← WSJ Heard on the Street hub  ·  Research hub  ·  Research library

WSJ AI & Business — How Shopify's Push into AI Is Working Out

Shopify went all-in on AI — an internal memo making AI use a "baseline expectation", AI questions in performance reviews, and a rule that teams prove a job can't be done by AI before they hire. Revenue has roughly doubled to a projected $15B+ while headcount fell from ~11,600 to 7,600 — the clearest public read yet on whether deep AI adoption produces real operating leverage.
2026-SEP-08 · WSJ AI & Business (weekly newsletter) · by Asa Fitch · written article · Read ↗ · transcript · actionable insights
One-line take: Shopify (SHOP) is the article's test case for whether an established company that plunges into AI — rather than dipping a toe — gets paid for it. The evidence Fitch assembles is operational, not price-based: revenue ~$7B in 2023 to a projected $15B+ this year against headcount cut from ~11,600 (end-2022) to 7,600 (end-2025), plus AI-driven traffic and orders tripling year-over-year in Q2 (President Harley Finkelstein, analyst call). The stock is the other half of the story — down >15% YTD, off 7% intraday Tuesday in a software selloff, and ~57× forward earnings. Next leg: agentic commerce, with Shopify offering its merchant catalog for OpenAI, Anthropic and other model developers to tap for product recommendations. Elsewhere in the issue: PC prices rising on on-device AI and AI-data-center-driven memory costs (HPQ, DELL); and a reversal of January's "AI eats software" panic — Salesforce (CRM) and ServiceNow (NOW) are recovering, because AI writes code well but maintains and updates software poorly. (A WSJ reporter's newsletter — stances below reflect how each name is framed in the reporting and by the executives/analysts it quotes, not a personal call.)

1. Stocks & names mentioned

A written WSJ AI & Business newsletter (no video), so the "At" column links to the article rather than a timestamp. SHOP is the subject; CRM/NOW carry the software-de-rating reversal; HPQ/DELL the PC-pricing item; the AI labs are private. Research legend: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis.

TickerNameResearchViewWhat the article saidAt
SHOPShopifyQT · SA · STK · FAPositiveThe article's verdict is that the all-in AI bet is working operationally: revenue ~$7B in 2023 to a projected $15B+ this year while headcount fell from ~11,600 (end-2022) to 7,600 (end-2025); AI-driven traffic and orders tripled year-over-year in Q2 (President Harley Finkelstein). Caveats stated plainly: stock down >15% YTD (−7% intraday Tuesday in a software selloff) and pricey at ~57× forward earnings. Next: agentic commerce, with Shopify's merchant catalog offered to OpenAI, Anthropic and other model developers.read ↗
CRMSalesforceQT · SA · STK · FAPositiveJanuary's fear that AI coding tools would quickly replace corporate software at a fraction of the cost "hasn't happened, and isn't likely to happen" — AI is great at writing code but poor at continuously maintaining and updating software. With two strong earnings reports since the panic, a recovery in software stocks is under way and could continue as investors accept that AI disruption takes far longer than assumed.read ↗
NOWServiceNowQT · SA · STK · FAPositiveNamed alongside Salesforce as the incumbent the market feared AI would leave "in the dust" in January. The article rejects the thesis on the same maintenance-and-updates argument and points to the recovery already afoot in software stocks.read ↗
HPQHP Inc.QT · SA · STK · FANeutralPC unit sales are sagging, but AI "might provide a respite … at least for manufacturers' pocketbooks": HP and other makers are souping up machines and selling them on on-device AI, and prices are up and projected to climb further. The other driver is a cost, not a win — memory prices have gone through the roof because AI data centers ate the supply.read ↗
DELLDell TechnologiesQT · SA · STK · FANeutralNamed with HP as a PC maker pricing AI-capable machines higher into a falling unit market, while carrying the AI-driven surge in computer-memory costs. Framed as a mix-and-price story, not a demand recovery.read ↗
ADBEAdobeQT · SA · STK · FANeutralWeek-in-AI highlight only: Adobe named a new chief executive as it aims to transform itself for the AI era. No stance argued.read ↗
OpenAIOpenAI (private)NeutralNamed as one of the AI developers Shopify wants tapping its merchant-product catalog for recommendations — the demand side of the coming "agentic commerce" shift, where agents rather than humans search for and recommend products.read ↗
AnthropicAnthropic (private)NeutralNamed alongside OpenAI as a model developer Shopify is making its merchant catalog easily available to, positioning Shopify at the center of agentic commerce.read ↗
Moonshot AIMoonshot AI (private, China)Neutral"The Number": valued at $50 billion in its latest funding round — one of several Chinese firms the article says are challenging the most advanced U.S. model developers.read ↗
Mistral AIMistral AI (private, France)NeutralWeek-in-AI highlight: the French AI developer was valued at more than $24 billion in a new funding round.read ↗

2. Key points

The question the piece is really asking

Outward AI — Sidekick, 2023

Inward AI — the "baseline expectation" memo

The scoreboard — market

The scoreboard — business (the operating-leverage evidence)

Agentic commerce — the next positioning bet

Lütke's own read

PCs — falling units, rising prices

The software-disruption panic unwound

Around the issue

3. In plain English

A jargon-free summary of how each name is framed in the article. (Plain-language companion to the table above; renders on the ticker's consolidated page.)

SHOP — Shopify Positive

Shopify sells the software small and mid-sized retailers use to run an online store — build the website, take the payment, see the numbers. Its CEO, Tobias Lütke, did something few big-company bosses have done: instead of handing staff some AI tools and hoping, he made using AI compulsory. A memo early last year said AI use was a "baseline expectation"; managers now get asked about their team's AI use in performance reviews, new projects are prototyped with AI, and — the sharpest rule — a team has to prove a job can't be done with AI before it is allowed to hire a person to do it.

The reason this article matters beyond Shopify is that it puts numbers on whether that works. Revenue has gone from about $7 billion in 2023 to a projected $15 billion-plus this year, while the number of employees fell from roughly 11,600 to 7,600. Doing twice the business with a third fewer people is the single cleanest public example of AI producing what investors call operating leverage — more output per employee — rather than just AI-flavored press releases.

The catch is the share price, and the article says so plainly: the stock is down more than 15% this year and trades around 57 times next year's expected profits, so a lot of good news is already in the price. The next bet is "agentic commerce" — the idea that shoppers will increasingly ask an AI assistant to find and recommend products rather than browsing themselves. Shopify is handing its merchants' product catalog to OpenAI, Anthropic and others so their assistants can recommend from it, and says AI-driven traffic and orders to its stores tripled versus a year ago in the second quarter. Fitch's framing: Shopify is the barometer — if all-in AI adoption pays anywhere, it should show up here first.

CRM — Salesforce Positive

Back in January investors panicked that AI would gut the big business-software companies: if AI can write code cheaply, why pay Salesforce every year when you could have an AI build you something similar? The stocks were sold off hard on that fear.

This article says that fear was wrong, and gives a specific reason rather than a hand-wave. Writing a first version of some code is the part AI is genuinely good at. Keeping software running, patched, updated and adapted year after year — which is most of what enterprise software actually is — is the part AI is bad at. With a couple of strong earnings reports since the panic, software shares are recovering, and the article thinks that recovery can continue as more investors work out that this disruption, if it comes at all, will take years rather than quarters.

NOW — ServiceNow Positive

ServiceNow is named beside Salesforce as one of the incumbents the market feared AI would leave behind. It gets the same verdict for the same reason: the "AI will rebuild your corporate software for pennies" story ran ahead of reality, because ongoing maintenance and updating — not the first draft of the code — is where the work and the money are. The de-rating that hit these names early in the year is unwinding.

HPQ — HP Inc. Neutral

Fewer PCs are being sold, yet the average PC costs more — and the article says AI is behind both halves. On the good side for HP, it and its rivals are loading machines with hardware that runs AI on the device itself and charging more for them, so revenue per unit rises even as unit numbers fall.

On the bad side, the memory chips inside every computer have become far more expensive because AI data centers are hoovering up the world's memory supply. So part of the higher price tag is not extra profit at all — it is a cost being passed on. That mix of a genuine premium-product tailwind and a real input-cost headwind is why this is a neutral mention rather than a buy case.

DELL — Dell Technologies Neutral

Dell appears in the same item as HP, and the picture is identical: it is selling AI-capable PCs at higher prices into a shrinking unit market, while paying much more for memory chips because AI data centers have taken the supply. The article treats rising PC prices as a fact about the AI build-out, not as a recommendation on Dell shares.


Summary derived from the WSJ AI & Business newsletter (a structured digest of the article is saved in transcript.txt; the piece itself is copyrighted and not reproduced in full) for personal study. Not investment advice. © The Wall Street Journal / Dow Jones for source material.