The $11 Trillion Earnings Night: Why Beating Won't Be Enough

Four Reports, One Closing Bell

On Wednesday afternoon, two days from now, the Federal Reserve announces a rate decision at 2:00 p.m. Eastern. Most quarters, that would be the event of the week.

This quarter, it's the opening act. At 4:00 p.m. the market closes, and within minutes Alphabet, Microsoft, Meta, and Amazon all release quarterly earnings - four companies worth more than $11 trillion combined, reporting into the same after-hours session. Apple follows Thursday. A quirk of the calendar has compressed what is normally two weeks of mega-cap earnings into a single evening.

These four are not just the biggest line items in the index. They are the demand side of the entire AI trade - the companies actually writing the checks for the data centers, the chips, and the power that every AI-adjacent stock is priced off. When they speak, NVIDIA's order book, the utility buildout, and half the S&P 500's earnings growth are all implicitly on the call.

We used Barebone to lay three things side by side: the four companies' January earnings reports and what happened next, the consensus numbers they have to clear this week, and what prediction markets are betting. The pattern that falls out is uncomfortable for anyone planning to trade the headline.

Last quarter, all four beat on revenue. Meta jumped 10%. Microsoft had its worst day since March 2020. Same season, same beat, opposite verdicts. Here are the receipts - and the framework for Wednesday night.

The Bar They Have to Clear

Start with the mechanics, because they decide everything else. A stock doesn't move on whether the quarter was good. It moves on the gap between the reported numbers and the predictions Wall Street analysts locked in beforehand - the consensus. Good quarter, higher bar, falling stock. It happens every season.

Here is the bar, as of this week's consensus tallies:

Company	Reports	Consensus revenue	Consensus EPS	Implied revenue growth
Alphabet (GOOGL)	Wed, after close	$107.0B	$2.63	+19%
Microsoft (MSFT)	Wed, after close	$81.4B	$4.06	+16%
Meta (META)	Wed, after close	$55.5B	$6.82	+31%
Amazon (AMZN)	Wed, after close	$177.2B	$1.64	+14%

(Microsoft's is fiscal Q3 2026 - the March quarter. The growth column compares each estimate to the same quarter's actual revenue a year ago, from company filings.)

Look at Meta's bar. Analysts already assume the advertising machine grows 31% year over year. That's not the reward for a good quarter - that's the entry ticket. Anything less is a miss, no matter how big the absolute numbers look. And underneath the headline figures sit the line items the market actually trades: the Street expects Azure, Microsoft's cloud unit, to grow around 38%, and wants to know whether Google Cloud can hold anywhere near the 48% it grew last quarter.

Will they clear the EPS bar? Almost certainly - and that's precisely the problem. Polymarket, the prediction market where traders bet real money on real-world outcomes, currently prices each of the four somewhere between roughly 90% and 98% likely to beat consensus EPS. History agrees: Microsoft has beaten or met EPS estimates in 17 of its last 18 quarters. Amazon and Meta have each done it in 13 of 18.

If "will they beat?" were the question, the week would already be settled. It's the wrong question.

January's Lesson: The Beat Pays Nothing

The last time these four reported - the December quarter, released between January 28 and February 5 - every single one beat on revenue, and three of four beat on EPS. Here's what that bought them:

Company	Reported	Revenue vs estimate	EPS vs estimate	Market reaction
Meta	Jan 28	$59.9B vs $58.4B - beat	$8.88 vs $8.16 - beat	+10% next day
Microsoft	Jan 28	$81.3B, +17% - beat	$5.16, +60% - beat	-10% next day
Alphabet	Feb 4	$113.8B vs $111.5B - beat	$2.82 vs $2.64 - beat	-3% after hours
Amazon	Feb 5	$213.4B vs $211.3B - beat	$1.95 vs $1.96 - missed by a cent	-11% after hours, -5.6% next day

Sit with the Microsoft line for a second. Revenue up 17%, a clean beat. EPS up 60%, a clean beat. The stock fell about 10% the next day - its sharpest daily decline since March 2020 - erasing $357 billion of market value and leaving the company worth $3.22 trillion by that Thursday's close.

The same night, Meta beat by a similar margin and spiked 10%. Two companies, one evening, both comfortably above consensus - and a twenty-point spread between their outcomes.

Amazon got it worst. A record $213 billion revenue quarter, the fastest AWS growth in three years at 24% - and the stock dropped 11% in after-hours trading, settled down 5.6% the next day, and kept bleeding: nine straight red sessions, the worst losing streak since 2006, more than $450 billion in value gone before it finally snapped in mid-February.

So if the beats didn't decide anything, what did?

The Third Number

A stock price is, mechanically, the market's model of all the cash a company will produce in the future, discounted back to today. The reported quarter is the past. What moves the model is anything that changes the future inputs - and that's what guidance is: management's own forecast, the number every analyst plugs into their spreadsheet the moment the call ends.

In January, the guidance that mattered wasn't next quarter's revenue. It was the spending line.

Alphabet told investors it would spend $175 - 185 billion on capital expenditures in 2026 - analysts had been modeling roughly $120 billion, and the company spent $91 billion in 2025. Amazon pointed to about $200 billion, against expectations near $147 billion. Mark Zuckerberg flagged up to $135 billion for Meta. Microsoft's report paired slowing Azure growth with no letup in AI spending. By early February, the four companies' combined 2026 capex plans approached $700 billion - most of it AI infrastructure.

That one line item explains the whole reaction table. Meta's spending came with accelerating ad revenue to pay for it: rewarded. Microsoft's came with a decelerating Azure: punished. Amazon's plan was 56% bigger than last year's spend, announced into a nervous tape: punished for nine straight days. Alphabet nearly doubled its budget and got away with a 3% after-hours dip - because Google Cloud grew 48% and the printed quarter was excellent.

This is how a company beats on revenue, beats on EPS, and still drops 10% before the call is over. The market wasn't grading the quarter. It was re-running the future-cash model with a new, much larger expense line - and deciding, name by name, whether the revenue trajectory justified it.

The script for Wednesday night writes itself: four companies will almost certainly beat, and then four spending-and-growth narratives will fight it out in the dark, after hours, with the whole AI complex as collateral.

The Swing Math

Now put sizes on it. The four companies reporting Wednesday are worth more than $11 trillion combined. Apply January's realized reactions to that base and the numbers stop looking like stock moves and start looking like macroeconomics:

A polite, 3%-average night reshuffles about $330 billion of market value. A 5% night moves over half a trillion. And a January-sized night - double-digit moves on the biggest names - swings more than $1.1 trillion, comfortably more than the entire annual economic output of Switzerland, which the World Bank put at about $885 billion in 2024.

That's the honest version of "every hedge fund is bracing for this week." Not because anyone expects all four to crash - Polymarket says the beats are near-certain - but because the distribution of outcomes on a single evening is measured in Switzerlands, and it lands two hours after a Fed decision.

Where This Framework Breaks

Before you treat Wednesday's after-hours tape as truth, three honest caveats - each one bought with January's data.

First: the one-night verdict often doesn't stick. Meta's triumphant 10% pop? Fully round-tripped - as of this week the stock trades within a few dollars of where it closed before January's report. Amazon's $450 billion punishment? The stock closed Monday at $261.12, up about 24% from where it closed the day after that February report. Of January's three big verdicts, only Microsoft's has held - it's still down roughly 11% this year. The night tells you how the crowd was positioned. It takes weeks to find out what the quarter actually meant.

Second: the mood can flip. January's market punished spending. But the spending came with receipts - Google Cloud growing 48%, AWS accelerating, every hyperscaler still reporting more AI demand than capacity. If this week's cloud numbers accelerate again, the same capex lines that tanked stocks in January could get read as capacity for demand that's already visible. Same number, opposite reaction. The market's read on AI spending is a mood, not a constant - and our sample size for "how do mega caps trade on AI capex" is exactly one cycle.

Third: the first print lies. Amazon was down 11% in the after-hours session and only 5.6% by the next close. After-hours markets are thin; the first hour's move is a handful of traders marking positions, not the market's considered opinion. If you're reading verdicts at 4:10 p.m., you're reading a rough draft.

What to Watch Wednesday Night

Strip out the noise and the watchlist is short:

1. The gap, not the print. Compare every headline to the table above. $107 billion from Alphabet isn't news - it's the assumption.
2. The cloud growth lines. Azure against 38%. Google Cloud against last quarter's 48%. AWS against 24%. These are the revenue streams that make the spending defensible, and they're where January's winners and losers actually diverged.
3. The forward spend number. January proved the capex line is the stock-mover, not the footnote. Any company that raises it had better raise the growth story with it.
4. The second session. Last quarter, the instant reactions overshot in both directions. The Thursday close told you more than the Wednesday tape.

Four of the most profitable businesses in history will almost certainly clear bars they've cleared 13 to 17 times in their last 18 tries. That part is priced. What isn't priced - what can't be, until roughly 4:05 p.m. Wednesday - is whether the third number, the one about the future, validates a trillion-dollar buildout or stretches the market's patience another quarter.

The beat is the formality. The model update is the trade.

---

Data: Barebone | Sources: Microsoft FY2026 Q2 earnings release (Jan 28, 2026), Meta Q4 2025 earnings release (Jan 28, 2026), Alphabet Q4 2025 earnings release (Feb 4, 2026), Amazon Q4 2025 earnings release (Feb 5, 2026), company IR earnings calendars, Polymarket earnings markets, World Bank | Prices as of April 27, 2026 close