The same word, opposite directions
In 2026, as AI computing cooled, Coca-Cola rose +22% and Nike fell -30%. Both are consumer names, both sit outside the AI chain, and yet they moved in opposite directions. As the consumer and service nodes this wiki tracks outside the chain passed twenty, one thing came into focus: under the single word "consumption," several different clocks are running at once.
What The AI money does not spread outside the chain drew was the boundary that the AI cycle is contained within the tech chain. This piece zooms in outside that boundary. The off-chain world is not one block; it is a gradient of clocks running from defensive to discretionary to luxury.
Defensives rise when AI cools
The clearest group is the pure defensives. Consumer staples and food-and-beverage names with virtually no AI exposure all rose in 2026: Coca-Cola +22%, Nestlé +20%, P&G +7%, PepsiCo +3%. They stayed in single digits while the semiconductor chain rose by triple digits, but in the phase where AI rested they became the place money rotates into. The yardstick is the stability of the dividend. Nestlé has raised its payout for 65 years, Coca-Cola and P&G for more than 60. This group's clock is the consumer economy and commodities and FX, not compute demand.
Discretionary and luxury fall on their own cycle
At the opposite pole from the defensives is discretionary consumption. Nike was -30% in 2026, pressed by weak Greater China sales and inventory correction, moving the opposite way while the chain climbed. Four of its last five years are negative. Luxury name LVMH came down from further up: from a +47% boom in 2021, it fell as China consumption and tourism cooled, to -20% in 2025 and -8% in 2026. Toyota rides the clock of a discretionary durable too. Its +63% in 2023 came from hybrid demand and a weak yen rather than AI propagation, and 2026 was -18%. Toyota sources NVIDIA DRIVE and Renesas chips and so touches the chain, but that is a component cost; the stock runs on the auto-demand cycle.
Depending on the cloud is not the AI clock
The most confusing group is the cloud-dependent companies. Netflix, Airbnb and Disney run their services on AWS with no data centers of their own, so they depend deeply on AI infrastructure. Yet their stocks move on subscription, travel and content demand. Netflix's +83% in 2024 was a subscriber recovery from the password-sharing crackdown and the ad tier, and 2026 was -17%. Airbnb tracked travel demand at +10%, and Disney moved on the content and park economy at -12%. Starbucks (+25%) and Uber run on their own operating clocks as well; Uber's +149% in 2023 was a re-rating on the turn to profitability and passing the rate peak. Depending on AI infrastructure and running on the AI cycle are two different things.
Payments and retail are proxies for consumption
The group that mirrors the consumption clock most directly is payments and retail. Visa and Mastercard earn payment fees, so they track consumer spending closely; in 2026 Visa was +4% and Mastercard -5%, a mild consumption curve. Even after putting AI into fraud detection and earmarking billions, their stocks did not break away from the consumer economy. Retail is the same. Costco at +11% showed steady membership spending, while Walmart, +74% in 2024 and then +1% in 2026, traced the retail-media heat cooling off. Korean consumption is more pressed: E-mart at -6% on a domestic slowdown and Coupang at -21%. Whatever the region or format, this group's clock is ultimately the consumer's wallet.
Korea and Asia consumption split on China and overseas volume
Line up the Korea and Asia consumer names and the same divergence reappears. Amorepacific runs on the China consumption clock: down three years through 2021-2023 on the content ban and China's slowdown, then a 2024-2025 rebound on US K-beauty (Laneige, COSRX) and Greater China's return to profit. Orion runs on overseas volume: the roughly 65% of revenue from China, Vietnam and Russia drove three straight double-digit years in 2024-2026. CJ CheilJedang is the opposite, a supposed defensive that fell nearly every year on grain and raw-sugar costs, weak domestic demand and heavy debt. The staples label does not guarantee a rise.
There is one exception, SK Telecom. A carrier, it stacked an AI-infrastructure story: GPUaaS, AI data centers, the SK Group AI Factory with 50,000 NVIDIA GPUs, and an SK hynix HBM alliance. Its stock rose +37.5% in 2025 even as a massive SIM-hack cut revenue and operating profit. AI expectations outweighed the incident. The exact inverse of consumer names that depend on the cloud yet do not run on the AI clock, SK Telecom became AI infrastructure itself and caught the AI clock. What sorts the clock becomes clear here: not whether you sit outside the chain, but whether you build AI infrastructure.
Verdict
Lay the off-chain returns out and "consumption" is not a single clock. Defensives (Coca-Cola +22) rise when AI cools; discretionary and luxury (Nike -30, LVMH -8) fall on their own cycle; and even the cloud-dependent (Netflix -17) run on subscription and travel clocks. Payments and retail mirror consumer spending directly. They share only one thing: none of them runs on the AI clock. When the phrase "AI beneficiary" spreads all the way to consumer names, this map shows where it stops. The story of capital markets running on the rate clock continues in The AI money does not spread outside the chain, and the full arrangement of clocks in Many clocks.