CoStar Group is a $25 billion business. Verisk Analytics is a $40 billion business. ICE Mortgage Technology runs the mortgage-data spine for most of the residential market. All three built durable businesses on the same thesis: become the single source of truth for an industry's transactional data, then sell that data back to everyone in the market who needs to make decisions on it. Industrialized construction has no equivalent. The data that would make one possible is finally being captured.

What gets measured today, and why it doesn't matter

Industrialized construction has surveys. The Modular Building Institute publishes one annually. Trade associations publish quarterly market sizing. Specialized analysts publish back-of-envelope deal flow. They all share the same structural weaknesses:

Quarterly lag — by the time the survey is published, the market has already moved on
Voluntary reporting bias — manufacturers who participate tend to be the ones with the best stories
Inconsistent units of measure — survey respondents define "module," "unit," and "delivered" differently
No price discovery — surveys collect volumes, not transaction-grade pricing
No quality signal — surveys don't distinguish a KeyScore-95 production line from a KeyScore-70 one

The institutional buyers who'd actually pay for industry data — hedge funds long-modular, REITs underwriting modular acquisition, investment banks researching securitization, LPs evaluating fund commitments — don't get what they need from surveys. So they either build their own datasets (slow, expensive, partial) or they price modular at a generic spread that has nothing to do with the underlying signal.

What the data layer underneath has to look like

A real industrialized construction benchmark requires a real underlying dataset. The dataset has to be:

Asset-level, not survey-level. The benchmark needs to know about specific modules, not aggregated manufacturer self-reports.
Continuously updated, not quarterly. Every module that enters the registry moves the index.
Standardized at the unit of measure. Every module gets coded the same way, measured the same way, scored the same way.
Quality-graded. KeyScore is the per-asset quality signal that lets the index distinguish a high-performance production line from a marginal one.
De-identified at the aggregation layer. Manufacturers and projects can't be re-identified from the index output. The data flows up; the attribution stays down.

This is exactly what CoStar built for commercial real estate. CoStar's underlying database tracks specific buildings, specific leases, specific transactions — at unit-level fidelity. The benchmarks they sell (cap rates, lease comps, occupancy indices) are derived from that underlying dataset. The benchmark business is only as good as the registry underneath.

What the Modular Index reports

Once the underlying registry exists, the index computes continuously from the data flowing in. The current build surfaces five core measures:

Average $/sqft across the registry — the closest thing to transaction-grade pricing the industry has had
Average KeyScore — industry-wide production quality, recomputed as modules register
Investment-grade share — percentage of modules at KeyScore ≥72 (BBB-equivalent or better)
Net-zero share — share of modules with net-zero certified production
Resilience mix — distribution across seismic classes, wildfire resistance grades, hurricane ratings

For Pro and Enterprise tiers, the index breaks down by sector, region, and HFA jurisdiction. For Enterprise-tier subscribers, the de-identified raw underlying data is available for the subscriber's own analytical work — the equivalent of CoStar's PowerLayer or Verisk's CAS access.

The five buyer cohorts

The buyers for a real industrialized construction benchmark fall into five clear cohorts:

Data platforms — CoStar, Verisk, ICE Mortgage Technology. Strategic partners or acquirers. The natural home for the benchmark once the dataset is at scale.
Investment banks — research desks tracking modular securitization as a credit strategy. The benchmark provides the spread data their analyst notes need.
Hedge funds + private credit — funds long-modular, REITs underwriting modular acquisition, private credit funds underwriting modular construction as a strategy. The benchmark provides the pricing and quality reference these funds underwrite against.
Institutional LPs — pension funds, sovereign wealth, foundation endowments evaluating commitments to modular fund managers. The benchmark provides the manager-versus-index comparison that LPs use for manager selection.
Government + policy — HFAs tracking housing-production index data, federal program offices monitoring industrialized construction velocity, advocacy groups making the case for production at scale.

Three licensing tiers, all driven by the same underlying dataset

Standard ($12K/yr) — Monthly snapshot, quarterly trend report. For research desks + analyst seats.
Pro ($36K/yr) — Sector breakdowns, REST API access, 5 named seats. For investment / data teams.
Enterprise ($120K/yr) — Per-region + per-jurisdiction breakdowns, custom benchmarks, raw de-identified underlying data, unlimited seats, dedicated support. For data platforms + portfolio quants.

The cold-start problem, and how it gets solved

An index requires data, and the data is only valuable at scale. That's the classic two-sided cold-start problem every CoStar- style business has to solve. Here's how this one gets solved:

The registry is useful before the index is useful. Manufacturers register modules for their own financing and insurance access reasons (per-module quality signal, pool-rating eligibility, resilience-priced premium). The registry data gets captured regardless of whether the index has any subscribers yet.
The first index tier ships at low cost. Even with limited registry volume, the Standard tier ($12K/yr) provides a usable benchmark for research desks and analyst seats. Early subscribers anchor pricing while volume catches up.
Volume compounds. Each additional manufacturer onboarded adds modules to the registry. Each module makes the index more representative. The flywheel turns.

What this is worth, structurally

CoStar built a $25 billion business by becoming the source of truth for commercial real estate transactions. Industrialized construction is structurally smaller, but it's also less crowded — no incumbent owns the data layer. The natural strategic home for a mature Modular Index is one of CoStar / Verisk / ICE, which is also the natural pool of acquirers in the medium term.

For LPs and fund managers reading this — the data layer is the interesting bet. Once the registry exists at meaningful scale, the benchmark business has a defensible CoStar-style margin profile: recurring licensing revenue, low marginal cost per subscriber, and natural lock-in (the only way to get unit-level modular data is to subscribe). Read the Modular Index product page → or see the live index in the demo →.

The benchmark business is only as good as the registry underneath. The registry is what's actually being built — the benchmark is what falls out as the registry grows.

Industrialized construction has no real benchmark — and what one would look like