Walmart’s customer card on Anthropic’s IPO website gives about 35% higher average order value for orders via Sparky, Walmart’s shopping assistant, than for other orders. Walmart’s own latest figure is 40%: on its 20 August earnings call, chief executive John Furner said Sparky users “spend 40% more per order than others who don’t”.
Walmart gave the about-35% figure on its 19 February and 21 May calls, so the card matches Walmart’s number on two of its three calls this year, by our count. The gap to the latest figure is 5 percentage points, and the website launched on 10 November, 82 days after the August call, by our arithmetic.
Business Pill · Verification Cost
An evergreen Business Engineer explainer on verification cost: why a claim that is cheap to make can be costly to check. It teaches the general idea only and does not discuss Anthropic or Walmart.
The key insight: As we read the two sets of figures, Anthropic’s IPO website shows the Sparky order-value gap Walmart gave in February and May, about 35%, while Walmart’s own latest call, on 20 August, put it at 40%, a gap of 5 percentage points by our arithmetic.
On the 19 February call, Furner said “the customers who use Sparky have an average order value that’s about 35% higher than non-Sparky customers”. David Guggina, chief executive of Walmart U.S., added: “Roughly half of our app users have used Sparky”.
On the 21 May call, Furner said of Sparky’s users: “Weekly active users are up over 100% just in the last quarter”. Furner repeated the about-35% order-value figure, and Guggina said “units purchased through Sparky have grown more than 4x since the previous quarter”.
On 20 August, Furner gave two new figures: “The number of customers using Sparky is up 70% from last year”, and users “spend 40% more per order than others who don’t”. Food Business News quoted the same words from the call on 24 August.
The order-value gap for Sparky as Walmart stated it on its 19 February, 21 May and 20 August 2026 earnings calls (about 35%, about 35%, 40%) and as shown on Walmart’s card on Anthropic’s IPO website, launched 10 November (about 35%). The February and May calls compare customers, the August call compares spend per order, and the card compares orders. The 5-point gap is our arithmetic.
How the IPO Card Compares
The Walmart card, in the website screenshots Anthropic filed with the SEC as a free writing prospectus, lists three figures: growth of more than 110% in Sparky weekly active users quarter over quarter, about 35% higher average order value via Sparky against non-Sparky orders, and more than 2 times the units sold via Sparky quarter over quarter. Our first read of the website counted the card’s two sales metrics as the only ones of 18 that measure what customers buy.
The card does not say which quarter its figures cover. Its weekly-user figure sits close to the “over 100%” Walmart gave in May, while its units figure, more than 2 times, differs from the more than 4 times Walmart gave for the same kind of measure in May.
The wording also differs. The card compares orders via Sparky with non-Sparky orders; Walmart’s February and May calls compared Sparky customers with non-Sparky customers, and the August call compared what users spend per order with others.
The card quotes Daniel Danker, Walmart’s EVP of AI acceleration, product and design: “Agents like Claude are helping us do both, from better understanding customer intent to accelerating repeatable development work.” The card does not say that Sparky runs on Claude. On the February call Furner named “partners like OpenAI and Alphabet” for new agentic shopping experiences.
Where Sparky Sits in Walmart’s Numbers
On the August call, chief financial officer John David Rainey said Walmart U.S. eCommerce grew 24% in the quarter and that “eCommerce sales now represent over 23% of our mix in Walmart U.S.”. Walmart’s earnings release gave global eCommerce growth of 23%.
None of the three calls we read gives Sparky’s share of Walmart’s orders or sales, so the order-value gap cannot be turned into a sales figure from these sources. Walmart’s agent terms already require an outside shopping agent to get a human’s approval before every checkout unless it shops through Sparky, as we reported in October.
The Structural Read
As we read Walmart’s three calls, its Sparky disclosures moved from engagement, with roughly half of app users having used Sparky in February, to purchase measures: units bought through Sparky in May and spend per order in August.
As we read the card next to the calls, the card’s three figures do not state a period, so they cannot be matched to one Walmart quarter; the weekly-user figure sits close to May’s and the units figure does not.
David Guggina, chief executive, Walmart U.S., earnings call, 19 February 2026
“Sparky is essentially helping us evolve from traditional search to intent-driven commerce.”
Three Implications
FOR IPO BUYERS The Walmart card on Anthropic’s website gives about 35%; Walmart’s own 20 August call gave 40% for spend per order.
FOR WALMART SHAREHOLDERS The three calls we read give Sparky growth and order-value figures but not Sparky’s share of orders or sales.
FOR OUTSIDE SHOPPING AGENTSWalmart’s agent terms require a human’s approval before every checkout unless an outside agent shops through Sparky.
The framework says: “The meaningful metric is work completed rather than conversations initiated.”
As we read it, Walmart’s own Sparky figures moved the same way across 2026, from how many app users had tried it to units bought and spend per order. The framework predates the IPO and does not discuss Walmart or Sparky.
Growth in units bought via Sparky versus the previous quarter: more than 4x, as Walmart said on its 21 May 2026 call, and 2x+, as shown on Walmart’s card on Anthropic’s IPO website, which does not state the quarter.
What Is Not Established
The filed screenshots do not say why the card uses about 35% rather than Walmart’s August figure of 40%, or whether the two measure the same thing. Walmart’s figures compare Sparky users with other customers; the transcripts we read do not say how much of the gap Sparky itself causes.
We did not read the downloadable Walmart case study linked from the card, and we did not contact Anthropic or Walmart for this piece. The deal’s price range and timetable are in our IPO map.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Anthropic’s IPO website shows Walmart’s Sparky orders at about 35% higher average order value, the figure Walmart gave in February and May. On its latest call, on 20 August, Walmart said Sparky users spend 40% more per order, 5 percentage points more by our arithmetic.
A note on sourcing. The Walmart card comes from the IPO website screenshots Anthropic filed with the SEC as a free writing prospectus, read in full. Walmart’s figures come from its own corrected transcripts of the 19 February, 21 May and 20 August 2026 earnings calls, published on its investor site and read for every mention of Sparky, and from its 20 August earnings release; Food Business News carried the same August quote.
The 5-point gap, the day count and the two-of-three count are our arithmetic. We did not contact Anthropic or Walmart for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
