FOR PEOPLE WHO WANT TO SEE WHAT BREAKS BEFORE IT BREAKS
Data centers are straining a grid already mid-upgrade, and private capital is moving to own the layer that AI cannot run without.
THE SETUP
The AI buildout is not only a chip story. It is a power story. Data centers need steady power that most regional grids were not built to deliver at this pace. Utilities were already mid-cycle on upgrades long before AI demand arrived. AI demand is landing on top of that pressure. The result is a fight for power assets and grid connections. Who owns the power stack may matter as much as who writes the code.
PMD Lens
The market is not just pricing in AI software. It is starting to price in the physical layer beneath it. Power now limits how fast AI can scale, and private capital is moving upstream to buy the utilities that run data centers. This is a structural shift in where returns get captured, not a one-cycle trade.
WHAT MOST PEOPLE WILL MISS
The AI boom is not only increasing demand for semiconductors and compute. It is also increasing demand for electricity infrastructure.
Power grids in many regions were already facing major upgrade cycles due to aging equipment and extreme weather risks.
AI-driven electricity demand is arriving on top of those existing pressures, accelerating the need for new capital investment across utilities and transmission networks.
The deeper implication is that electricity access and infrastructure financing may become the limiting factor for how quickly AI infrastructure can expand.
PREMIER FEATURE
Great Companies Don’t Stay Under the Radar Forever
Most great stocks look boring when the real opportunity begins.
They’re still building scale, executing quietly, and being ignored by most investors.
The original market leaders didn’t become obvious overnight.
Our analysts believe the same pattern is forming again.
In These 7 Stocks Will Be Magnificent in 2026, we highlight companies that may look unremarkable today…
But you’ll soon see they all have the traits that historically define future market leaders.
SIGNALS IN MOTION
The signals below are not forecasts. They are mechanisms already in motion. Each one reveals the same pattern: duration is being financed before economics are fully proven.
Signal 1: Rising Bills Reflect Older Problems, Not Just New Demand
American electricity bills have risen nearly 30% since 2020, with AI often blamed. The drivers are older. In many states, rate growth trailed inflation, while spikes in the Northeast and California came from wildfire mitigation and wholesale market shifts. Utilities were already replacing transformers and upgrading transmission before data center demand accelerated. AI expanded an existing investment cycle rather than starting it.
Investor Signal
The electricity system is entering a multi-decade upgrade cycle. Capital will flow toward grid upgrades and new capacity regardless of AI demand growth.
Signal 2: xAI Starts Building Its Own Power Supply
xAI has been operating 27 gas turbines at its Colossus 2 site in Mississippi, producing roughly 495 megawatts of electricity. Regulators are now considering permits for 41 permanent turbines. The significance is strategic: when grid supply lags demand, hyperscalers begin building generation themselves rather than waiting for utilities.
Investor Signal
When hyperscalers build power plants, energy has become a structural constraint. Control of power supply now shapes build pace and competitive position in AI.
FROM OUR PARTNERS
Central Banks Are Lying About Gold
Jerome Powell says gold isn’t money. The Fed says inflation is under control.
Last year, they bought more gold than at any time since 1967. China dumped $100B in U.S. debt, then bought gold. Poland, Hungary, Singapore, Turkey… all loading up.
This isn’t a trend. It’s a panic.
After the U.S. froze Russia’s assets, the world learned a hard lesson: there’s only one asset no one can freeze.
Gold.
I’ve just released an urgent report on one stock positioned to benefit as this rush accelerates.
Signal 3: Rate Uncertainty Clouds Infrastructure Financing
Infrastructure valuations depend heavily on long-term interest rates. Oil moving above $100 has complicated expectations for central bank policy, with markets pricing divergence between the Federal Reserve and the European Central Bank. Energy shocks typically hit Europe harder as a net importer. If that divergence breaks, rate expectations shift.
Utilities and infrastructure assets behave like long-duration investments, making them sensitive to financing costs.
Investor Signal
If markets misprice the rate path, infrastructure deals built on stable long-term costs carry hidden risk. Stress-test duration in any active position.
DEEP DIVE
Private Equity Moves Upstream Into the Grid
The New Acquisition Target
Private equity has spent the last decade buying data centers. It is now buying the utilities that power them. PitchBook counts four of the seven largest PE power generation buyouts in the past three years. A fifth, Stonepeak's $5 billion Cleco deal, is pending. BlackRock's Global Infrastructure Partners, Blackstone, and JPMorgan are all acquiring generation assets. The IEA forecasts data center power demand will double by 2030, four times faster than any other sector. That demand is driving the deals.
The Regulatory Fault Line
Utilities are not like other infrastructure assets. They operate inside frameworks designed to protect ratepayers, not maximize returns. Blackstone's $11.5 billion TXNM bid has drawn opposition in New Mexico. Consumer groups argue PE ownership extracts returns from regulated assets. The deal requires approval from the state commission and several federal bodies.
State utility commissions are designed to protect affordability. Regulators already rejected a prior TXNM bid from Avangrid in 2024 after concluding it would raise consumer costs. That decision may preview future resistance.
Technology companies are approaching from the other direction. Alphabet bought Intersect Power for $4.75 billion to bring clean power in-house, while xAI is building gas plants in Mississippi. Private equity and hyperscalers now compete for the same scarce input: reliable power near compute clusters.
The Cascade
AI demand rises. Data centers compete for power and grid connections. Private capital buys generation assets near data hubs. Prices rise. Political pressure follows, and regulators step in.
Returns depend on pricing power, and pricing power depends on regulatory tolerance. That tolerance depends on how much cost consumers will bear. Utilities operate under rules that move slowly. The decision in New Mexico may signal how far PE can push before commissions push back. Texas, Virginia, and the Midwest are next.
The upside is real. So is the exit risk if the regulatory framework tightens.
Investor Signal
Private equity is moving into electricity infrastructure to capture AI-driven demand. The return thesis depends on regulatory tolerance already under pressure. Underwrite this with political risk fully priced in.
FROM OUR PARTNERS
The New #1 Stock in the World?
A tiny company now holds 250 patents tied to what some call the most important tech breakthrough since the silicon chip in 1958.
Using this technology, it just set a new world speed record, pushing the limits of next-generation electronics.
Nvidia has already partnered with this firm to bring its tech into advanced AI systems.
This little-known company could soon become impossible to ignore.
THE PLAYBOOK
Track electricity availability near major data center hubs. Power access may determine where AI infrastructure expands next.
Monitor private equity acquisitions of utilities and power generation assets tied to data center regions.
Watch for regulatory pushback as electricity demand and prices rise.
Evaluate infrastructure investments that combine stable regulated cash flows with exposure to rising electricity demand.
THE PMD REPOSITION
The AI capital cycle is expanding beyond software and semiconductors into the physical infrastructure that powers computation. Electricity systems are emerging as a strategic layer of the technology economy. Private capital is increasingly competing with utilities and technology companies for control of these assets. The next phase of the AI investment cycle may be defined not by algorithms, but by who controls the electricity required to run them.
