The Magnificent Seven (aka 7 biggest tech companies), just put up one of the best earnings quarters in two decades.

Revenue beats across the board.
EPS surprises that would've triggered rallies in any normal cycle.

Then Meta told Wall Street it was raising 2026 capex guidance to $125–$145 billion… and the stock dropped 6% after hours.

The numbers were not the problem. The check was.

Alphabet printed $109.9 billion in Q1 revenue, beating consensus by nearly $3 billion.

EPS came in at $5.11 versus a $2.62 estimate.
That's an 82% YoY jump.

Google Cloud grew 63% to $20 billion, accelerating from 48% in Q4.

Microsoft delivered $82.9 billion in revenue, up 18%, with its AI business running at an annualized $37 billion, up 123% YoY.

Amazon did $181.5 billion with a record 13.1% operating margin.

AMD's data center segment hit $5.8 billion, up 57%, and the stock ripped 16% the day after.

By any historical standard, this was a blowout.

For five years the Mag 7 ran on a simple story:
Print free cash flow, return it to shareholders.

That story is dead.

Amazon has committed roughly $200 billion in 2026 capex.
Alphabet raised guidance to $180–$190 billion.
Microsoft flagged $190 billion, $25 billion in component cost inflation.
Meta raised by $10 billion on both ends.

Combined, the four hyperscalers are on track to spend $650 to $700 billion this year, which is simply more than the GDP of most European countries.

Amazon's free cash flow has already collapsed to $1.2 billion trailing twelve months. That’s a 95% decline YoY.

Meta's will be cut roughly in half.

The new model:

Spend like a utility and hope revenue catches up before the buyback budget dries up.

And it might.

Microsoft's AI revenue tripled.
Google Cloud is accelerating.
AMD's data center business outgrew the rest of the company.

It's not an irrational bet. It's just not free anymore.

The Mag 7 now represents roughly 25% of S&P 500 market cap.

Six straight winning weeks for the index, the longest streak since 2024.
But the 10-year Treasury closed May 8 at 4.38% and the 30-year is at 4.95%.

Bonds are pricing in either persistent inflation, persistent issuance, or both.
Neither helps high-multiple tech.

If AI monetization stalls even one or two quarters, the index-level concentration means the whole market goes with it.

When everyone owns the same trade, the trade is fragile…
And right now, the S&P 500 is structurally a leveraged bet on whether $650 billion in capex starts paying back on schedule.

Equal-weight S&P indices have been quietly outperforming since April.

That's not a coincidence.

read more

An OTC telecom company it's on track for $450 million in revenue this year. Last year it did $3 million. The math in between involves three acquisitions, a NASDAQ uplisting plan, and 500 patent filings.

Warren Buffett compared markets to a church with a casino attached. Greg Abel ran his first annual meeting as CEO. Buffett said Americans have never been in more of a gambling mood than right now.

A $2 trillion corner of finance just got its first real warning label. The Financial Stability Board flagged private credit for high leverage, opaque valuations, and zero stress-testing in a downturn.

What did The White House say about Trump's drug pricing deals? Seventeen pharma companies have signed on. Few details are public.

Everyone's piling into AI chips, but the real bottleneck is electricity.

And the pipeline of signed offtake agreements between data center operators and small modular reactor projects has nearly doubled — from 25 gigawatts at end of 2024 to 45 gigawatts today.

Those aren't forecasts.
Those are real deals between those companies.

The International Energy Agency (IEA) flagged data center growth as a primary driver of global electricity demand for the first time.

Data center electricity use surged 17% in 2025, and AI-focused facilities grew even faster.

Hyperscaler capex surged past $400B last year and is set to jump another 75% in 2026.

Tech companies now account for 40% of all corporate renewable power purchase agreements.

But renewables alone can't deliver baseload at the scale these facilities need.

The NuScale Power Corporation (SMR) pipeline tells the story:
Conditional nuclear offtake deals have nearly doubled in 18 months.

Constellation Energy issued 2026 EPS guidance of $11–$12, raised its buyback to $5 billion, and earmarked $3.9 billion for growth capex.

Cameco, the largest pure-play uranium producer, cited rising electricity demand directly in its latest earnings call.

Vistra, Talen Energy, and GE Vernova — key players in the energy sector — have all been running.

Sovereign wealth funds and pensions have been quietly building positions in nuclear-adjacent names since Q4 2024.

Independent power producers with nuclear assets, uranium miners, SMR developers, and grid infrastructure names are all catching institutional flows.

The AI capex story gets the headlines.
But the power story — one layer down — doesn't depend on AI revenue ever monetizing.

The data centers get built either way. The grid build-out happens either way.

Even if half the AI capex turns out to be malinvestment, the infrastructure names still print.

Because the electricity demand is already real and already contracted.

read more

Friday's jobs report looked clean at the top line.

Payrolls up 115,000 versus 62,000 expected.
Unemployment held at 4.3%.

Wall Street's take: economy still chugging.

Buried deep in the Bureau of Labor Statistics (BLS) release: information services lost another 13,000 jobs in April.

That sector has now shed 342,000 positions since November 2022.
That's an 11% decline over 41 months without a single positive quarter.

A giant recession running in parallel with all-time-high stock prices in the same industry.

Software publishing, data processing, telecom, broadcasting, cloud infrastructure, web search portals. Every line item is bleeding.

Telecom lost 3,000 jobs in April alone.

Motion picture and sound recording dropped 6,000.
Computing infrastructure and data processing shed another 4,000.

AI is the obvious culprit, but the cuts started before AI was profitable.

The layoff announcements at Microsoft, Google, Meta, and Salesforce ($650 billion in AI capex) cut the people who built the systems AI now replaces.

One quarterly earnings call talks about AI tailwinds.
The 10-K talks about operational efficiency.

Both are true.

445,000 people moved from full-time to part-time work in a single month because their hours were cut or they couldn't find full-time positions.

Involuntary underemployment jumped 10% in 30 days.

The real unemployment rate, including discouraged and underemployed workers, rose to 8.2%.

Average hourly earnings are up 3.6% YoY.

Year-ahead inflation expectations are at 4.5%.

That gap — negative 0.9 percentage points in real terms — is the cleanest explanation for why consumer sentiment just hit 48.2, the lowest reading in the University of Michigan survey's 73-year history.

People are just doing the math.

The official payroll numbers and the sentiment surveys have diverged this sharply maybe three times in the post-war era.

Defensive consumer staples, dollar stores, and beaten-up healthcare names have all seen rotation since April.

Short-duration Treasuries are paying 3.9% with no equity risk.

The money is already moving…

The question is whether the hard data catches up to what households have been saying for months.

But data isn't waiting for anyone to say it out loud.

read more

(by feargreedmeter)

Source: Stock Analysis