Private QE is here—and it’s not coming from the Fed.

It’s coming from the balance sheets of Microsoft, Amazon, Alphabet, Meta and friends, in the form of hundreds of billions of capex a year being firehosed into a few narrow sectors.

Think data centers. AI chips. Power. Metals.

Not via banks. Not via bond markets. Just direct spending that behaves a lot like a new money printer for whoever sits closest to the flow.

That’s the new Cantillon Effect: the people and sectors closest to this “Private QE” often see the upside first. Everyone else just sees higher prices and wonders why their “cheap” stock won’t move.

Over the next few years, the hyperscalers are set to act like mini-central banks.

Street estimates from Goldman Sachs and Morgan Stanley say combined capex for Microsoft, Amazon, Alphabet, Meta and others will run around $500–$530B in 2026—roughly the size of Austria’s entire economy.

That 2026 forecast is up from earlier projections of about $314B, implying almost 29% year‑over‑year growth in spend as the AI arms race accelerates (Janus Henderson).

Where is that money going?

That spend doesn’t trickle in—it hits as large, long-dated contracts.

If you’re a supplier, that can feel better than QE. You’re getting committed revenue, not just lower interest rates.

The Cantillon Effect is simple:

Whoever touches new money first tends to win. Everyone else gets higher costs later.

Here, “new money” = Big Tech capex.

In this setup, markets often trade as if you’re not just betting on earnings anymore—you’re betting on who’s closest to the tap.

AI clusters need insane amounts of stable power. “Sometimes-on” wind and solar won’t cut it for 24/7 GPU farms.

That’s why you’re seeing deals like Microsoft’s move to revive capacity at Three Mile Island via a 20‑year power purchase agreement to restart Unit 1 as the 835‑MW Crane Clean Energy Center, targeting a 2028 restart.

Constellation Energy’s stock jumped roughly 22% on the announcement day and more than 30% for the month, and by early 2026 it was trading in the $330–$350 range versus ~$250 pre‑deal—an example of what can happen when a company suddenly finds itself plugged into Private QE (Motley Fool / Macrotrends).

Then there’s copper.

High-density AI data centers need far more wiring, cabling, and cooling infrastructure than legacy setups. That’s copper-heavy.

Commodity trading giant Trafigura estimates that AI and data centers alone could add about 1 million metric tons of copper demand per year by 2030 (Trafigura).

Case studies already show a single Microsoft hyperscale site in Chicago consuming 2,177 tonnes of copper for its buildout, highlighting how metal‑intensive this infrastructure really is (Argentina Lithium research).

As hyperscalers lock up supply:

Again: the people nearest the new spend tend to win. Late receivers just see margin pressure.

Macro veterans have been saying the quiet part out loud: liquidity drives markets more than fundamentals.

As Stanley Druckenmiller has argued for years—most famously in a 2015 talk at the Lost Tree Club—“Earnings don’t move the overall market; it’s the Federal Reserve Board… It’s liquidity that moves markets.”

On January 14, 2026, the Federal Reserve officially ended Quantitative Tightening and pivoted to an “ample reserves” technical expansion regime, using Reserve Management Purchases to keep bank reserves from falling below roughly $2.9T and to offset the collapse of the Fed’s overnight reverse repo facility to about $6B.

In practice, the New York Fed is now a steady buyer of around $40B/month in short‑term Treasuries to maintain liquidity (SVB / Fed).

That balance sheet shift came on top of three straight 25 bps rate cuts in Q4 2025, which brought the funds rate down to 3.50–3.75% (Great Liquidity Pivot).

Translation: the Fed stopped draining the pool and quietly turned the tap back on.

Big Tech then stepped in and turned on its own hose—this “Private QE”:

If you’re not tracking where that hose is pointed, it can feel like flying blind.

You’ll hear: “This is just the dot-com bubble all over again.”

The comparison:

Fair. But there are key differences:

In plain English: Microsoft, Amazon, and Meta can, if they choose, spend on capex for much longer than most skeptics can stay short.

This isn’t a “buy X, sell Y” note. It’s a lens.

Here are a few ways people are starting to use it:

If there’s only one habit to experiment with: skim capex and contract announcements before you read the earnings commentary.

Which sector do you think benefits most from this new “Private QE”?

Hit reply with your pick (and why) and I’ll feature a few in the next issue.

If this helped you see markets less as “stock stories” and more as liquidity maps, forward it to a friend who’s still hunting for the next meme stock.

They don’t necessarily need another ticker. They might just need a new lens.