Eighty seconds. That is how long it took on Wednesday night for Microsoft, Alphabet, Meta, and Amazon to all report Q1 earnings. Every Bloomberg terminal in the world was frozen on the same chart. And in those eighty seconds, the AI trade that has carried the market for two years quietly broke into two pieces.

Alphabet shot up 6% in after-hours.

Meta fell more than 5%.

Microsoft dropped about 3%.

Amazon dropped about 3%.

All four beat headline estimates. All four raised AI capex. All four said demand is strong. And the market split them down the middle. That split is the story of 2026.

The Numbers That Actually Matter

Forget EPS for a second. The four companies just collectively told you they will spend close to $700 billion on AI infrastructure this year alone. Not over a decade. In twelve months.

Microsoft: $190B in CY2026 capex (including $25B from higher component pricing — read: they are paying up just to secure parts)
Alphabet: $175-185B
Meta: $125-145B (raised AGAIN from the $115-135B range they gave a quarter ago)
Amazon: ~$200B

That is bigger than the GDP of Switzerland. It is bigger than every dollar venture capital has ever deployed in any single year on record. And it is being committed by four companies whose boards apparently looked at “OpenAI may not be able to pay its compute bills” and answered with “build more anyway.”

The market is not punishing the spend. The market is punishing the lack of proof.

Why Alphabet Won and Meta Lost

Alphabet beat on revenue, beat on margin, raised capex, and the stock went up 6%. Why? Because Search ad revenue accelerated. Cloud revenue accelerated. YouTube accelerated. Every dollar of AI capex showed up as a dollar of higher revenue somewhere on the income statement. The capex story rhymes with the monetization story. That is what investors want.

Meta beat on revenue, beat on EPS, raised capex, and the stock fell 5%. Why? Because guidance for Q2 was flat, and the capex bump implied higher depreciation expense pulling down 2027 earnings. Plus Reality Labs is still a money pit. The Street is now asking the uncomfortable question: if you are going to spend $145B on AI infrastructure this year, when does it show up in revenue acceleration the way Alphabet’s does?

Microsoft and Amazon got the same treatment. Cloud growth was good, not great. Capex was huge. The math implies margin compression unless monetization picks up sharply. So the stocks went down.

This is the chasm. On one side, AI capex is being rewarded because it visibly generates revenue. On the other side, AI capex is being treated as a tax on future earnings.

The Fed Just Made It Harder

While everyone was watching tech, the Fed quietly made the financing equation worse. They held rates at 3.5-3.75% for the third meeting in a row, with four dissenting votes. The last time four FOMC members dissented was October 1992. The dot plot still projects only one cut for the rest of 2026.

Translation: rates stay higher for longer. Discount rates on those distant AI cash flows do not come down. The companies that need cheap money to justify $200B of capex are not getting it. The ones that can already monetize ad inventory and cloud at scale (read: Alphabet) are fine. The ones that cannot (read: Meta’s metaverse + Reality Labs spending) are not.

This is why a hawkish hold is much more bearish for unproductive capex than for productive capex.

What This Means for Your Money

There are four practical takeaways here.

One: the basket trade is dead for 2026. “Buy QQQ and forget about it” worked when all of Big Tech rose together on the same AI tide. That is over. The dispersion between Alphabet and Meta this morning is roughly 11 percentage points in 24 hours. Half of the Mag 7 is going to outperform the index. The other half is going to drag it down. Stock picking matters again, and so does the discipline to actually do it.

Two: cash flow per dollar of capex is now the only metric that matters. Pull up each Mag 7’s 2024 capex, divide by 2025 incremental revenue, and rank them. The companies with the best ratios are going to absorb $700B in spending without breaking. The ones with the worst ratios are going to look like 1999 telecoms three years from now. Yes, that comparison is harsh. It is also the comparison the Street is starting to make.

Three: power, not chips, is the real bottleneck. All four hyperscalers basically said “we are constrained by power and component supply, not demand.” This is a generational signal. The companies selling power generation, transmission, cooling, and rare earth components are about to enter a multi-year supply squeeze. The energy trade and the AI trade are converging into one trade. Anyone still treating utilities as “boring defensive” is a quarter behind.

Four: Apple tonight is the wildcard. Apple reports after the close today. Unlike the other four, Apple does not have a hyperscaler capex problem. They have a services-margin story and a China-iPhone-cycle story. If Apple guides above consensus on services and Q3 iPhone, the Mag 7 narrative becomes “old guard rotates back in, hyperscalers consolidate.” If Apple disappoints on services, the entire Mag 7 risks looking like a tired trade.

The Bigger Picture

Step back. $700 billion is being committed in a single year, by four companies, to build the infrastructure for a technology whose largest pure-play customer (OpenAI) just told its CFO it might not be able to pay its compute bills. Either this is the most overbuilt cycle in the history of capital markets, or it is the most underappreciated buildout since the early internet.

Wealtharian’s view: it is both, and the spread between the winners and the losers is going to make some people generationally rich and others permanently poorer. The people who make money from this point forward will be the ones who stop treating “AI” as one trade and start treating it as fifteen different trades — chips, power, real estate, software, monetization, model providers, agents, infrastructure, cooling, networking, and so on.

The lazy money already made its return on AI. From here, you actually have to think.

Want to track how your own portfolio is responding to all this? The Wealtharian Wealth Tracker lets you watch your net worth, FU money progress, and concentration risk in one place. Free to use. Try it here.