What retailers and publishers just taught telecom about data monetization
Americans used more wireless data in 2025 than ever before, and the growth is still slowing. CTIA’s annual report counts 159.3 trillion MB, up 20%. A year earlier the increase was 32%, and the year before that 35%. Meanwhile US operators put almost $30 billion into network capacity in 2025 alone.
Every operator knows this pattern. Traffic growth slows, the network bill stays. Selling more gigabytes is running out of road, so the next margin has to come from somewhere else. In September, five companies outside telecom showed where.
Kroger: growth from data it already had
Kroger Precision Marketing grew profit 24% in Q2 2026, its strongest retail media profit growth since 2021. CEO Greg Foran credited “stronger collaboration between our merchandising and media teams”, which expanded advertising inventory and improved visibility and conversion for brand partners.
There was no new traffic and no new data source. Two internal teams started working from the same signal, and brands on the other side got a measurable outcome. Operators hold richer first-party signal than any grocer. What most of them lack is the link between commercial, inventory and measurement.
Walmart: from reports to recommendations
At Walmart Data Ventures, Mark Hardy described the next step for the Scintilla insights platform: moving “from summarizing findings or highlighting key points to, now, [making] recommendations.” Suppliers get visibility and direction. They do not get a raw export.
The sequence is first-party data, then insight, then a recommendation of what to do next. Many operators are still at stage two, shipping dashboards. The value sits at stage three.
Instacart: the value leaves, the data stays
Ryan Hamburger, Chief Commercial Officer at Instacart, was direct about the company’s new AI shopping integrations: “We’re not sharing the same level of data with OpenAI as we would with the shopping assistant embedded on our platform.” The reason: “we want to make sure that data remains a competitive advantage for us and for our retail partners.”
Instacart runs its own assistant on the full signal and gives external models a deliberately thinner layer. For any data owner, that is the right instinct.
Google: data priced by contribution
Google is piloting a pay-per-value program in Search Console. Publishers get paid when their content significantly informs an answer in Gemini, AI Overviews or AI Mode. The methodology is not disclosed, and critics call the payouts small. Luke Stillman of Madison and Wall noted that publishers have “relatively little leverage over how AI changes content discovery and distribution.” David Buttle, founder of Spur, pointed to the bigger shift: the program establishes “the notion that usage events convey value.”
Data is moving from flat-fee access to payment for contribution. Whoever controls the attribution controls the price.
Axios: own the client relationship
Axios hit its full-year revenue goal in early September and is launching Axios Direct, content feeds for investment firms, companies with their own AI models and, eventually, personal agents. CRO Jacquelyn Cameron says “about 52% to 54% of our revenue this year will be direct through client”, and that a product built on trust is best started client-direct.
The playbook for operators
Put the five together and the approach to subscriber data becomes clear:
1. Monetize the signal you already have. Kroger connected existing teams and data. Operators need a working loop around the asset they own, not a new asset.
2. Sell decisions. Walmart moved from summaries to recommendations. Buyers pay for what to do next.
3. Keep raw data inside the perimeter. Instacart shares outcomes and a thinner layer. For telcos, with regulators and data sovereignty in play, this is a requirement.
4. Own the attribution. If data is priced by contribution, the party that proves the lift sets the price. Measurement has to be yours, and it has to be transparent.
5. Own the relationship. Axios keeps more than half its revenue client-direct. Selling through someone else’s exchange hands over both margin and context.
This is how afina DMP is built. Behavioral signals the operator already holds become targeted offers, measured inside a closed loop, with an effect of around +5% ARPU for participating operators. No raw dataset leaves the operator’s perimeter.
Traffic growth is slowing and the network bill is not. Operators will close that gap with better decisions on the data they already have.
Sources
Telecoms.com, Andrew Wooden, “US mobile data growth slowed in 2025” (CTIA), 10 Sep 2026
Modern Retail, Mitchell Parton, Kroger ad business, 11 Sep 2026
Modern Retail, Mitchell Parton, Walmart Scintilla, 16 Sep 2026
Modern Retail, Mitchell Parton, Instacart AI assistants, 9 Sep 2026
Digiday, Jessica Davies & Sara Guaglione, Google pay-per-value, 14 Sep 2026
