• THE GRID
181 Gigawatts of Gas. That's What "Speed to Power" Really Means.
One hundred eighty-one gigawatts. I read that number twice when BloombergNEF published its new data center report on Wednesday. That's how much gas-fired power has been proposed to sit right next to data centers. Not on the grid. Next to it. Private plants, on private land, burning gas to run AI.
More than 90% of those plants are in the U.S. That's 165 gigawatts of gas on American soil. Not for cities. Not for factories. For AI training clusters and inference farms.
"The extraordinary economics around AI are making speed to power one of the most important priorities for defining constraints on data center developers."
Speed to power. That's the phrase worth circling. Grid hookups take years. Permits stack up. Transmission lines don't exist yet. Utilities can't move fast enough for a business that doubles in size every twelve months. So developers skip all of it. They build gas plants. Fast and dirty.
If all those U.S. plants get built and run at 60% use, BNEF says they'd spew about 345 million tonnes of CO₂ per year. Compared to drawing from the grid, those on-site plants would add 152 million tonnes of extra emissions. That would raise U.S. power-sector emissions by roughly 10%.
On-Site Gas
181 GW
| U.S. Share
165 GW
|
DC Demand by 2035
2,354 TWh
| Extra Emissions
152M tonnes
|
A Drone Expert Is Making Flying Cars Real
In 2016, after years working on military drones, Doron Merdinger began sketching a flying car. He saw how quickly the tech had advanced. Soon, people would fly from driveways. So he built Doroni.
It’s great timing, too. The urban air mobility market is worth $4.8B today. By 2040, Morgan Stanley puts it at $1 trillion. By 2050, $9 trillion.
And with 600+ reservations for Doroni’s aircraft and $240M+ in potential revenue already, the demand for this innovation is undeniable.
Now, after unveiling the full-sized showroom prototype of their H1-X personal aircraft in March, they’re preparing to start test flights with the FAA next year.
By 2028, they’re targeting the first commercial deliveries. By 2032, they’ve got a $1.4B annual revenue target. And you can claim an early-stage stake in Doroni. Invest in Doroni at $3.72/share today.
BNEF sees total data center power demand nearly quadrupling to 2,354 TWh by 2035. That's 7% of all the electricity on Earth. More than Japan uses today. More than India used five years ago.
Developers say these plants are temporary. Bridge power until the grid catches up. But gas plants have 20-year lifespans. Pipelines get built around them. Temporary has a way of becoming permanent.
The math is plain. AI makes money fast. Gas plants get built fast. The grid does not move fast. So the gap gets filled with fuel.
And that creates a three-front problem. Voters are mad about data centers near their homes. Regulators are cracking down on rate hikes. And now climate groups have their own number to wave around: 181 gigawatts of new fossil fuel, all for Big Tech.
The AI boom was always going to be an energy story. Now it's an emissions story. And that changes the politics of every permit in the pipeline.
Your Download Link Will Expire
If you still haven't downloaded my free "Safe Trade Options Formula" guide...
...please take a few seconds and download it right now before your new temporary download link expires.
I eventually plan to charge money for this training, so do yourself a favor and download it now...
That way, no matter what it costs in the future, you'll have a free copy on your computer.
Make sense?
• VOLTAGE
North America's First SMR Is Already Late and Over Budget
Seven point seven billion Canadian dollars. That was the price tag for the first BWRX-300 reactor at Darlington, just east of Toronto. Was.
Ontario Power Generation said Tuesday it will miss both the C$7.7 billion budget and the October 2030 target date for its first unit. OPG didn't say by how much. Just that it won't make it.
This is the reactor the nuclear bulls have been watching. The BWRX-300, built by GE Vernova and Hitachi, is the design behind the whole Western SMR push. It's the one Ontario bet on. It's the one X-energy's rivals measure against.
The full project calls for four reactors at a total cost of C$20.9 billion. If Unit 1 is already over, the others will be, too.
I'm not against nuclear. Not at all. But the SMR thesis runs on one idea: that small reactors are cheap and fast enough to compete with gas. Darlington is the first real test. And the first test is failing on both counts.
This matters beyond Canada. U.S. companies like X-energy, NuScale, and Oklo all need investors to believe the next build goes smoother. Every cost overrun at Darlington makes that pitch harder. It feeds the old fear: nuclear always runs late and over budget.
OPG says it will post updated costs and timelines early next year. Watch those numbers closely.
The SMR story isn't dead. But it just got a lot harder to sell on schedule and on budget.
• WIRED IN
Signals From the Grid
Follow the money this week and it lands on three names. A utility. A chipmaker. A cloud builder. Each one just told us where power spending goes next.
DTE Energy boosted its five-year utility plan to $44 billion — a $10 billion jump. The driver? A locked-in 2.4 GW of data center load from Oracle and Google in Michigan. DTE also moved to sell its non-utility Vantage unit for about $3 billion, sharpening its focus on the regulated grid side. Every dollar from the sale goes back to wires and poles.
Meanwhile, the companies that make the gear inside those data centers are surging. Infineon reported fiscal 2026 data center power revenue of €2.2 billion — then said fiscal 2027 revenue guidance is close to €4 billion. That's an 82% jump in a single year. Factory output, not demand, is now the ceiling on growth.
CoreWeave posted Q3 revenue of $3.7 billion, up 168% year over year. It raised full-year 2026 guidance to $13–$13.3 billion and now expects more than 1.85 GW of active power by year-end. Capex guidance ticked up too: $37–$39 billion. This company alone is spending more on power than most mid-sized utilities.
Utilities are raising capex. Chip makers can't build fast enough. And the hyperscalers keep writing bigger checks. The capital cycle is still accelerating.
*Disclaimer: This is a paid advertisement for Doroni Regulation A offering. Please read the offering circular at invest.doroni.io
Rockets, EVs and humanoid robots all need critical metals. One relatively unknown company trading around $5 could be positioned to benefit. Get the free report and the ticker by text.
