• THE GRID
Old Reactors, New Megawatts
October 6 was a good day for Constellation Energy. The stock surged 12.2%.
No earnings beat. No buyback. Just one deal — with Google.
Google signed a 20-year deal to buy 890 MW of new nuclear power. Not from a new reactor. From upgrades to 11 reactors across 6 plants in three states.
Those plants already run. They already feed PJM's grid.
Constellation will spend more than $4.3 billion on new turbines and better steam gear. The goal: squeeze more power from each core.
The first new megawatts arrive by 2028.
| New Nuclear Capacity 890 MW | CEG Investment $4.3B | |
| Reactors Upgraded 11 | First Uprate Online 2028 | |
Trump Clears Way for New ‘Bank of Musk’
Something strange is happening. A few of the most powerful people in Congress are scrambling to stop Elon Musk’s latest innovation. Led by Senator Elizabeth Warren, who is fighting it tooth and nail. Meanwhile, Musk is rolling it out to millions of Americans… with the full blessing of President Trump and the U.S. Treasury. And starting January 18, he could take it one step further... issuing his own digital dollars. When that happens, his backers could make a fortune.
That same Monday, the DOE placed its own bet. It offered Vistra a loan of up to $4.2 billion.
The targets: Perry and Davis-Besse in Ohio, Beaver Valley in PA. Those three plants back Meta's 20-year deal for 2,609 MW of nuclear power.
Two weeks earlier, Google cut a separate deal with Georgia Power. Upgrade parts of Hatch and Vogtle. Add 96 MW to the grid.
Three deals. Two weeks. Over $8.5 billion in capital.
Nearly 1,000 MW of new clean firm power — all from plants already running.
Constellation closed at $300 after the Google deal. Vistra jumped 10.8%. Talen Energy rose 12.4%. If you own nuclear plants in PJM right now, your phone is ringing.
This is Big Tech's shortcut around the grid queue.
Building a new reactor takes a decade. Linking a new project to PJM can take just as long. An uprate skips both steps.
The site exists. The wires exist. The license exists.
You pay engineers to run the plant harder. Power arrives in two to four years — not ten.
The three biggest data center power buyers — Google, Meta, and Amazon — all signed uprate deals this year. They're not waiting for new tech. They're tapping plants that have run for decades. And the market took note.
One analyst put it well: 890 MW of uprates "is roughly equal to building a large new reactor." But the wait is half as long. Or less.
I think this is the most important power story of the month. Uprates don't grab headlines. They're less flashy than an SMR pitch or a fusion bet. But they put real watts on a grid that's straining right now.
The fastest path to new nuclear power doesn't start with a new design. It starts with an old reactor and a fat check.
I'm Giving This Away (Seriously)
Quick question...
Would you spend 60 seconds to save yourself $29.97?
That's what my "Safe Trade Options Formula" book costs on our website right now. But today, I'm giving it away — no charge, no strings.
Inside, you'll discover a 5-part trade that makes every trade... a SAFE TRADE.
It's the same book that sells for $29.97 on our website right now, but is yours on the house today.
All you have to do is click here and tell me where to send it.
p.s. This won't stay free forever — $29.97 is a fair price and I'll be going back to it soon.
Grab your copy here before that happens.
• WIRED IN
Four Signals From the Power Trade
Schneider Electric agreed to buy industrial software maker PTC for $22.6 billion — its biggest deal ever, at a 42% premium. PTC builds tools for product design and factory work used by big firms across many fields. Schneider wants to merge PTC with its own energy platform — one system from data center blueprint to power bill. The deal is expected to close by Q3 2027.
While Schneider builds through acquisition, Bloom Energy keeps climbing on fuel cell demand. Shares hit $280.50 on October 9 — up 223% in 2026. Brookfield expanded its data center financing with Bloom to $25 billion. Those fuel cells ship faster than any new turbine or reactor. Q3 earnings land on October 22 — I'll be watching margins closely.
Not every energy play is working. NuScale Power — the biggest pure-play SMR stock — has crashed 49% to $7.18 this year, now down 87% from its 52-week high. It posted $75,000 in quarterly sales, filed to sell $750 million in new shares, and faces a Pomerantz fraud probe. UBS downgraded the stock and RBC slashed its target to $7.
The pain spreads beyond SMRs to the broader utility sector. 24/7 Wall St. called utility stocks "the most oversold since 2023." Yet EIA still projects record US power demand in both 2026 and 2027. Oversold stocks plus rising demand — that mismatch is what I'm watching into earnings season.
Money is chasing companies that ship power now. The blueprints can wait.
• VOLTAGE
Bigger Than Housing
A number crossed my screen last month. I almost scrolled past it.
In Q2 2026, US spending on data centers and computing hardware passed home building. First time in the history of the BEA's data.
Fortune broke the story on September 20 using BEA data. America now puts more cash into server farms than into homes.
Housing drove this economy for decades. It built suburbs. It fueled bank lending. It shaped where power lines ran.
Now data centers have taken that role.
Goldman Sachs sees Big Tech capex reaching $800 billion this year. That's well above the $540 billion the IEA says the world needs each year just to keep oil and gas flowing.
Every dollar of that pulls power. More servers mean more racks. More racks mean more cooling. More cooling means more watts.
The EIA projects record US demand of 4,288 billion kWh this year.
Rystad found that global data center capex hit $770 billion in 2025 — more than the world spent on oil and gas drilling. The 2026 numbers will be higher. Morgan Lewis called this shift "a new US economic milestone" just last week.
A house doesn't need a substation. A 500 MW data center campus does. That's the difference this shift brings to the grid.
I'm not calling a crash. The AI demand is real. But when one sector's spending tops housing in the US economy, the grid becomes the load-bearing wall beneath all of it.
If power can't keep pace with capital, the capital gets stranded. That's the risk hiding in plain sight.
