Seventy S&P 500 Companies Have Guided Above Consensus for Q3. Forty-Four of Them Are Tech.

by Grace Whitaker · September 7, 2026

Earnings Editor, Earnings Dispatch

US equity markets are shut on Monday for Labor Day, and no S&P 500 company is scheduled to release results. That makes it a reasonable moment to close the books on the second-quarter reporting season and look at what the two forward-looking scorecards — analyst estimate revisions and company-issued guidance — actually say about the quarter now under way.

The season itself is essentially finished. FactSet's Earnings Insight edition dated September 4, 2026 reports that for the second quarter, with 99% of S&P 500 companies having reported actual results, 87% have reported a positive EPS surprise and 77% a positive revenue surprise. That edition does not print five-year or ten-year averages alongside those two beat rates, so no historical benchmark for them is quoted here and the beat rates are not characterized as high or low by historical standards.

The more interesting numbers sit further forward. FactSet's report for the week carried as its topic of the week the observation that analysts have been marking third-quarter estimates up rather than down. The bottom-up Q3 EPS estimate — the aggregation of individual analysts' per-share estimates for the index constituents — "increased by 1.2% (to $89.69 from $88.64) from June 30 to August 31," in the report's words. That figure is dated: it measures the two months to August 31. The same edition separately carries a current reading of the same series — an increase of 1.3%, to $89.78 from $88.64, since June 30 — so the two numbers are the same estimate measured to different cutoffs, not a discrepancy, and the August 31 version is the one used throughout this article.

That direction is the story, not the magnitude. FactSet's own historical benchmarks for the first two months of a quarter are all negative: an average decline of 1.7% over the past five years, 2.1% over ten years, 2.6% over fifteen years and 3.1% over twenty years. Those are the first-two-months benchmarks specifically; the same edition prints a separate, larger set of averages for the change across a full quarter, and the two series are not interchangeable. A 1.2% increase against a twenty-year norm of a 3.1% cut is a swing of more than four percentage points relative to the long-run pattern. FactSet describes the third quarter as the second straight quarter in which the bottom-up estimate rose during a quarter's opening two months.

The revision was not broad. Four of the eleven sectors saw their Q3 bottom-up estimate rise between June 30 and August 31, led by energy at plus 11.8%. Seven sectors saw theirs fall, led by materials at minus 9.1%. An index-level increase built on four sectors is a different fact from an index-level increase built on eleven, and the aggregate number does not distinguish between them.

The second gauge is guidance, and here the reading is more emphatic. By FactSet's September 4 count, 111 companies in the index have issued EPS guidance for the third quarter. Of those, 41 issued negative guidance and 70 issued positive guidance. The report states that "the number of companies issuing positive EPS guidance is well above the 5-year average of 43 and well above the 10-year average of 41." The same edition does not print a comparable historical average for the negative count, so the 41 figure is reported here without a benchmark.

Then comes the concentration. FactSet writes that "44 of the 70 companies (or 63%) in the S&P 500 issuing positive EPS guidance for Q3 2026 are in the Information Technology sector." Nearly two thirds of the positive guides come from one of eleven sectors. Strip information technology out and, by simple subtraction, the positive count falls to 26, below the five-year and ten-year averages the report cites for the whole index. The headline — more companies guiding up than usual — and the underlying distribution are pulling in different directions.

Two definitional points matter before any of this is read as a statement about profit direction. First, FactSet's classification is relative, not absolute. The report defines guidance as "a projection or estimate for EPS provided by a company in advance of the company reporting actual results," classified as negative if the figure or range midpoint "is lower than the mean EPS estimate the day before the guidance was issued" and positive if it is higher. A company can guide to a year-over-year earnings decline and still be counted as a positive guider, provided the number it published cleared the prevailing mean. The count measures a company against the sell side, not against its own prior-year result.

Second, the basis. The definition quoted above carries no accounting label: it refers to an EPS projection by the company and to the mean EPS estimate it is measured against, without specifying whether either is stated on a GAAP or a non-GAAP basis. That matters on this beat, because the distance between the two bases has been unusually wide this season — FactSet's own late-August work shows GAAP EPS actuals at two of the index's largest members carrying multibillion-dollar investment gains booked in other income, detailed below. A tally that compares a company's projection to a compiled mean inherits whatever basis each side used, and the count does not carry that basis forward with it.

The denominator deserves attention too. The 111 guiders are drawn from an index of roughly 500 companies, which means a little over one member in five contributed to the guidance scorecard at all. The remainder either do not issue quarterly EPS guidance, guide on lines other than EPS — revenue, margin, segment volume — or had not guided as of the report's cutoff. The guidance count is a real signal about the companies inside it and silent about everyone else.

Compilers also disagree on what the quarter is worth. FactSet's September 4 edition puts the estimated year-over-year earnings growth rate for Q3 2026 at 28.5% on expected revenue growth of 11.9%. Zacks Investment Research, in a preview published August 21 by Sheraz Mian, projected S&P 500 third-quarter profits rising 22.3% on 10.8% higher revenues. That is a gap of more than six percentage points on earnings between two widely quoted index aggregates, reflecting different constituent sets, different estimate pools and different cutoff dates; the Zacks piece reviewed here does not specify an accounting basis for its figures. Any story describing a company as beating or missing "consensus" this quarter should name which consensus it means.

There is also a base-effect problem sitting underneath every index growth rate quoted this cycle. In a separate August 28 note, FactSet reported that the "Magnificent 7" companies delivered second-quarter earnings growth of 118.5% while the blended rate for the other 493 members was 31.8%. The note attributes much of that to investment gains at two companies: Alphabet's GAAP EPS actual for the second quarter "included a gain of $98 billion in other income primarily due to net unrealized gains on equity securities," and Amazon.com's GAAP EPS actual "included a gain of $53.4 billion in other income primarily due to investments in Anthropic." Excluding those two companies, FactSet says the group's growth rate "falls to 43.2% from 118.5%." Those gains are GAAP items, and they sit in the reported base that later comparisons will be measured against.

Some of the year's arithmetic moves for the same reason. FactSet reports that the CY 2026 bottom-up EPS estimate rose 6.1%, to $361.38 from $340.49, between June 30 and August 31. A calendar-year bottom-up figure blends quarters already reported with estimates for quarters still ahead, so a move in it during a reporting season is not the same thing as an upgrade to future quarters. The Q3-only estimate, which contains no actuals, rose 1.2% over the identical window.

For the quarter itself, FactSet's estimated net profit margin for Q3 2026 is 14.9%, "which is below the previous quarter's net profit margin of 17.0%, but above the year-ago net profit margin of 13.0% and above the 5-year average of 12.4%." On the growth side, the report says "all eleven sectors are expected to report year-over-year earnings growth, led by the Energy, Information Technology, Communication Services, and Materials sectors" — a broader expected distribution than the estimate revisions or the guidance counts would suggest on their own. Looking further out, the September 4 edition projects Q4 2026 earnings growth of 26.1% and CY 2026 growth of 31.5%. The report puts the forward 12-month P/E for the index at 19.5, below its five-year average of 19.8 and above its ten-year average of 19.0.

The week ahead is thin by design. FactSet's September 4 edition states that in the upcoming week four S&P 500 companies are scheduled to report second-quarter results and two are scheduled to report third-quarter results, across four trading sessions. Those are calendar entries, not outcomes: none of those companies has reported, and nothing about their numbers is known until they publish. The rest of the index will not begin reporting the September quarter in volume until the middle of October, at which point the two forward gauges described above stop being predictions and start being tested against filings.