The intelligence graph connecting private capital, companies and real assets. Public records ingested on their own cadence, resolved to one identity by shared identifier, connected by typed relationships with a date and a source on every edge, and read by detectors that say what changed and what it means. Facts, derived metrics, signals, anomalies and inferences are labelled as such.
Reading the graph
United States properties, parcels, loans, leases and transactions from assessor rolls, deed registries, HUD and FHA programme tapes, CMBS filings and bank call reports, resolved into owners, lenders, managers and servicers with a date on every event.
Data as of , computed read only from the real estate core by the hourly Data Factory snapshot from federal programme debt and subsidy; securitised debt and banks; assessor and registry layers; federal tenancy. Sources refresh on their own cadence: factory tick every 15 minutes; each source on its own cadence, daily to annual.
Multifamily is the national backbone (HUD, LIHTC, FHA and CMBS tapes); ownership of record is statewide in Massachusetts and in New York City; recordings are NYC ACRIS 2024 on; other commercial property types are CMBS-tape only.
Observations: planner estimate of reg_observation rows (46M rows, never counted exactly); the head observation id is exact. Entities: count of reg_entity rows not merged into another entity, by type. Relationships: count of reg_relationship rows by rel_type. Dated events: count of reg_event rows; the published subset is rx_pub_event after the rights gate. Loans: count of reg_loan rows by instrument type. Sources: reg_source rows; sources with observations from reg_observation_rollup. Cadence: factory tick every 15 minutes; each source on its own cadence, daily to annual.
Series computed Sep 18, 2026, 3:43 AM UTC; every point is a count or a sum over published rows, never an estimate.
A capital path is a chain of measured links: a plan’s commitment to a fund, the fund’s manager, the company or property underneath; or a lender, a facility, its borrower and the sponsor behind it. The paths are being measured, and each one publishes only when every hop is an identity or a bridge that passed its blind labels.
1,225 loans on DFX's tape (31 FHA-insured multifamily, 1,194 CMBS) carry a scheduled maturity in 2027, $22.9B of current balance ($24.6B at origination). Largest state exposures: NY $4.7B, CA $4.4B, unknown $1.9B, TX $1.1B. Largest lenders or servicers: CITI REAL ESTATE FUNDING $1.6B, NATIXIS REAL ESTATE CAPITAL $1.3B, GSMC $1.3B.
1,514 loans on DFX's tape (41 FHA-insured multifamily, 1,473 CMBS) carry a scheduled maturity in 2028, $27.5B of current balance ($29.6B at origination). Largest state exposures: CA $4.4B, unknown $3.8B, NY $3.8B, FL $2.0B. Largest lenders or servicers: CITI REAL ESTATE FUNDING $3.2B, JPMORGAN CHASE BANK NATIONAL ASSOCIATION $1.6B, UBS $1.2B.
1,287 loans on DFX's tape (47 FHA-insured multifamily, 1,240 CMBS) carry a scheduled maturity in 2029, $25.6B of current balance ($26.8B at origination). Largest state exposures: unknown $5.4B, NY $4.5B, CA $3.0B, TX $1.8B. Largest lenders or servicers: CITI REAL ESTATE FUNDING $4.3B, fdic:33124 $2.5B, BARCLAYS CAPITAL REAL ESTATE $1.4B.
Of $55.6B maturing through 2028 on the programme and CMBS tape, the eight largest lenders or servicers carry CITI REAL ESTATE FUNDING $5.2B (266 loans), JPMORGAN CHASE BANK NATIONAL ASSOCIATION $2.7B (98 loans), GOLDMAN SACHS MORTGAGE $2.4B (104 loans), NATIXIS REAL ESTATE CAPITAL $1.9B (80 loans), UBS $1.7B (140 loans). The largest alone is 9% of the total.
Between 2026-09-18 and 2027-09-18, 1,431 LIHTC compliance periods on the HUD tape reach their end date. Most in CA (174), NY (140), TX (73), NJ (57), VA (48), IN (48).
Between 2026-09-18 and 2027-09-18, 2,796 HUD subsidy contracts on the HUD tape reach their end date. Most in CA (212), NC (191), OH (156), MA (129), NY (117), FL (99).
The record above stays free. What you pay for is the depth behind it.
No account. What a visitor and a search engine both see.
Enough to be useful on its own, and it is how we learn what you came for.
or 10 months paid yearly. The working tier: the depth and the reach behind every lock on this vertical.
For the people who work this market every day rather than look things up in it.
All nine verticals, and the joins between them. The one thing here that no list you can buy anywhere else contains.
The managers and vehicles that own these assets.
$199 / month
The credit side of the same balance sheet.
$249 / month
The offices buying property directly.
$249 / month
Each is its own subscription. Three or more and the whole Factory costs less, at $599 a month for all nine.
Every row on this page carries its source and its date. The methodology says what is measured, what is inferred, and what DFX does not claim. The Data Factory is this and eight other graphs, joined.