After Monday’s close, earnings night delivered two sharply different reports: ad-tech company The Trade Desk and electronics manufacturing services company Sanmina both reported, with one surging after hours before giving back most of its gains, while the other held on to its advance.
First, The Trade Desk. Third-quarter revenue was $736 million, down 0.5% year over year, but about $86 million above the market’s roughly $651 million estimate and above management’s own floor of at least $650 million. Adjusted earnings per share were $0.33, versus an expected $0.25. Adjusted core profit—operating profit excluding one-off expenses such as restructuring charges—was $266 million, compared with an estimate of about $161 million. Fourth-quarter guidance also topped expectations: revenue of at least $755 million, versus about $690 million expected, and adjusted core profit of at least $280 million, against an estimate of about $190 million. The company also increased its share repurchase authorization by $731 million, bringing the remaining authorization to $1 billion. Shares had already climbed 15.5% on Monday to close at $14.93, then briefly rose another 26% after hours.
But the excitement lasted less than an hour. As soon as the conference call began, most of the gains evaporated. In a snapshot at 7:19 p.m. Eastern time, the stock was up only about 4% after hours. Chief Financial Officer Nate Olmstead struck a cautionary note: the company did not expect the usual seasonal lift from the third to the fourth quarter—even though the fourth quarter is typically peak advertising season and would normally bring substantially higher sales. The market remains challenging, and some large advertisers are under pressure. About three percentage points of third-quarter revenue growth also came from a one-time benefit from a strategic partner, which will not recur. Political advertising accounted for a single-digit percentage of third-quarter revenue. More concerning was GAAP net income—the actual profit calculated under the strictest accounting rules—which came in at $38 million, or $0.08 per share, versus $116 million, or $0.23 per share, in the same period last year. The company also recorded about $44 million in restructuring costs this quarter and had just laid off roughly 15% of its workforce in September. CEO Jeff Green himself acknowledged that the company remains “below our own growth expectations.”
Sanmina took a different path that same night. Fourth-quarter revenue was $3.86 billion, up 84% year over year and above the $3.48 billion estimate. Non-GAAP earnings per share were $4.43, versus $3.27 expected—a 35% beat. Non-GAAP operating margin was 9.0%. Guidance for fiscal 2027 was also strong: revenue of $17.5 billion–$19.5 billion and non-GAAP earnings per share of $15.50–$17.50. At the midpoint, that implies revenue growth of about 27% and earnings-per-share growth of about 24%. Shares rose 11.1% on Monday to close at $149.64, then added about 20% after hours, largely holding on to the gain.
The difference between the two reports isn’t whether they beat expectations, but the quality of those beats. The Trade Desk is down 60.67% year to date. On October 7, it hit a 52-week low of $11.62, about three-quarters below its 52-week high of $47.96. Expectations had already been driven to the floor, so a beat was almost a given. What the market really wants to know is whether the company can deliver on its $755 million fourth-quarter revenue outlook, and whether growth is coming from its own business rather than one-off benefits and political advertising. Sanmina’s 84% growth includes a boost from its acquisition of ZT Systems—the companies were not yet consolidated in the year-ago period, leaving a low comparison base. But its orders point to real capacity demand in cloud and AI infrastructure, and its fiscal 2027 outlook gives investors next year’s numbers rather than explanations.
The dividing line of this earnings season is taking shape: beating expectations alone is no longer worth much. What matters is whether a company can beat again next quarter, and whether its growth is sustainable. The Trade Desk is telling a recovery story of “the stock fell too far, expectations were too low, and management is lowering the tone.” Sanmina is telling a delivery story: “AI infrastructure spending is actually reaching contract manufacturers’ books.” The real test is still ahead: the U.S. October Consumer Price Index is due tonight at 9:30 p.m., followed after the close by CoreWeave’s earnings. Whether the economics of AI infrastructure add up will only become clear after that test.
There are three risks: The Trade Desk’s after-hours pullback may be only a knee-jerk reaction to the conference-call wording, and tonight’s opening will be the real verdict; Sanmina’s 84% growth reflects an acquisition-related comparison effect, and the integration costs and debt tied to ZT Systems still need to be monitored; and if the Consumer Price Index comes in above expectations, the rationale for last night’s broad rally will be repriced.
#美股财报 #ArtificialIntelligence
First, The Trade Desk. Third-quarter revenue was $736 million, down 0.5% year over year, but about $86 million above the market’s roughly $651 million estimate and above management’s own floor of at least $650 million. Adjusted earnings per share were $0.33, versus an expected $0.25. Adjusted core profit—operating profit excluding one-off expenses such as restructuring charges—was $266 million, compared with an estimate of about $161 million. Fourth-quarter guidance also topped expectations: revenue of at least $755 million, versus about $690 million expected, and adjusted core profit of at least $280 million, against an estimate of about $190 million. The company also increased its share repurchase authorization by $731 million, bringing the remaining authorization to $1 billion. Shares had already climbed 15.5% on Monday to close at $14.93, then briefly rose another 26% after hours.
But the excitement lasted less than an hour. As soon as the conference call began, most of the gains evaporated. In a snapshot at 7:19 p.m. Eastern time, the stock was up only about 4% after hours. Chief Financial Officer Nate Olmstead struck a cautionary note: the company did not expect the usual seasonal lift from the third to the fourth quarter—even though the fourth quarter is typically peak advertising season and would normally bring substantially higher sales. The market remains challenging, and some large advertisers are under pressure. About three percentage points of third-quarter revenue growth also came from a one-time benefit from a strategic partner, which will not recur. Political advertising accounted for a single-digit percentage of third-quarter revenue. More concerning was GAAP net income—the actual profit calculated under the strictest accounting rules—which came in at $38 million, or $0.08 per share, versus $116 million, or $0.23 per share, in the same period last year. The company also recorded about $44 million in restructuring costs this quarter and had just laid off roughly 15% of its workforce in September. CEO Jeff Green himself acknowledged that the company remains “below our own growth expectations.”
Sanmina took a different path that same night. Fourth-quarter revenue was $3.86 billion, up 84% year over year and above the $3.48 billion estimate. Non-GAAP earnings per share were $4.43, versus $3.27 expected—a 35% beat. Non-GAAP operating margin was 9.0%. Guidance for fiscal 2027 was also strong: revenue of $17.5 billion–$19.5 billion and non-GAAP earnings per share of $15.50–$17.50. At the midpoint, that implies revenue growth of about 27% and earnings-per-share growth of about 24%. Shares rose 11.1% on Monday to close at $149.64, then added about 20% after hours, largely holding on to the gain.
The difference between the two reports isn’t whether they beat expectations, but the quality of those beats. The Trade Desk is down 60.67% year to date. On October 7, it hit a 52-week low of $11.62, about three-quarters below its 52-week high of $47.96. Expectations had already been driven to the floor, so a beat was almost a given. What the market really wants to know is whether the company can deliver on its $755 million fourth-quarter revenue outlook, and whether growth is coming from its own business rather than one-off benefits and political advertising. Sanmina’s 84% growth includes a boost from its acquisition of ZT Systems—the companies were not yet consolidated in the year-ago period, leaving a low comparison base. But its orders point to real capacity demand in cloud and AI infrastructure, and its fiscal 2027 outlook gives investors next year’s numbers rather than explanations.
The dividing line of this earnings season is taking shape: beating expectations alone is no longer worth much. What matters is whether a company can beat again next quarter, and whether its growth is sustainable. The Trade Desk is telling a recovery story of “the stock fell too far, expectations were too low, and management is lowering the tone.” Sanmina is telling a delivery story: “AI infrastructure spending is actually reaching contract manufacturers’ books.” The real test is still ahead: the U.S. October Consumer Price Index is due tonight at 9:30 p.m., followed after the close by CoreWeave’s earnings. Whether the economics of AI infrastructure add up will only become clear after that test.
There are three risks: The Trade Desk’s after-hours pullback may be only a knee-jerk reaction to the conference-call wording, and tonight’s opening will be the real verdict; Sanmina’s 84% growth reflects an acquisition-related comparison effect, and the integration costs and debt tied to ZT Systems still need to be monitored; and if the Consumer Price Index comes in above expectations, the rationale for last night’s broad rally will be repriced.
#美股财报 #ArtificialIntelligence