run with it. The average nuclear plant has a capacity of about 1 gigawatt, and assuming a utilization rate of 90%, that's 900 megawatts. Multiply that by 365 days per year and 24 hours per day, and the result is roughly 8 terawatt hours of generation per year. If the electricity demand from data centers increases by roughly 550 terawatt hours globally in the next 5 years, that's about 70 nuclear plants that would be needed by 2030. Yes, only about 20% of that power demand will be from AI, but that will only accelerate over time.

高盛估计，2024年全球数据中心的用电需求为500太瓦时，并将在2030年超过一倍，达到约1,050太瓦时。这个数字也许比较保守，但我们就用这个来计算。平均来看，一个核电站的发电能力约为1吉瓦，假设利用率为90%，即900兆瓦。把这个数字乘以365天和每天24小时，结果大约为每年8太瓦时的发电量。如果未来5年内，全球数据中心的电力需求增加约550太瓦时，那么到2030年，大约需要70座核电站来满足需求。是的，其中只有大约20%的电力需求会来自人工智能，但这一比例会随着时间推移而加速增长。

Additionally, while it's true that AI data centers can use energy from any type of generation, and will be, AI data centers and nuclear power are a great match for each other. That's why Microsoft announced that they're restarting the 3-mile island nuclear plant to power their AI data centers. They could have instead chosen to invest in solar, wind, and batteries or natural gas, but they didn't. Why is that? Some people would suggest it's because the plant is eligible for hundreds of millions of dollars from the zero emissions tax credits. However, other forms of energy generation are also eligible for tax credits, such as biomass, wind, and geothermal.

此外，虽然AI数据中心确实可以使用任何类型的能源，但是AI数据中心与核能发电是非常合适的搭配。这就是为什么微软宣布重新启动三里岛核电站来为他们的AI数据中心供电。他们本可以选择投资于太阳能、风能、储能电池或者天然气，但他们没有。为什么会这样呢？有些人认为这是因为核电站可以获得数亿美元的零排放税收减免。然而，其他形式的能源发电，比如生物质能、风能和地热能等也同样有资格获得税收优惠。

So, it still doesn't explain why nuclear is the preferred generation for AI data centers. Others might assume that restarting 3-mile island would be quicker than building new generation of other types, but that's not necessarily the case. It looks like 3-mile island isn't going to be back online until 2028, which is 4 years from 2024 when Microsoft announced the deal to restart the facility. Four years is enough time to bring completely new generation online from other sources. With all that in mind, in my view, the reason why Microsoft is using nuclear power for their AI data centers is threefold. First, as I said earlier, fully-depreciated nuclear power plants offer some of the cheapest power available, and it's likely going to be years before solar and wind with firming can compete at that price point. Microsoft could also build new natural gas generation, but that runs $45 to $108 per megawatt hour, or $0.11 per kilowatt hour.

所以，这仍然没有解释为什么核能是AI数据中心的首选发电方式。有人可能认为重启三里岛核电站比建造其他类型的新一代电站更快，但事实并非如此。看起来三里岛核电站要到2028年才能重新上线，而微软宣布重启该设施的协议是在2024年，也就是说还有四年时间。在这四年时间内，完全可以通过其他来源建成全新的发电设施。考虑到这些因素，我个人认为微软选择核能为其AI数据中心供电的原因有三个。首先，如我之前所说的，已经完全折旧的核电站可以提供极为低廉的电力，太阳能和风能还需要很多年才能在价格上与之竞争。微软也可以选择建造新的天然气发电设施，但其成本为每兆瓦时45到108美元，或每千瓦时0.11美元。

Second, nuclear power has the lowest carbon emissions of any energy generation technology, which reduces the public relations risks of building out power-hungry data centers. Yes, they will catch some flack from the people who still have apprehensions about the safety of nuclear, but the majority of Americans now support nuclear power. Third, as I said a moment ago, AI data centers and nuclear are a great match for each other. Both have high capital costs and need to run as much as possible to maximize their commercial viability. AI guarantees large and steady power demand 24 hours a day, seven days a week, and 365 days a year, and those are exactly the economics where nuclear works best. The last issue with Rethink X's analysis is that we're seeing massive revisions upward and the growth of the energy market in general.

其次，核能是所有发电技术中碳排放量最低的，这降低了建设高能耗数据中心的公关风险。虽然仍然会有人对核能的安全性心存疑虑并提出批评，但大多数美国人现在支持核能。第三，正如我刚才提到的，人工智能数据中心和核能是很好的组合。两者的资本成本都很高，并且需要尽可能多地运行以实现商业上的可行性。人工智能保证全天候、每周七天、一年365天都有大量而稳定的电力需求，这正是核能最佳的经济效益条件。关于Rethink X分析的最后一个问题是，我们看到能源市场整体的增长以及预测的重大上调。

After years of stagnation, the energy market in the US is expected to expand rapidly due to not just the all-devouring energy needs of AI data centers, but also broad-based electrification for technologies like EVs. Due to that, energy demand is expected to double or triple by 2050. As I showed in the last video of the grid storage series, Rethink X correctly forecasted that the price of solar, wind, and batteries is going to plummet thanks to the effects of scaling. If electricity demand was expected to remain flat in the next 5 to 10 years, I'd say their forecast that nuclear would be displaced in the next 5 to 10 years by cheap solar, wind, and batteries would have a higher likelihood of occurring because less demand means more price competition and the cheapest power would win out.

在经历多年的停滞后，美国的能源市场预计将迅速扩张，这不仅是因为人工智能数据中心对能源的巨大需求，还因为技术如电动汽车的广泛电气化。因此，预计到2050年，能源需求将翻倍甚至三倍增长。正如我在电网储能系列的最后一期视频中展示的，Rethink X准确预测，由于规模效应，太阳能、风能和电池的价格将大幅下降。如果在未来5到10年内，电力需求保持平稳，我会认为他们对核能将在便宜的太阳能、风能和电池面前被取代的预测有更大的实现可能性，因为较低的需求意味着更激烈的价格竞争，最便宜的电力将最终胜出。

However, if power demand ends up growing by 2 to 3 times in the next 25 years, that could have three impacts. First, despite lower generation costs thanks to solar, wind, and batteries, electricity prices could increase if the build-out of generation capacity and therefore electricity supply ends up chasing electricity demand. Second, on that note, a shortage of electricity generation, or higher prices, might slow the speed that older electricity generation technologies are removed from the grid, or even see them reopened, which is exactly what's happening with Microsoft and Three Mile Island.

然而，如果未来25年电力需求增长2到3倍，可能会产生三个影响。首先，尽管太阳能、风能和电池技术降低了发电成本，但如果发电能力的扩建速度赶不上电力需求的增长，电价可能会升高。其次，由于电力短缺或电价上涨，可能会减缓淘汰旧发电技术的速度，甚至可能导致这些旧技术重新启用，这正是微软和三里岛核电站正在发生的情况。

Third, higher electricity prices could mean there's a market opportunity to not only restart moth-bulb nuclear facilities, but also spur the build-out of next-generation nuclear facilities that have high initial investment costs. In summary, contrary to the bear case, nuclear power isn't going away anytime soon. Since all of the nuclear facilities in the United States produce energy from fully-depreciated equipment at $32 per megawatt-hour, which, with all costs considered, is some of the cheapest available. Yes, by 2030, the price of solar, wind, and battery systems are expected to drop by over 50%, which, even including firming costs, will bring them into price parity or make them cheaper than nuclear.

第三，电价上涨可能意味着存在一个市场机会，不仅可以重启闲置的核电设施，还可以推动建设高初始投资的下一代核电设施。总的来说，与悲观的预测相反，核电不会在短期内消失。由于美国的所有核电设施都是通过已折旧完的设备生产能源，成本为每兆瓦时32美元，这在所有成本考虑下是非常便宜的能源。确实，到2030年，太阳能、风能和电池系统的价格预计会下降超过50%，即使考虑到稳定成本，这些能源也会与核电价格持平，甚至比核电更便宜。

But there are other factors at play that'll likely keep existing nuclear plants operational for decades to come. The first is the AI boom, which is well-suited to using nuclear power, and that demand will only skyrocket in the coming decades. Microsoft's agreement at Three Mile Island is, for example, 20 years. The second is that electricity demand in general is expected to double or triple by 2050, which could mean that electricity supply lags demand, meaning higher prices, meaning that more expensive generation technologies could remain operational for longer. However, nuclear power from new generation facilities is a different story because in the United States it costs $142 to $222 per megawatt-hour.

但还有其他因素可能使现有的核电站在未来几十年内继续运营。首先是人工智能的蓬勃发展，它非常适合利用核能，这种需求在未来几十年内只会大幅增长。比如，微软在三里岛的协议就是20年。其次，预计到2050年，电力需求总体上将翻倍或三倍增长，这可能导致电力供应跟不上需求，从而导致价格上涨，这意味着更昂贵的发电技术可能会继续运行更长的时间。然而，新建核电设施的情况有所不同，因为在美国，每兆瓦时的成本为142到222美元。

So the key question to answer, which I hope to do in the next two videos of the series, is in the coming decade whether new nuclear plants and next-generation nuclear reactors have any real chance in the face-up competition from ultra-cheap solar, wind, and batteries. 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. 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.

所以，我希望在接下来的两期视频中回答的关键问题是，未来十年里，新的核电站和下一代核反应堆能否在与超便宜的太阳能、风能和电池的竞争中占据一席之地。如果你喜欢这个视频，请考虑通过描述中的链接支持这个频道。同时也可以考虑关注我在X上的账号。我经常用X来作为分享想法的平台，X上的订阅者通常和我的Patreon支持者一样，可以提前一周看到我的视频。顺便说一下，特别感谢我的YouTube会员、X订阅者和所有在片尾名单上的赞助者。我感激你们所有的支持，谢谢你们的收看。