The headline numbers
Adobe Analytics forecasts US online holiday sales of 275.1 billion dollars for the period running November 1 through December 31, 2026, a 6.7 percent increase over last year. Cyber Week, from Thanksgiving through Cyber Monday, is expected to account for 47.5 billion dollars of that total, or 17.3 percent of the entire season. Within that window, Adobe forecasts Cyber Monday at 15.1 billion dollars, Black Friday at 12.9 billion, and Thanksgiving Day itself at 6.9 billion, each growing faster than the season's overall rate.
October spend is forecast at 95.8 billion dollars, up 8 percent year over year, with Prime Day on October 6 and 7 contributing 9.9 billion dollars at discounts peaking at 19 percent. Those October figures matter because they confirm something retail planners have suspected for a few seasons now: the holiday shopping window keeps stretching backward into early fall, which means systems built to handle a tight November-to-December peak are increasingly being tested a month earlier than they were designed for.
AI traffic is the story within the story
The number drawing the most attention in Adobe's release sits behind the sales total rather than in it: the traffic shift feeding those sales. Adobe forecasts that visits to US retail sites originating from AI chat tools and AI-enabled browsers will rise 130 percent year over year this holiday season. The growth is uneven across the calendar: Thanksgiving is forecast to see a 159 percent jump, Black Friday 95 percent, and Cyber Monday 88 percent, meaning AI referral traffic is growing fastest on exactly the days retail infrastructure is already under the most strain.
Adobe does not disclose the baseline those percentages are measured against, so none of this converts cleanly into absolute visits or dollars. That is a real limitation for anyone trying to size the engineering problem precisely. What the figures do establish is direction and relative magnitude: AI-originated traffic is growing many times faster than overall online traffic, and it is concentrated on the days when checkout systems, inventory feeds, and personalization engines are least forgiving of added load or unfamiliar request patterns.
Why shoppers say they trust AI more, not less
Adobe paired the traffic forecast with a July 2026 survey of 5,000 consumers, and the results cut against a common assumption that AI-assisted shopping produces more second-guessing and more returns. Seventy-seven percent of AI-assisted shoppers reported feeling more confident in their purchases, and 69 percent said they are less likely to return items bought with AI assistance. If that pattern holds through the actual holiday season, it has direct operational consequences: lower return rates ease pressure on reverse logistics exactly when capacity is tightest.
That confidence effect is worth treating with some caution rather than full confidence of its own. Self-reported purchase satisfaction from a survey fielded months before the actual holiday rush does not guarantee the same behavior holds under real pressure, with real inventory constraints and real shipping deadlines. Retailers building return-rate assumptions into Q4 staffing and capacity plans should treat the 69 percent figure as a directional input worth testing against their own early-season data, not as a number to bake directly into a forecast model.
Grocery and mobile are quietly rewriting channel mix
Adobe's category breakdown shows grocery growing faster than any other major segment at 10.3 percent, ahead of toys at 9.6 percent and cosmetics at 9.5 percent, and well ahead of apparel at 4.7 percent. Electronics remains the largest category in absolute terms at 63.3 billion dollars, but grocery's growth rate signals that online grocery has moved from a pandemic-era convenience into a durable, growing holiday channel, which carries its own implications for fulfillment capacity during the exact weeks warehouses are also handling peak general merchandise volume.
Mobile's share of spend is set to hit a record 63 percent during parts of the season, up from 61.6 percent last year, continuing a multi-year shift that most retailers have already priced into their roadmaps. The number worth flagging for CTOs is the overlap: mobile commerce, grocery growth, and AI-driven traffic are all accelerating on the same calendar days. Systems architected for any one of those trends in isolation are going to be tested on all three simultaneously during Cyber Week.
A footnote correction, zero change to the forecast
On October 16, Adobe republished its holiday forecast release with a correction. The fix addressed a single factual error, the start year attributed to Amazon's Prime Day event, and changed no sales or traffic figures whatsoever. On its own, a corrected footnote is a minor editorial event. In context, it is a useful marker of how quickly these numbers have already been absorbed into retailer planning: a report is being treated as a stable reference document barely two weeks after its original release, well before the season it forecasts has even started.
That speed matters because it compresses the window retailers have to challenge or stress-test Adobe's assumptions before acting on them. A forecast published in late September and corrected in mid-October is functioning, in practice, as a planning input for budgets and staffing decisions that are likely already finalized. Technology leaders relying on third-party seasonal forecasts should build in a deliberate step to validate the assumptions against their own first-party traffic data as early data starts arriving, rather than treating the published numbers as fixed.
What this means for the decision retailers are making right now
Adobe's numbers put a specific, practical decision in front of CTOs: whether current infrastructure can absorb a 95 to 159 percent surge in a referral channel most retailers only started instrumenting in the past year or two, given that the arrival of AI-driven traffic itself is no longer in question. Traffic from AI agents and chat tools often behaves differently from traditional browser traffic, with different session patterns, different bot-detection risk, and different attribution challenges for marketing and analytics teams trying to credit the right channel.
Retailers that have already built dedicated handling for agentic and AI-referred traffic, including bot-detection tuned to distinguish legitimate AI agents from malicious scrapers, are in a materially better position heading into Cyber Week than those treating it as a subset of normal organic traffic. Adobe's forecast is a strong argument for finishing that work in the next several weeks, not after the first week of November makes the gap in capacity planning obvious in real time.
