Inrupt's move toward \"agentic wallets\", what it signals, and what our own note got wrong

It is not a retreat from pods. The wallet sits on top of one. Our own internal note said otherwise, and checking it is the first thing this page did.

What is actually being built, from their own material

A data wallet that lives on top of a user's Solid pod, holding consented, current customer data that applications can use interoperably.

An agentic layer on top of that, integrating Solid's permission model with a model-context protocol, so an automated assistant can read a person's curated data under permissions granted for specific purposes rather than by handing over a blanket credential.

And a named, multi-year partnership with a global payments company, with a completed proof of concept and a sandbox where merchants, financial institutions and model providers can try it.

The same check also surfaced their separate developer-preview offering, so it was distinguishing their products rather than blurring them, that one is compared elsewhere.

What our own note got wrong, published rather than quietly corrected

It described this as a pivot away from Solid pods. The pod is the substrate. The wallet is built on it and the permission model is Solid's. A company doubling down on the layer we also depend on is close to the opposite of the story our note told.

It also attributed a motive, that the move suggests a search for a revenue model. We are not publishing that. We cannot see inside another company's finances, we did not verify it, and speculation about a competitor's motives is not analysis; it is gossip with a chart.

And it carried a recommendation about pursuing them for partnership. Commercial strategy. Not a customer's business, and not this page's.

What the move actually signals

That the company closest to Solid commercially believes the value is in permissioned data for automated agents, not in selling storage.

That is a substantive bet and it is worth taking seriously. Storage is a commodity; the interesting question is what may read your data, on whose authority, for which purpose. It is the same question access control and consent records are circling from a different direction.

It also suggests the buyer is an enterprise rather than an individual. A payments sandbox for merchants and institutions is not a consumer product, and that matches the wider adoption picture: real institutional deployments, thin consumer uptake.

What it means for us, stated plainly and not in our favour

Two things, and the second is uncomfortable.

First, it is mild good news for the layer we run on. Investment in the Solid permission model by the party with the most commercial exposure to it is better for us than the alternative. We did not cause it and take no credit for it.

If that is where the value turns out to be, we are not in the market for it, and saying so is more useful to a reader than a paragraph about our roadmap.

What we are not going to say

That their developer preview being long-lived proves anything. It is their stated product stage and we have no insight into their plans.