At a glance: 49 S&P 500 names report today — headlined by GE Vernova and Amphenol before the open, Microsoft, Alphabet, KLA Corporation, ServiceNow, T-Mobile US and Fortinet after the close.

A methodology note: for several of today’s scorecard names, the reported EPS reflects each company’s headline figure on a slightly different basis than the consensus estimate it is measured against, so no year-on-year change is shown for those names.

The tape ahead

Microsoft closes the day after the bell with the quarter that will decide whether the market’s AI-capex arithmetic still works: consensus EPS of $4.72 (up 14.3% YoY) and revenue of $90.74B (up 16.8% YoY) sit against guidance for roughly 45% constant-currency Azure growth, an acceleration from last quarter’s 43%, with management saying it remains capacity-constrained. The debate is less about the top line than the spend behind it — capex is expected to run above $50B in the quarter, a figure optically eased by a change to how data-center leases are depreciated (spread over 25 years rather than 15) — and how broad-based that Azure growth really is, given estimates that a large share of it traces to a handful of frontier-model customers. Margins stayed healthy last quarter (gross margin 67.2%, down 0.4 points; net margin 39.7%, up 1.4 points), and trailing EPS growth has actually accelerated on a one-year view (31.6%) against a high-teens-to-low-20s pace over three, five and ten years. The 5.1% implied move is well above Microsoft’s own 0.5% average swing, against a perfect eight-for-eight beat streak and an 11.8% beat last quarter — a lot of confidence priced into a name that rarely misses. Read-across extends to the broader hyperscaler and AI-infrastructure complex on whether capacity constraints are easing and whether capex guidance for next fiscal year moves again.

Alphabet follows after the close with cloud as the swing factor: consensus EPS of $3.01 (up 4.9% YoY) and revenue of $127.91B (up 25.0% YoY) sit against a Cloud business some estimates had nearly doubling last quarter and a reported backlog above $500B, with at least one bank looking for Cloud revenue to roughly double again year on year this print. The more interesting line is net margin, which jumped 36.7 points to 93.65% last quarter — a figure that has nothing to do with Search or Cloud profitability and everything to do with marking Alphabet’s SpaceX stake to fair value after SpaceX’s IPO; that stake’s implied value has since fallen by roughly a third, so whether this quarter books a partial reversal of that gain rather than a repeat is worth watching alongside the operating numbers. Gross margin eased slightly (61.65%, down 0.8 points), even as full-year capex guidance of $195-205B leaves open whether management raises it again heading into 2027. Trailing EPS growth has been remarkably steady, running from 24.8% over ten years to 34.4% over one. The 6.0% implied move is moderately above Alphabet’s own 1.7% average swing, against a perfect eight-for-eight beat streak and a 214.2% beat last quarter. Read-across extends to the broader hyperscaler group, and specifically to Microsoft reporting the same evening, on whether cloud demand and capex discipline can coexist.

GE Vernova reports before the open into the clearest AI-power-buildout test of the day: consensus EPS of $4.01 (up 144.5% YoY) and revenue of $12.00B (up 20.4% YoY) carry a flagged comparison — last year’s $1.64 print still carried steep Wind-segment losses and predates the equipment-order surge now flowing through revenue, so the headline growth rate says more about how depressed last year’s base was than about a clean trajectory. The underlying business has been accelerating regardless: equipment backlog was running roughly 60% above year-ago levels last quarter on strong Power-segment orders, with management guiding Power revenue growth of 17%-19% and margins near 17%-18%, while Wind is expected close to breakeven after last year’s losses. Margins back up the improvement at the segment level (gross margin 21.3%, up 2.2 points), even as net margin fell sharply (6.0%, down 44.8 points) — a swing worth watching this quarter. The 6.0% implied move is modestly below GE Vernova’s own 6.7% average swing, against a 50% beat rate and a 22.4% miss last quarter — a far less reliable record than most of today’s other names, especially with shares having been weak into the print on financing-related concerns tied to the broader AI power buildout. Read-across extends to the broader power-equipment and grid-electrification complex, and to the credibility of AI-datacenter power demand as a multi-year capex driver.

KLA follows after the close with its own cautious guide as the frame: consensus EPS of $1.18 (up 34.1% YoY) and revenue of $4.04B (up 25.8% YoY) run essentially in line with management’s own guide issued alongside last quarter’s report, when shares fell despite a beat on concerns the guidance range was unusually wide and on Chinese domestic competition in chip equipment. Process-control and inspection tools remain the center of the debate — whether leading-edge logic and memory capex, increasingly AI-driven, keeps absorbing KLA’s capacity as fast as peers add it. Margins held steady (gross margin 61.4%, up 0.2 points; net margin 37.3%, up 2.1 points), and trailing EPS growth has been strong across most windows — 20.5% over one year, 22.3% over five and 23.3% over ten — though softer at 14.9% over three. The 7.6% implied move sits moderately above KLA’s own 3.1% average swing, against a perfect eight-for-eight beat streak and a 5.1% beat last quarter. Read-across extends to the broader semiconductor-capital-equipment group on whether leading-edge spending momentum continues into 2027.

Amphenol reports before the open with AI-datacenter interconnect demand still the dominant storyline: consensus EPS of $0.73 (up 55.3% YoY) and revenue of $9.49B (up 53.3% YoY) sit above management’s own Q3 sales guide of $9.3-9.4B, itself already higher than where analysts had been modeling before the guide — the Street has since pushed estimates past even that raised bar. The read is almost entirely about durability: the AI-exposed IT datacom segment, built on high-speed optical and power interconnects, accounted for a bit over 40% of sales as of the first quarter, so most of the business still rides on non-AI end markets even as the AI-exposed piece draws most of the attention (and the stock’s premium valuation), with the recently closed CommScope acquisition adding further broadband exposure. Margins compressed slightly (gross margin 36.8%, down 1.5 points; net margin 12.2%, down 6.3 points), even as trailing EPS growth has been exceptional — 74.3% over one year, moderating to the high-20s over three and five years (29.7% and 27.9%) and a more modest 18.7% over ten. Amphenol has no weekly options around this print; the nearest expiry is the November monthly, so the 10.5% straddle spans 23 days rather than pricing the print alone, and shouldn’t be read against the 5.7% average one-day move the way a same-week straddle would. The beat rate stands at 100% with a 13.2% beat last quarter. Read-across extends to the broader AI-datacenter interconnect and optical supply chain on whether demand signals are still accelerating.

ServiceNow closes the day after the bell with a guide the Street has already called conservative: consensus EPS of $1.02 (up 6.3% YoY) and revenue of $4.09B (up 20.2% YoY) compare against management’s own Q3 subscription-revenue guide of roughly $3.975-3.98B (about 20.5% growth) and 19.5% cRPO growth, with an acknowledged FX headwind of about $35M. The bigger swing factor is comparability: last quarter’s beat was flattered in part by U.S. federal on-premise revenue that pulled forward from this quarter, so some deceleration in the growth rate is already expected and shouldn’t be read as demand softening on its own. AI remains the long-term pitch — annual contract value tied to AI offerings passed $1B last quarter — against a full-year subscription guide raised to $15.76-15.78B. Margins compressed (gross margin 70.7%, down 4.4 points; net margin 7.5%, down 5.0 points), even as trailing EPS growth has been enormous on a multi-year view (73.6% over three years, 70.1% over five) against a more modest 22.1% over one year. The 10.4% implied move is well above ServiceNow’s own 3.3% average swing, against a perfect eight-for-eight beat streak and a 5.1% beat last quarter. Read-across extends to the broader enterprise-software and AI-agent-platform group on whether federal and public-sector demand normalizes without denting the private-sector growth story.

T-Mobile closes the day after the bell against a stock that has drifted to a multi-year low into the print, with at least one bank citing a softer service-revenue outlook even as full-year guidance was left intact after last quarter. Consensus EPS of $2.91 (up 20.7% YoY) and revenue of $23.14B (up 5.4% YoY) sit against management’s own color for the quarter: postpaid net account additions guided to roughly 250,000, a deliberate step-down reflecting a temporary churn increase tied to rate-plan changes, inside a full-year postpaid account-addition range of 950,000-1.05 million that was itself raised last quarter. Margins improved (gross margin 64.8%, up 3.0 points; net margin 14.2%, up 3.4 points), and trailing EPS growth has been solid over longer windows (29.6% over five years, 27.0% over ten) and extraordinary over three (67.6%), though essentially flat over the past year (0.6%). The 5.5% implied move is modestly above T-Mobile’s own 4.1% average swing, against an 88% beat rate and a 14.9% beat last quarter. Read-across extends to the broader wireless group — Verizon and AT&T among them — on whether the softer postpaid-adds guide marks a one-quarter blip from rate-plan changes or a more durable deceleration in subscriber growth.

Fortinet closes the day after the bell with the firewall-refresh cycle as the central debate: consensus EPS of $0.86 (up 16.2% YoY) and revenue of $2.06B (up 19.6% YoY) sit inside management’s own Q3 guide of $2.01-2.10B revenue and $0.83-0.87 non-GAAP EPS, itself raised alongside a full-year EPS guide increase to $3.41-3.47. Product revenue growth accelerated through the first half of the year, and the question for this print — and for 2027 — is how much of that reflects a large 2026 refresh cohort that analysts expect to be followed by smaller cohorts next year, testing whether growth holds up once the current wave cycles through. Margins stayed strong (gross margin 80.2%, roughly flat; net margin 29.6%, up 0.7 points), and trailing EPS growth has accelerated sharply over the long run, from 7.2% over one year to a 33.0% pace over five years and 74.8% over ten. The 11.1% implied move is well above Fortinet’s own 1.3% average swing, against a perfect eight-for-eight beat streak and a 20.5% beat last quarter. Read-across extends to the broader network-security group on whether refresh-driven product growth decelerates in 2027 as the current cohort cycles through.

Also reporting today

Beyond the marquee eight, 41 more S&P 500 names round out the day. Vertiv’s data-center power and cooling business headlines on strength, with consensus-implied EPS up nearly 47% YoY, while TEL and ODFL both imply growth above 25%. Phillips 66 (+367.9%) and Carvana (+109.5%) both carry flagged, depressed year-ago bases rather than clean trajectories. On the other side, MGM Resorts and Biogen both imply sharp year-on-year declines (down 66.7% and 58.8% respectively).

Scorecard — reported since we last wrote

Forty-two names have scored since last Tuesday’s edition, and the split between who got bought and who got sold on a print is the story. Boeing missed badly on EPS (-70.0%) and narrowly missed on revenue too (-0.8%), then fell 10.7% in the regular session — the cleanest read of the batch. A cluster of after-hours reporters beat and still got sold: Seagate beat on both lines (EPS +3.4%, revenue +2.0%) but fell 5.9% — reported EPS up 194.3% YoY, though that comparison runs against a low year-ago base, not a clean growth read — Teradyne beat on both lines (EPS +11.2%, revenue +2.8%) and fell 5.5% on the same low-base flag (EPS up 169.4% YoY), CoStar Group beat EPS (+11.8%) but missed revenue (-0.8%) and fell 8.2%, and NXP Semiconductors came in in line on both lines — with reported EPS still up 35.0% YoY — yet fell 7.7%, a reminder that an in-line print the tape sells off is a more interesting outcome than a narrow miss would have been. Xylem was the sharpest mover of the day: in line on EPS but a revenue miss (-3.3%) sent shares down 14.2% in the regular session, the steepest reaction on today’s scorecard. On the other side, Pentair’s EPS beat (+10.3%) drove a 12.8% rally, and UPS, Johnson Controls, Royal Caribbean and PPG all beat on EPS and rallied between roughly 5% and 7% — though UPS’s beat came against reported EPS still down 2.9% YoY, a reminder that a beat and a growing business aren’t the same thing. Robinhood missed on EPS (-7.3%) and fell 7.9% after hours, with reported EPS down 16.4% YoY.

Rest of the week

Thursday is the week’s heaviest day by both market cap and event risk: Amazon reports after the close (7.5% implied move) and Eli Lilly before the open (6.1% implied, with consensus implying a year-on-year decline), followed by Mastercard, Intel, Merck, Caterpillar, Linde and Gilead. Friday closes the week with the integrated oils and AbbVie — ExxonMobil, Chevron and AbbVie all show enormous consensus-implied YoY swings, which look like distorted year-ago comparisons rather than a genuine inflection, worth treating with caution against the headline growth rates when they print next week.

Informational only — not investment advice. Figures are consensus/estimates and option-implied values from public sources and may be revised.