S&P 500 Profit Growth Tracking Above 47% as Disney, CVS, Uber and DoorDash Report Wednesday

Numbers That Both Sides Agree On, But Read Completely Differently

Every side in this argument accepts the same figure. FactSet, the data firm that Wall Street relies on to track this stuff, says S&P 500 profits grew 47.4% in the second quarter, as of July 31^[1]. That's up sharply from the 38.0% growth rate estimated just a week before. About 300 of the 500 companies in the index have reported so far, and roughly 85% of them beat analysts' forecasts.

Nobody disputes that number. What they dispute is what it means. And FactSet's own data hands both sides their best argument in the very same report: strip out just two companies, Alphabet and Amazon, and growth drops from 47.4% to 28.8%^[1]. The average earnings surprise falls even harder, from 31.4% to 9.2%.

That one line is the most important fact in this story, and it's the fact that keeps getting left out. Whether you read this quarter as proof the American economy is booming or as proof a handful of AI companies are propping up everyone else's index fund depends almost entirely on whether you know that exclusion exists.

This week gives readers a way to check the story without waiting for FactSet's next update. Walt Disney, CVS Health, Uber and DoorDash all report earnings Wednesday, August 5^[4][5][6]. None of them sell AI chips or run cloud data centers. They sell theme park tickets, prescriptions, rides and food delivery — the parts of the economy that depend on regular people spending money, not on the AI buildout.

What "Beat the Estimate" Actually Means

Before getting into who's right, it helps to know what these numbers are actually measuring, because the terms do a lot of quiet work.

"Blended" growth means FactSet is mixing real, reported results from companies that have already announced with analyst estimates for the roughly 200 that haven't yet. That's why the number moves every week: as more real results replace estimates, the figure shifts. It's a snapshot in motion, not a final tally.

"Beat rate" is even trickier than it sounds. It doesn't measure how well a company actually did. It measures whether a company beat what analysts predicted it would do. And analysts routinely lower their forecasts in the weeks before a company reports, often with help from the company itself, which likes to guide expectations down to numbers it knows it can clear. An 85% beat rate is partly a story about profits. It's also partly a story about forecasts getting set low enough to jump over.

None of that makes the 47.4% growth figure fake. Second-quarter profits at big U.S. companies are genuinely strong, and no serious analyst disputes that^[1]. But "genuinely strong" and "broadly shared across the index" are two different claims, and this is where the real argument starts.

Three Companies, 70% of the Growth

Analysts estimate that three companies, Alphabet, Amazon and Meta, will supply about 70% of the S&P 500's expected earnings growth for all of 2026^[3]. That is a staggering concentration for an index that's supposed to represent 500 different businesses across every sector of the American economy.

Here's why that matters beyond just being a striking stat: the S&P 500 is weighted by company size, so the biggest firms move the index average the most. When a handful of trillion-dollar companies grow fast, they can drag the whole index's growth rate up even if hundreds of smaller companies are barely moving. Every person with a 401(k) invested in an S&P 500 index fund is riding this concentration, whether they chose to bet on AI or not^[3].

Part of what's driving those three companies' numbers is capital spending, or capex — money spent building data centers, buying chips and constructing server farms. Under accounting rules, that spending isn't counted as a cost against this quarter's profit. It gets recorded as an asset and written off gradually over years instead. That accounting choice makes today's profits look better than they would if the spending were counted all at once.

The catch is that the spending is currently outrunning the revenue it's supposed to generate. Microsoft's capex rose 84% while its Azure cloud business grew 40%^[2]. Alphabet's capex roughly doubled against 82% cloud growth. Announced 2026 spending plans run to $175 billion to $185 billion at Alphabet, roughly $200 billion at Amazon, and $125 billion to $145 billion at Meta^[2]. If all that spending doesn't eventually pay for itself in revenue, the writedowns show up later, not now.

The Case That the Numbers Prove the Policy Worked

Set the concentration argument aside for a moment, because the administration and allied conservative outlets have a case worth hearing on its own terms. Forecasters warned that tariffs would crush corporate profits and reignite inflation. Instead, profits came in far above forecast, and mentions of "recession" on earnings calls collapsed, according to reporting from the comparable 2025 earnings season whose framing conservative outlets are applying again this year^[9][10].

Their argument isn't just about vibes. It points to specific policy mechanics. The 2025 tax law made the earlier corporate tax rate cuts permanent and added a provision letting companies immediately deduct research and development spending, instead of spreading that deduction over several years^[9]. That matters more than it sounds like it should. If a company can write off its R&D costs right away, its tax bill drops this year, which pushes reported after-tax profit up this year. In this reading, that's exactly what a pro-investment tax code is designed to do, and the AI spending boom is proof it's working.

Since the November 2024 election, the S&P 500 has returned more than 30% through July 21, 2026^[8]. That's the kind of number a White House press shop can quote without any asterisks attached, and it has become one of the administration's most repeated talking points heading into the 2026 midterms.

The tell, though, is what doesn't make it into that version of the story. The White House's own past messaging on this selects only the series moving up and to the right — profits, the market — while leaving out payrolls, labor-force participation and consumer spending^[10]. Crediting a single policy for an index-wide number, without ever mentioning that two companies account for nearly half of it, is the same move on both sides of that messaging.

A Job Market That Isn't Telling the Same Story

Meanwhile, the numbers coming out of the labor market paint a very different picture, and this is where labor-focused economists build their case. The Bureau of Labor Statistics reported that nonfarm payrolls rose by just 57,000 in June 2026, after a downward revision cut May's number to 129,000^[13]. Unemployment sits near 4.2%. Real consumer spending, adjusted for inflation, grew at just a 0.4% annualized rate in the first quarter of 2026 — the weakest pace in five years^[14].

Economists on this side call it a "K-shaped" economy: one branch, made up of asset owners and AI-adjacent companies, rising fast, while the other branch, wage earners and labor-heavy industries, stays flat or falls^[14]. Their evidence is the household data, not the earnings data, and their point is a careful one. Record corporate profits sitting next to a stalling job market isn't necessarily a contradiction. It might just be what growth concentrated in automation and AI spending looks like from the ground.

This is exactly why Wednesday's reporters carry more weight than another tech earnings call would. Disney's parks and streaming service, CVS's pharmacies and insurance business, and Uber and DoorDash's ride and delivery volumes all depend directly on people choosing to spend money they don't have to spend^[5][6]. If the profit boom is real and broad, it should show up in how much people are paying for a Disney vacation or a DoorDash order. If it's narrow and concentrated in a few AI giants, this is where that gap would show.

Bullish strategists have a counter that doesn't require dismissing the labor data at all. It's a valuation argument. The S&P 500's forward price-to-earnings ratio, a measure of how many years of future profit investors are paying for in today's share price, actually fell from 22.2 at the end of 2025 to about 20.1 in mid-2026^[7]. That happened because the index price rose about 9% while forward earnings estimates rose about 21% — profits grew faster than the stock price did. Goldman Sachs projects S&P 500 earnings per share of roughly $340 in 2026, up about 24%, and notes the earnings boom isn't purely American: analysts expect 54% earnings growth in emerging markets in 2026, versus 24% in the U.S.^[15]. Asia-based brokerages serving investors deciding whether to keep money in American assets tend to frame the same data more cautiously, listing earnings growth, AI returns and Federal Reserve policy as three distinct risks to watch rather than three reasons for confidence^[11].

Wednesday will add real data to all three arguments at once. Whatever Disney, CVS, Uber and DoorDash report won't settle whether this is a policy success, a labor-market warning, or a valuation story carried by two or three companies. But it will be the first hard evidence in weeks about whether the boom reaches past the businesses selling AI infrastructure and into the businesses selling things to ordinary households.