⚛ Why Amazon Is Betting Billions on Nuclear: What X-Energy's IPO Filing Actually Says

📋 What Showed Up in Our Pipeline This Week

On April 27, 2026, a 424B4 filing from X-Energy, Inc. landed in our IPO Intelligence pipeline. A 424B4 is a final IPO prospectus — the official document a company files with the SEC right before its shares begin trading. It's the legal version of "here's everything you need to know before you buy." And it's required to tell you the bad stuff, not just the good.

X-Energy priced its IPO at $23 per share on April 23 — above the $16–$19 target range, meaning demand was strong enough that they could charge more than planned. On its first trading day, April 24, the stock surged 27%. By May 3, it was trading at $30.54, giving the company a market cap of nearly $9 billion.

So: what is this company, why does Amazon care, and what does the filing actually say that the press release left out?

📚 The Lesson: Why Solar and Wind Can't Power an AI Data Center

Before we get into X-Energy specifically, there's a concept worth understanding: baseload power. It's the reason nuclear energy is suddenly the hottest topic in tech — and it's also why Big Tech's pivot to renewables ran into a wall.

Here's the problem with solar and wind: they only generate power when the sun is shining or the wind is blowing. An AI training cluster running 24 hours a day doesn't have that luxury. You can't tell a GPU farm "sorry, the weather's overcast today, come back tomorrow." Data centers need power that is always on, always reliable — what the energy industry calls baseload or "firm" power.

Nuclear power has always been baseload power — it runs continuously, unaffected by weather, and produces zero carbon emissions while operating. The problem was that traditional nuclear plants cost $10–$20 billion, took 15–20 years to build, and had a history of spectacular cost overruns. Nobody wanted to build one.

That's where Small Modular Reactors — SMRs — come in. Here's the key difference:

X-Energy's reactor, the Xe-100, is a high-temperature gas-cooled SMR. It uses helium gas — not water — as a coolant, which lets it operate at much higher temperatures. That means it can produce both electricity and industrial heat, making it useful for chemical plants, hydrogen production, and other heavy industries that can't easily electrify.

The fuel is also unusual: TRISO-X pebbles, where each uranium kernel is coated in layers of carbon and ceramic. The coating is so effective at containing radioactive material that even in a worst-case scenario, the fuel itself doesn't melt down. That's a fundamentally different safety profile from the reactors most people associate with nuclear accidents.

⚡ The Story: How AI Created a Nuclear Renaissance

For most of the past 40 years, "nuclear energy" and "growth opportunity" were not phrases that appeared in the same sentence. Three Mile Island (1979), Chernobyl (1986), Fukushima (2011) — each incident made nuclear harder to finance and harder to permit. The U.S. hasn't completed a new nuclear plant since the 1990s.

Then came AI. And with AI came a power demand problem that nobody fully anticipated.

Training a large language model uses roughly the same amount of electricity as thousands of homes. Running a data center full of GPU inference clusters uses even more — continuously, year-round. Microsoft, Google, and Amazon had all made aggressive commitments to run their operations on 100% renewable energy. Those commitments are running directly into the physics problem described above: you can't run a data center on intermittent solar and wind without massive battery storage, and battery storage at that scale doesn't yet exist at viable cost.

Amazon's solution: go nuclear. In 2023, Amazon's Climate Pledge Fund invested directly in X-Energy. The deal included an agreement to potentially deploy up to 5 gigawatts of power using X-Energy's technology — enough to power several large data center campuses. Critically, Amazon gets to count that power as clean energy toward its climate commitments, because nuclear produces zero carbon emissions during operation.

The first real-world project isn't a data center though — it's a chemical plant. Dow agreed to host the first commercial Xe-100 deployment at its facility in Seadrift, Texas. The U.S. Department of Energy backed it with $1.2 billion in support. Construction is expected to begin in the next few years, with first power delivery in the early 2030s.

By the time X-Energy filed its IPO prospectus, it had raised over $1.4 billion from private investors including ARK Invest, Citadel's Ken Griffin, Ares Management, Centrica in the UK, and Ontario Power Generation in Canada. The IPO raised another $1 billion — pricing above range because demand from institutional investors exceeded supply.

The IPO Journey: How the Story Changed Between March and April

Our pipeline actually tracked X-Energy through its full IPO process. The company filed its initial S-1 registration statement on March 20, 2026 — five weeks before shares started trading. Then it filed the final 424B4 prospectus on April 27. Comparing the two tells you something interesting about how companies position themselves during an IPO roadshow.

Here's what changed in the narrative between the S-1 and the 424B4:

The underlying technology hadn't changed. What changed was the pitch. Between March and April, X-Energy's investment bankers — J.P. Morgan, Morgan Stanley, and Jefferies — ran the IPO roadshow and learned what institutional investors were most excited about. The answer: AI. So Amazon moved from a supporting detail to the headline, and the AI data center narrative moved from context to the core thesis.

The result: the IPO was upsized. They originally planned to raise less than $1 billion. Demand was strong enough that they sold 44.25 million shares at $23 — above the $16–$19 target range — raising $1.018 billion. An upsized IPO priced above range is one of the strongest signals that institutional investors competed to get in before the first trade.

🔍 What the 400-Page Prospectus Actually Says

Here's the thing about IPO prospectuses: they are legally required to tell you every material risk the company faces. Not the marketing risks. Not the "macro environment is uncertain" boilerplate. The actual, specific risks that could cause you to lose money. The SEC requires this, and it's one of the most useful things in any 424B4 filing — if you know where to look.

Here's what X-Energy's prospectus disclosed that the press releases glossed over:

None of this means X-Energy is a bad company or a bad investment. It means you're buying a long-horizon bet on a technology that is genuinely promising but genuinely unproven at commercial scale. The filing is honest about this. The headlines were not quite as honest.

That gap — between what the press release says and what the SEC filing discloses — is exactly where informed investors find their edge.

📲 This Is Exactly What Stockadora Is For

X-Energy's 424B4 prospectus is 400 pages long. It was filed with the SEC four days before the IPO — the same window when most retail investors are reading headlines, not documents. By the time the stock opened at $23 and surged to $29 on day one, most people had no idea what a Tax Receivable Agreement was or what "pre-revenue nuclear startup" actually meant for their investment timeline.

Our pipeline read that prospectus and surfaced this:

That's not pessimism — it's the filing, translated. And it's exactly the context you need before deciding whether this belongs in your portfolio.

💡 The Lesson Worth Keeping

X-Energy's IPO is exciting for real reasons. The AI power demand problem is real. Nuclear's advantages for baseload power are real. Amazon's backing is real. The $1.2 billion in DOE support is real. These aren't hallucinated talking points — they're in the filing.

But here's the pattern worth remembering for any IPO, not just this one: every IPO prospectus tells two stories simultaneously. The front half is the pitch — the market opportunity, the technology, the partnerships, the vision. The back half is the disclosure — the risks, the financial gaps, the dependencies, the things that could go wrong. Retail investors typically read the pitch. Institutional investors read both.

The Tax Receivable Agreement is a perfect example. It's not buried in footnotes — it's in the prospectus, clearly labeled. But it requires understanding what a TRA is before you can evaluate whether it matters for your investment thesis. Most headline coverage of the X-Energy IPO didn't mention it once.

IPO prospectuses are some of the most information-dense documents in public markets. They're also free, publicly available, and required by law to disclose the things that could hurt you. The challenge has always been the reading — not the access.

That's what Stockadora is here to fix.