A historic U.S. agreement, the most important energy move in decades, just slipped under the radar, and once you understand what it solves, the whole AI boom starts to look different. |
In my last few essays, we've spoken about how Nvidia's latest quarter wasn't just strong. It bent the curve of what a modern company is expected to deliver: |
$57 billion in quarterly revenue — a company record. $51.2 billion from data centers alone — another record. Guidance of $65 billion next quarter.
|
Those numbers tell one story: AI is scaling at industrial speed. |
But the market's reaction told a different one. Instead of soaring, Nvidia closed down. |
Why? |
Because investors have finally started to understand the real constraint in the AI boom. |
It isn't GPUs. It isn't model size. It isn't cloud capacity. |
It's electricity. |
AI workloads run continuously and not in cycles. They demand steady, unbroken, industrial-grade power. And the grid is struggling to keep up. |
A single hyperscale data center consumes the electricity of a small city, often equivalent to 100,000 households. The world is now building 500–700 data centers per year, including nearly 200 hyperscale facilities. |
That means we're effectively adding a new city to the global grid every 12–18 hours. |
And that grid hasn't seen a meaningful upgrade in decades. |
Countries and utilities are now being forced to admit the strain: |
Georgia is seeking ~10 GW of new capacity, with a bulk of that driven by AI/datacenter load. Florida approved a multi-year rate increase, in part to upgrade infrastructure stressed by industrial-scale power demands. Virginia and North Carolina are delaying new data-center approvals due to substation saturation. Ireland halted new Dublin data centers after they hit 30%+ of regional electricity consumption. Singapore froze data center approvals because they were overwhelming the grid. The UK warned London's grid cannot support the current expansion pace.
|
As the World Economic Forum put it: |
"Investors are increasingly pricing in geopolitical risk and long-term supply security." |
Translation: demand isn't the problem. Power is. |
AI's Energy Diet Requires Baseload, and That's Where the Gap Is |
The question we have to focus on is what kind of power an AI-driven world actually requires. |
And once you map those needs onto the grid, a missing piece becomes obvious. |
AI doesn't need "a lot" of electricity. It needs a specific type of electricity: |
Baseload. |
Baseload is the minimum level of electricity the grid must supply continuously, 24/7, to stay stable. |
If baseload falters, the grid falters. |
To qualify as baseload, a power source must: |
deliver steady output around the clock, operate independently of weather or time of day, offer frequency stability and voltage support, handle industrial loads without flicker or fluctuation.
|
Here's how the current energy mix stacks up in terms of baseload: |
✔ Natural Gas |
Reliable but price-volatile, carbon-heavy, and dependent on pipelines. |
✔ Coal |
True baseload but rapidly being phased out. |
✔ Hydro |
Excellent baseload, but can't scale due to geography. |
✔ Geothermal |
Same as hydro, excellent Baseload but limited to a handful of regions. |
✘ Solar & Wind |
These are critical parts of the energy mix, but not baseload. They depend on storage that doesn't scale to AI-level demand. |
✘ Batteries |
Great for stability. Impossible for 24/7 industrial loads. |
When you step back, a pattern emerges: |
An AI-driven world needs a reliable, carbon-free baseload foundation. |
The energy mix will keep growing, but it still lacks a scalable, steady, around-the-clock power source. |
That gap is what brings nuclear back into the conversation. |
Nuclear won't replace the growing energy mix, but it will serve as the stabilizing force that the rest of the system can build on. |
Because when it comes to baseload, nuclear doesn't compete with renewables; it fills a critical void. |
| 🚀 Get Moonshot Alerts Instantly | Markets move in minutes. Your inbox can't keep up. | Join the Moonshot VIP Text Service and get real-time alerts, portfolio updates, and first-move intelligence directly to your phone. | Be first to know. Anywhere. | 👉 [Join the Text List →] | |
|
|
America Reenters the Nuclear Game |
This brings us to one of the most important industrial-policy moments in modern U.S. history. |
In October, the U.S. government announced an $80 billion strategic partnership with Westinghouse, the largest nuclear initiative in 50 years. |
The plan includes: |
Building eight AP1000 reactors across the United States. Deploying Westinghouse's AP300 Small Modular Reactors (SMRs). Reviving domestic supply chains that have been dormant since the 1980s. Accelerated permitting and strategic financing. Government participation is linked to long-term performance.
|
All of this signals that nuclear capacity is now being treated the way semiconductors were treated under the CHIPS Act: critical to national competitiveness, national security, and national economic growth. |
It also signals that the U.S. is trying to regain ground ceded to Russia and China. |
While the U.S. paused nuclear development for decades, Russia built reactor fleets across Eastern Europe, the Middle East, and Africa. China is building reactors faster and cheaper than anyone else and is advancing all six Gen-IV designs simultaneously. |
The U.S. is finally responding. |
| 🚀 Get Moonshot Alerts Instantly | Markets move in minutes. Your inbox can't keep up. | Join the Moonshot VIP Text Service and get real-time alerts, portfolio updates, and first-move intelligence directly to your phone. | Be first to know. Anywhere. | 👉 [Join the Text List →] | |
|
|
SMRs: Nuclear Without the Historical Baggage |
Small Modular Reactors (SMRs) are the bridge between today's strained grid and the future energy system required for AI baseload, electrification, and industrial reshoring. |
They solve the biggest issue with traditional nuclear: scale and construction risk. |
SMRs are: |
Factory-built, not site-built. Modular, allowing phased deployment. Compact, fitting on retired coal sites or industrial corridors. Faster to permit and construct. Flexible, able to sit near data centers, manufacturing hubs, or AI clusters.
|
They provide the same clean, continuous baseload without the volatility of megaprojects. |
And because SMRs use the same fuel architecture as AP1000 reactors, they benefit from a shared fuel cycle and supply chain. |
This matters because it means once the supply chain is built, scaling becomes exponentially easier. |
Why Nuclear Matters in a Multi-Source Future |
The future energy mix will be broad: solar, wind, batteries, natural gas, hydro, and nuclear all play roles. |
But the physics of the grid are clear. |
Renewables grow fast but need stability underneath. |
Gas fills gaps but is volatile and carbon-intensive. |
Hydro and geothermal help, but only where geography allows. |
AI-era demand is different. It requires continuous power, industrial-grade reliability, frequency stability, zero downtime, carbon-free scaling…and at volumes the existing grid cannot provide. |
Nuclear doesn't replace renewables, but it enables their growth. |
It becomes the foundation that allows everything else to scale. |
| 🚀 Get Moonshot Alerts Instantly | Markets move in minutes. Your inbox can't keep up. | Join the Moonshot VIP Text Service and get real-time alerts, portfolio updates, and first-move intelligence directly to your phone. | Be first to know. Anywhere. | 👉 [Join the Text List →] | |
|
|
The Fuel Reality: The Uranium Squeeze |
Nuclear reactors don't run on optimism. They run on uranium. |
And the nuclear buildout has fuel implications that the market is barely beginning to understand. |
From the Westinghouse plan alone: |
One AP1000 reactor requires ~1.2 million pounds of uranium to start. |
A fleet of eight requires ~10 million pounds. |
Annual operations require ~3 million pounds. |
As if that weren't enough, we are already in a structural uranium deficit. |
The World Nuclear Association reports that reactor demand has exceeded primary mine supply every single year since 2016, with the gap plugged by stockpiles and secondary sources that are now rapidly shrinking. |
UxC estimates the shortfall at ~55 million pounds per year, even before accounting for SMRs or the new U.S. fleet. |
Some of the largest Western producers have warned repeatedly of a "durable long-term structural deficit" driven by years of underinvestment and a surge in long-term contracting. |
And that deficit widens with every new reactor, SMR, and international order. |
Regulatory filings show that the eight AP1000s require the United States government to commit to roughly 8–10% of global annual uranium mine supply before a single SMR is added to the grid. |
Few commodities on earth have demand trajectories as clear as this. |
What This Means for Investors |
If nuclear becomes the backbone of the AI era — and every signal from policy, utilities, and global energy markets points in that direction — then the investment map shifts dramatically. |
The opportunities will live in the supply chains, the fuel cycles, and the companies positioned to meet the kind of long-term, infrastructure-heavy demand that AI and electrification guarantee. |
The market is only beginning to understand that the next dominant theme is not just AI, it's the power behind AI. And the companies that provide the materials, components, and fuel that make nuclear possible stand at the center of a multi-decade capital cycle. |
This cycle favors: |
Uranium producers with long-life, tier-one assets, capable of supplying utilities in a contracting environment. Fuel cycle companies positioned in conversion, enrichment, and fabrication. Vehicles that track uranium demand, which benefit directly from rising long-term contracting and shrinking secondary supply. Manufacturers and suppliers inside the SMR and reactor-component ecosystem, who will scale alongside the Westinghouse buildout.
|
If you're serious about positioning yourself at the foundation of the most consequential shift in the global energy system in half a century, follow this trend now. |
A Brand New Recommendation |
Later this week, I'm revealing a brand-new recommendation exclusively for Premium Members — one that sits at the crossroads of the nuclear pivot, the uranium squeeze, and the AI-era baseload crisis we've been tracking. |
I can't share the name here because that's reserved for Premium Members, but the positioning is unlike anything we've discussed this year. |
If you've been following the structural shifts in this essay and feel the gravity of where the world is heading, you'll want to be there when it drops. |
If you're not a Premium Member, scroll down and decide if it's right for you. |
Premium Members: watch your inbox this week. We have a huge multi-decade trend in front of us, the full power of the United States behind us, and an unstoppable AI wave which make this upcoming recommendation a long-term, high-conviction hold. |
More to come. |
— Double D |
|
🔓 Premium Content Begins Here 🔒 |
|
In this week's Premium Section, you'll find a brand new recommendation for this explosive stage of the AI and nuclear baseload supercycle. | I hope you've been paying attention because many of our picks are currently beating the S&P by up to 3-to-1 this year. | Most financial newsletters charge $500, $1,000, even $5,000 per year. Why? Because they know they can. | I don't. | I built my wealth the old-fashioned way, not by selling subscriptions. | That's why I priced this at $25/month, or $250/year. | Not because it's low quality, but because I don't need to charge the typical prices other newsletters charge. | One good trade, idea, or concept could pay for your next decade of subscriptions. | The question isn't 'Why is this so cheap?' The question is, 'Why would I charge more?' | 👉 Upgrade to Premium Now | P.S. If this newsletter were $1,000 per year, you'd have to think about it. | You'd weigh your options. You'd analyze the risk. | But it's $25 a month. | That's the price of a bad lunch decision. | And remember, just one good idea could pay for your subscription for a decade. | 👉 Upgrade to Premium Now | |
|
|
|
Having trouble using Hometalk? 
