Most Investors Are Missing the Nuclear Stack
大多数投资者忽略了核能栈
Data centers ate 415 TWh in 2024 — 1.5% of all electricity on Earth. By 2030 that doubles to 950 TWh, more than Japan's entire grid. AI isn't just a compute problem; it's an energy problem. That's why every nuclear ETF returned 45–88% in 2025. But "buy a nuclear ETF" isn't a strategy — these five funds aren't interchangeable. Each one sits on a different rung of a four-tier stack: miners, fuel, reactors, utilities.
数据中心2024年消耗415 TWh——占全球用电量1.5%。到2030年翻倍至950 TWh，超过日本整个电网。AI不只是算力问题，更是能源问题。这就是为什么2025年每只核能ETF都涨了45–88%。但"买核能ETF"不是策略——这五只基金并非可互换。每只都坐在四层栈的不同梯级上：矿企、燃料、反应堆、公用事业。
The 4-Tier Nuclear Stack
4层核能栈
Nuclear isn't one industry — it's a ladder that builds on itself one rung at a time. Each tier moves for different reasons and pays off on different timelines.
核能不是单一行业——它是一架自下而上的梯子。每一层因不同的原因移动，按不同的时间表回报。
Tier 1 — The Miners (Uranium Supply)
第1层——矿企（铀供应）
Three ETFs, three flavors of the same trade. URA is the balanced index, URNM is the pure producer torque, URNJ is the early-stage accelerator.
三只ETF，同一交易的三种口味。URA是均衡指数，URNM是纯生产商杠杆，URNJ是早期加速器。
Global X Uranium — The Benchmark
Global X Uranium——基准
Most established, most liquid, most balanced single-ticker exposure. YTD +69%, fee 0.69%, dividend 1.69%. Led by Tier 1 (Cameco, NexGen, UEC) but also carries some fuel exposure and a handful of utilities — diversification without juggling tickers. Analysts: moderate buy, 10.5% upside over 12 months. Sits where conviction meets liquidity.
最成熟、最具流动性、最均衡的单只代码敞口。YTD +69%，费率0.69%，股息1.69%。以第1层为主（Cameco、NexGen、UEC），但也包含部分燃料敞口和少量公用事业——多元化无需多个代码。分析师：适度买入，12个月上行10.5%。在信念与流动性的交汇点。
Sprott Uranium Miners — Tier 1 Torque
Sprott Uranium Miners——第1层杠杆
YTD +45%, fee 0.75%, dividend 2.33%, $1.7B AUM. Established miners plus direct uranium exposure via the Sprott Physical Uranium Trust. No utilities, no enrichment, no SMRs. Tied directly to the supply side: production cuts, contract renewals, the long-term squeeze from growing demand. Accelerates when uranium tightens, cools when sentiment pauses. The straight line into Tier 1.
YTD +45%，费率0.75%，股息2.33%，$17亿AUM。成熟矿企加上通过Sprott实物铀信托的直接铀敞口。无公用事业、无浓缩、无SMR。直接挂钩供应端：减产、合同续签、需求增长带来的长期紧缺。铀紧时加速，情绪暂停时冷却。进入第1层的直线。
Sprott Junior Uranium Miners — The Accelerator
Sprott Junior Uranium Miners——加速器
YTD +45%, fee 0.80%, dividend 2.97%. 100% Tier 1 junior plays — Denison, F3 Uranium, Fission, Global Atomic — companies with real assets in the ground but years from supplying a reactor. Moves faster than producers when sentiment turns; falls harder when it cools. Analysts see 25% upside over 12 months — Wall Street takes future-supply tightness seriously. Treat like seasoning, not the meal.
YTD +45%，费率0.80%，股息2.97%。100%第1层初级玩家——Denison、F3 Uranium、Fission、Global Atomic——有真实地下资产但距供应反应堆还需数年。情绪转好时比生产商涨得快；冷却时跌得更狠。分析师看到12个月上行25%——华尔街认真对待未来供应紧张。当作调料，不是主菜。
Tiers 2 + 3 — Fuel and Reactors (The Middle)
第2+3层——燃料与反应堆（中间层）
The part most ETF lineups skip. Without the fuel cycle and the reactor designs, miners and utilities can't connect. One fund covers both.
大多数ETF阵容跳过的部分。没有燃料循环和反应堆设计，矿企和公用事业无法连接。一只基金覆盖两者。
Global X Uranium & Nuclear — The Middle Layer
Global X Uranium & Nuclear——中间层
YTD +60%, fee 0.85%. The rare ETF sitting between mining and the grid. Tier 2: Centrus, Cameco fuel services, specialty material players building out the enrichment bottleneck. Tier 3: NuScale, BWX Technologies, X-energy — the SMRs and micro reactors meant to drop next to a data center. Without these layers, nuclear can never scale to AI's demand. Analysts: 9.5% upside over 12 months — measured, supply-chain-aware, not chasing hype.
YTD +60%，费率0.85%。罕见的位于矿企与电网之间的ETF。第2层：Centrus、Cameco燃料服务、建设浓缩瓶颈的特殊材料玩家。第3层：NuScale、BWX Technologies、X-energy——可放在数据中心旁的SMR和微反应堆。没有这些层，核能永远无法扩展到AI需求规模。分析师：12个月上行9.5%——克制、关注供应链、不追炒作。
Tier 4 — The Utilities (Cash Flow Today)
第4层——公用事业（今日现金流）
The only tier that already shows up on a power bill — and on hyperscaler procurement contracts.
唯一已经出现在电费账单和超大规模企业采购合同中的层。
VanEck Uranium & Nuclear — Steady Compounding
VanEck Uranium & Nuclear——稳健复利
YTD +59%, fee 0.56% (lowest in the category), dividend 0.47%. Holdings read like the US grid: Constellation, Duke Energy, Entergy, Dominion — operators already running gigawatts and signing multi-decade hyperscaler power deals. Doesn't spike, doesn't collapse — just compounds. Utilities get paid to keep the lights on, not to find new ore. Analysts: 11.59% upside over 12 months. Where rising AI demand turns into revenue you can count on.
YTD +59%，费率0.56%（类别最低），股息0.47%。持仓像美国电网：Constellation、Duke Energy、Entergy、Dominion——已运行千兆瓦并与超大规模企业签订数十年电力合同的运营商。不飙升、不崩盘——只是复利。公用事业靠保持灯亮赚钱，不是找新矿。分析师：12个月上行11.59%。AI需求上升变成可指望的收入。
Five ETFs, Five Roles
五只ETF，五种角色
None of these ETFs match the full stack perfectly. That's the point — pick by what role you want it to play.
没有一只ETF完全匹配整个栈。这正是关键——按你想让它扮演的角色挑选。
Own the Stack, Not a Ticker
持有栈，而不是单个代码
Frameworks Beat Hype
框架胜过炒作
Don't buy nuclear by guessing which ticker is about to run. Buy it knowing which tier you're in. Mining, fuel, reactors, utilities — they don't move for the same reasons and they don't pay off at the same time. Each ETF leans differently, and that's the feature, not the bug. Own the stack intentionally, not emotionally.
不要靠猜哪个代码要爆来买核能。要清楚自己买的是哪一层。矿业、燃料、反应堆、公用事业——它们因不同原因移动，按不同时间表回报。每只ETF倾向不同，这是特点不是缺陷。有目的地持有栈，而不是情绪化。
Why Single-Tier Won't Work
为何单层行不通
If AI doubles, triples, or quadruples energy demand the way every forecast suggests, miners alone won't solve it. Utilities alone won't either. Innovation without fuel is just a press release. When nuclear scales, it scales across all tiers — not just one. The headline of 2025 was uranium price; the real story is fuel capacity, grid integration, SMRs, permitting, and long-cycle capital.
如果AI按每份预测所示让能源需求翻倍、三倍、四倍，仅靠矿企解决不了。仅靠公用事业也不行。没有燃料的创新只是新闻稿。核能扩张时，是所有层一起扩张——不只一层。2025年的头条是铀价；真正的故事是燃料产能、电网整合、SMR、审批与长周期资本。
Discipline Over Conviction
纪律胜过信念
The point of a framework isn't to sound smart — it's to stay disciplined when everyone else is chasing the loudest corner of nuclear. Want exposure to the part that actually turns uranium into electricity? That's the entire stack: miners at the bottom, fuel above them, reactors stacked on top, utilities plugged into the grid.
框架的目的不是听起来聪明——是当其他人都在追逐核能最喧嚣的角落时保持纪律。想要敞口于真正把铀变成电力的部分？那就是整个栈：底部矿企、上面燃料、再上反应堆、最顶端是接入电网的公用事业。