🤖 The AI Trade Nobody's Watching

Good morning, power brokers.

Everyone spent two years buying the chip. Almost nobody bought the thing that turns the chip on.

Today: why the AI story is quietly becoming an electricity story, and the second-order trade hiding one step downstream of Nvidia.

TLDR is the free daily email with summaries of the most interesting stories in startups, tech, and programming. The stuff worth knowing, minus the doomscrolling.

Issues are curated by ex-Google and Anthropic engineers and land in your inbox before your morning coffee. A 5-minute read, and you walk into the day already knowing what your team is still catching up on.

Tech is just the start. We also cover AI, marketing, dev, and more. Pick the briefs that match your work.

Free, daily, and read by 7M+ subscribers. Subscribe and let the experts do the digging for the tech news that matters.

Subscribe for Free

America's power companies just unveiled a $1.4 trillion spending plan through 2030, a 27% jump in a single year.

The reason is almost entirely one thing: feeding AI's data centers.

Here's why that number is so big.

A modern AI data center doesn't sip power.

It drinks like a small city. Add them all up and data centers are on track to burn roughly 1,050 terawatt-hours a year.

Put another way: if the world's data centers were their own country, they'd be the 5th-largest electricity user on Earth, sitting between Japan and Russia.

That's the power bill for the AI boom.

And here's the catch nobody saw coming.

You can build a data center in about a year. But plugging it into the grid now takes multiple years.

The waiting line for a new connection has stretched so long it's become the real bottleneck.

So the tech giants are done waiting.

They're building their own power.

Oracle just expanded a deal with Bloom Energy to install up to 2.8 gigawatts of on-site fuel cells, power boxes that skip the grid entirely.

Bloom's stock jumped 15% on the news.

And they're going nuclear.

Meta signed deals for up to 6.6 gigawatts of nuclear power across four companies.

As of this year, every single major tech giant (Amazon, Microsoft, Google, Meta) has now signed at least one nuclear deal to feed AI.

(A gigawatt → powers a million homes)

The AI trade quietly stopped being about chips and became about electricity, and almost nobody has repriced for it.

For two years, the entire AI story fit on a bumper sticker: buy Nvidia, buy the chipmakers, ride the boom.

That worked. But a boom always runs into a wall, and this one just found its wall. Power.

Think about the chain: a chip is worthless until it's plugged in.

And right now the plug is the scarce part: the line to connect a new data center to the grid runs years long, while the demand for compute grows by the month.

When the bottleneck moves, the money moves with it.

AI demand is being routed into the energy sector, a slower, stickier, far less crowded corner of the market than semiconductors.

And it shows up in layers, most of them one step downstream of the names everyone already owns.

On-site power, the fuel cells that let a data center skip the grid, is why a sleepy name like $BE ( ▼ 6.43% ) popped 15% on a single Oracle order.

Then the nuclear operators already running plants today: $CEG ( ▲ 1.16% ) and $VST ( ▼ 1.38% ), now signing 20-year deals to sell electrons to hyperscalers.

Then the new-build bets: small modular reactors from $OKLO ( ▼ 0.17% ) and the turbines and grid gear from $GEV ( ▼ 1.87% ), the unglamorous picks-and-shovels of the power buildout.

When everyone crowds the obvious winner, look one step downstream at what that winner can't function without.

Nvidia can't sell a single chip that no one can power.

We've been calling power the hidden tax on AI for a while now.

This is the month the bill came due.

Know someone who only owns Nvidia and thinks that's the whole AI trade? Forward this to them →

Second Order is for information and entertainment only and is not financial advice.

We are not licensed financial advisors.

Any tickers or companies mentioned are our opinions, not recommendations to buy or sell.

Do your own research and consult a professional before making investment decisions.