• THE GRID
Ninety-Seven Gigawatts of Hunger, Thirty-Two Gigawatts of Air
Ninety-seven gigawatts. I read that number three times.
That's Morgan Stanley's new estimate for total U.S. data center power demand through 2028. Their September 21 research note, "Powering AI," hiked the figure 43% from a prior estimate of 68 GW.
Then they did the math on supply. They counted every behind-the-meter gas plant. Every fuel cell project. Roughly 19 GW of spare grid capacity.
And after all of that... they still came up 32 GW short.
Thirty-two gigawatts of nothing. More than New York City's entire grid. Just gone.
MS Demand Est.
97 GW
| Net Shortfall
32 GW
|
Meta Q3 Capex
$34.9B
| Meta FCF Q3
−$2.1B
|
Forget SpaceX. Elon's next BIG bet is a radical "light-speed" device that's turning AI into "Accelerated AI" – and making it 100 times faster and 100 times more energy efficient.
The mainstream hasn't caught on yet. But "Accelerated AI" stocks are breaking out as we speak: 133%, 217%, or even 320% or more… and it's just getting started.
A year ago, the bottleneck was chips. GPU shortages. HBM allocation. Packaging capacity at TSMC. Those problems got solved. Or at least got better.
Now the constraint is physical. It's copper and concrete. Substations and transformers.
You can't download a substation. You can't ship a transmission line overnight.
And the cash keeps coming. Meta spent $34.9 billion on data center infrastructure last quarter alone. That's more than NASA's entire annual budget — in a single quarter.
Full-year 2026 capex is guided at $135 billion to $145 billion. CFO Susan Li told analysts that Meta is still developing its 2027 budget. Over at Alphabet, CFO Anat Ashkenazi told analysts that 2027 spending will "significantly increase" from there.
Meta's free cash flow flipped to negative $2.1 billion in Q3. The company doesn't care. It's racing to build the compute for its Muse AI platform. Profit can wait. Capacity can't.
The money is there. The watts are not.
"The upcoming infrastructure CapEx cycle will create islands of wealth, and literal power."
What Byrd means is straightforward. The companies that lock up power win. Everyone else waits.
It used to be about who had the best chips. Now it's about who has the best extension cord.
Morgan Stanley says this gap will favor deep-pocketed tech giants. Smaller players get squeezed out.
If you can't get power, you can't train models. If you can't train, you can't compete. The gap is a moat — and only the biggest wallets can cross it.
The EIA says U.S. power use will hit record highs in both 2026 and 2027. Gas-fired capacity planned for data centers nearly doubled in H1 this year, to 189 GW. The grid still can't keep up.
The AI race is no longer a chip race. It's a power race. And right now, the grid is losing.
Elon Musk’s Next Partnership: Buy This Stock Now
In April 2026, Elon Musk announced he was partnering with Intel.
The result?
Shares took off…
And were trading 177% higher just two months later.
I believe Elon is about to announce a NEW partnership with this little-known company…
To work on what he called "the most epic AI project in history."
But you need to act fast…
Because Elon has already approached this company.
I believe he's going to make it official by October 30…
Which should send shares higher than anyone can imagine.
• VOLTAGE
Tesla's Battery Factory Can't Build Fast Enough
Elon Musk said something on last week's earnings call that didn't get enough press.
"We're actually almost sold out of all Megapacks for 2027, and we're ramping production as quickly as possible to meet demand."
Read that again. It's October 2026. Tesla has nearly sold every utility-scale battery it can build through the end of next year. Demand is outpacing a company that runs three Megafactories on two continents.
The Q3 numbers back it up. Tesla deployed 13.7 GWh of energy storage — up 10% from a year ago. The energy segment posted $2.88 billion in revenue.
And gross margins jumped to 33.1%, up from 20.4% just one quarter earlier. That's a margin swing most hardware firms only dream about.
Two things drove it. First, Tesla's new Megafactory near Houston began production in Q3. That adds to plants in Lathrop, California, and Shanghai. Second, Tesla shipped its first Megapacks built with LFP battery cells made in Nevada — a push to bring the supply chain home.
Who's buying all these Megapacks? Utilities adding storage as data center load surges. Grid operators who need peaking power fast. And data center developers who want backup on-site.
Tesla's Q3 Update says energy storage demand is "increasingly aligned with SI infrastructure demand."
Revenue dipped 16% year over year — Tesla says tariff-related customer refunds were partially the cause. But the margin jump and record backlog tell the real story. This is a supply problem, not a demand problem.
When the world's biggest battery maker says it's sold out, the grid's power shortage isn't a forecast anymore. It's a fact.
• WIRED IN
Signals From the Wire
Trump says data centers must make their own power. On October 21, the president posted that new data centers are "mandated by me to make their own electricity." He also called for a "dividend" to local communities. No executive order has followed. But the signal is clear — the White House now backs the behind-the-meter trend already in motion. Gas-to-data-center pipeline builders like Williams and Enbridge just got political cover. Watch for utilities to push back.
Hitachi Energy and Ansaldo Energia signed an MoU on October 16. The deal bundles Ansaldo's gas turbines with Hitachi's switchgear and transformers into one turnkey package. The pitch: one-stop power for data center developers who can't wait years for a grid hookup. This is two industrial giants betting that the build-your-own-plant model isn't a phase. It's the new default.
Vistra fell 9.81% on October 22 — its worst day in months. Meta's Q3 report showed AI spending surging but cash flow going negative. Morgan Stanley trimmed its VST target from $227 to $226. The stock is now down 28% over the past year. The irony is thick — AI power demand is supposed to be the bull case for generators like Vistra. But investors fear the capex flood won't turn into the long-term power contracts Vistra needs. Cash is flowing to build data centers. It's not yet flowing to pay for the power.
The pattern across all three signals is the same: the money wants power, but power doesn't want to wait for the grid.
The best sector of 2026 isn't tech. It's energy, up around 40% this year while most investors chased chip stocks. Get the #1 energy stock free, and we'll text you the name. Show Me the Energy Stock.
