Everyone Wants to Be the Layer Underneath

Two companies with nothing in common made the same move this year.

Alation spent fourteen years building the data catalog category, then rebranded around AIOS and said, in their own FAQ, that you should build your agents wherever you like. Claude Code, n8n, Microsoft 365, doesn't matter. AIOS just makes sure whatever you build is working from governed data. Glean came at it from the other side, having spent seven years as the assistant people actually open, and then shipped a Model Hub and an MCP Gateway so customers can pick their own models, clouds, frameworks, and front ends without rebuilding underneath.

One climbed up from the catalog. One stepped back from the interface. Both landed in the same place: be the substrate.

I've written about each of these separately. What I didn't say, because I hadn't worked it out yet, is that the pattern is not a strategy. It's a hedge, and it has a bill attached.

The short version

Alation climbed up from the data catalog and told customers to build agents anywhere. Glean stepped back from the assistant and shipped a Model Hub and an MCP Gateway. Opposite ends of the stack, same landing spot: be the substrate, not the assistant. The logic is sound, because the interface layer is where bundling happens and no independent assistant beats "it's already included in your Microsoft license." The problem is that the layer underneath is where Microsoft and Google already live, so you are often retreating into the incumbent's house. And it has a bill: Glean's 45% wDAU/wMAU ratio exists because it owns a surface people open, and plumbing has no such ratio, because nobody opens plumbing. Go fully underneath and you trade a person with a habit for a procurement team with a spreadsheet. The useful question is not which layer to be, it is what has to happen for you to get paid again next year. A human choosing to come back makes you a surface. A system calling you makes you a substrate. What stays scarce is neither the connectors nor the index, it is the permission model and the record of who asked what.

Why everyone is walking downstairs

The logic is easy to reconstruct. The interface layer is where bundling happens. Copilot ships inside a license the customer already signed, Gemini shows up in Workspace, and no independent assistant wins an argument against "it's already included." So if you own something real, you move it somewhere the bundle can't reach. Permissions, lineage, the org-specific context that took eighteen months of connector work to assemble. That stuff is expensive to reproduce and boring to demo, which is exactly what makes it defensible.

Fine. Except the layer underneath is also where the incumbents already live. Microsoft has the Graph. Google has the workspace. The data was in their systems before anyone built a knowledge graph on top of it. "Retreat to infrastructure" only works if you're retreating toward something the incumbent can't reach, and a lot of the time you're retreating directly into their house.

The part nobody prices in

Here's the cost, and it's the reason I don't think this is a strategy.

Glean holds a 45% wDAU/wMAU ratio, more than double the enterprise SaaS norm. Nearly half of monthly users show up on a given day. That number exists because Glean owns a surface people open. It is the single most valuable thing they have, more than the ARR headline, because it's the thing that survives a budget review.

Plumbing does not have a wDAU/wMAU ratio. Nobody opens plumbing. If you fully become the layer underneath, your relationship moves from a person who has a habit to a procurement team with a spreadsheet, and the conversation stops being about whether people would miss you and starts being about your unit price versus the next vendor's. That's a fine business. It is a very different business, and it renews differently.

So both companies are doing both. Glean kept its assistant and opened the gateway. Alation kept the catalog and put governance on top. That isn't decisive positioning, it's optionality, purchased because nobody knows yet which layer commoditizes first.

Optionality is rational at their scale. They can afford to staff two products and let the market pick. Most of us can't, and copying the posture without the balance sheet is how you end up with a substrate nobody plugs into and a surface nobody opens.

Pick your renewal mechanism

The useful question isn't which layer to be. It's what has to be true for you to get paid again next year.

If your renewal depends on a human choosing to come back, you're a surface. Your job is habit: return rate, time to first useful answer, whether the thing shows up where the work already happens. Openness is a feature you offer, not an identity you adopt, and every hour spent making yourself neutral is an hour not spent making yourself sticky.

If your renewal depends on a system calling you, you're a substrate. Your job is reliability, latency, permission fidelity, and the completeness of what you cover. Nobody loves you and that's the correct outcome. Your metric is call volume and breadth of integration, not delight.

If you're honestly both, you have two products with two different renewal mechanics, and pretending otherwise means one of them gets run as a side effect of the other. Usually the substrate wins internally, because it has cleaner metrics and louder customers, and the surface quietly rots.

What actually stays scarce

One more thing worth saying out loud, because the substrate crowd keeps pointing at the wrong asset.

It isn't the connectors. MCP is eating that, and Glean shipping a gateway is the tell. It isn't the index, and it certainly isn't the model. What stays scarce is the permission model plus the record of who asked what and what they did with the answer. Permissions are scarce because they encode org structure that no external system can re-derive. The query record is scarce because it's the only place the organization's real questions get written down, and it compounds.

If you're building toward the layer underneath, that's the asset. Everything else is a feature someone will bundle.

Pick one thing this week: write down the single event that has to happen for you to renew next year. If it's a person opening something, stop calling yourself infrastructure. If it's a system making a call, stop optimizing your onboarding.

Related answer: Should you be the platform underneath or the application on top?