Where the tenants are going, your comps can't see yet.

Your submarket read is a photograph of last quarter. Here's how to read the windshield.

Every submarket read I see starts the same way: pull the comps, average the rents, note the absorption. All of it true, all of it three months old. And the tenants who actually decide where a submarket is going already left a trail your comps won't pick up for another quarter.

- Sasha Deneux

The Take: where the tenants are going, your comps can't see yet.

Here's how a submarket read usually works. You pull the rent comps, the absorption numbers, the supply pipeline, the cap-rate context, and you build a picture of where the market is. It's a good picture. It's also a photograph of where the market was, one to two quarters ago, because that's how long it takes for a signed lease to show up as a comp.

In a flat market, that lag doesn't cost you much. In a market that's actually moving, it's the whole game.

Look at what's moving it right now. AI and tech firms are taking a bigger and bigger share of office leasing, and they're not signing where the models expect them to. They're clustering in smaller mixed-use pockets and converted industrial space, not the Financial District towers your submarket comps are built around. By the time that demand shows up in the rent comps, the good space in that pocket is gone and you paid last quarter's price for it.

So the operators pulling ahead aren't the ones with the cleanest backward comps. Everyone has those now. They're the ones reading the windshield instead of the mirror: where permits are clustering, which corridors are seeing businesses open versus close, how brokers describe a district today versus six months ago. Softer signals, but earlier ones.

The comps tell you what a submarket was worth. The forward signals tell you where it's going. You need both, and most reads only have one. So let's build the one that has both.

The Teardown: the submarket intel report, end to end

Picture the report an acquisitions team actually needs before it bids in a new submarket. Not a forty-tab data dump. A read: here's where this submarket is, here's where it's going, here's what would change my mind. Here's how to build it.

Start with the fundamentals, sourced, not scraped-and-hoped. Pull the rent comps, the absorption trend, the supply pipeline, and the cap-rate context, each with its source and date attached. Rents from a named comp set, not "market's around fifty a foot." Absorption from the actual quarterly, not a vibe. This is the backward-looking half, and it has to be clean before anything else matters.

Then normalize it into one shape. Comps come from five places in five formats. Get them into a single schema so you can actually compare across the submarket instead of squinting at PDFs. This is where most reads die, in the reconciliation, and it's exactly the part AI does well.

Then layer the forward signals, and label them as signals. Permit activity clustering on a corridor. Business licenses opening versus closing. Tenant and broker language shifting, describing the district differently than they did two quarters ago. None of these are comps. All of them lead comps. The trick is presenting them as leading indicators to test, not facts to underwrite.

Then write the read, with the thesis on top. Where is this submarket headed, what's the evidence, and what specifically would flip the call. Two paragraphs a principal can act on, not forty pages nobody opens.

And the honest caveats. The soft signals are directional, not predictive, so never price off them alone. Every number is sourced or flagged to confirm. A submarket read is decision-support, not a forecast: it tells you where the weight of the evidence points, and where you're still guessing.

What comes back is a report your IC can actually use: the fundamentals clean, the forward signals surfaced, and a thesis that says where the market's going before the comps catch up.

Signal

MIT's 2025 State of AI in Business found 95% of organizations are getting zero return on generative AI, despite $30 to $40 billion in enterprise spending. Why it matters: the firms seeing returns aren't the ones using AI, they're the ones who redesigned a workflow around it. A submarket read is a perfect candidate, if you rebuild the process instead of bolting a chatbot onto it.

New platforms are centralizing public rental data into real-time, unit-level comparisons of rents, fees, and concessions, the manual comp pull that used to eat an analyst's afternoon. Why it matters: clean comps are becoming table stakes, available to everyone. When the inputs commoditize, the edge moves to the read you build on top of them.

Commercial Observer laid out how AI can turn soft submarket signals into testable ones: scanning public meetings for recurring infrastructure concerns, tracking whether local businesses are opening or closing, and flagging when brokers describe a district differently than they did six months ago. Why it matters: these lead the comps. The teams that test them, instead of dismissing them as anecdote, see the shift before it's priced in.

AI and tech firms now account for roughly a fifth of all US office leasing, up from about a tenth a few years ago. Why it matters: the single biggest force redrawing office submarkets is a tenant class your historical model has almost no data on. Where they sign next is the read that matters, and it isn't in last year's comps.

From NextAutomation

A submarket engine that pulls the fundamentals clean, layers the forward signals, and writes the read your IC can act on is exactly the kind of loop we build with acquisitions teams. If you'd rather find where AI pays back before you commit to a build, that's what the AI Opportunity Audit is for: two weeks, your opportunities ranked by dollar impact, and two working proofs on your own deals. Bring a submarket you're trying to get a read on.

Book a call

The submarket you can see in the comps is the one everyone else can see too. The edge is reading the one that's forming, and having the discipline to label the soft signals as signals.

Next week: the IC memo. Turning your underwriting into the thesis, business plan, risks, and returns narrative your committee actually reads.

- NextAutomation Team