Five years after the catastrophic collapse of Champlain Towers South in Surfside, Florida, a final investigative report from the National Institute of Standards and Technology (NIST) confirms that the building showed visible and measurable signs of failure for weeks before it gave way, raising questions about whether emerging structural monitoring technology could have saved 98 lives.
NIST released its findings on June 23, 2026, concluding that two connections between garage columns and the pool deck began failing in early June 2021, nearly three weeks before the building collapsed at 1:22 a.m. on June 24. Investigators documented visible cracking in planter walls, accelerating water infiltration in the parking garage, and a pool-deck section that had fully detached from the slab in the hours before the collapse. The report also found that the building's structural inadequacy was present from construction, with some locations providing less than half of the required code-level strength. Forty years of salt-air corrosion, water intrusion, and deferred maintenance compounded those original deficiencies until the structure had no margin left.
Estructura, a structural intelligence company based in San Juan, Puerto Rico, with offices in Miami and Lima, says the tragedy illustrates precisely why continuous, AI-powered structural monitoring is a life-safety necessity. The company deploys what it describes as the only vertically integrated combination of its kind: GeoSIG precision ground sensors paired with the GeoSMART AI-based software platform, and TerraIntel satellite InSAR imaging that detects millimeter-scale ground deformation and subsidence invisible to on-site inspection. According to Estructura, applied to Champlain Towers South, that combination would have produced a cascade of alerts weeks before the collapse. TerraIntel’s satellite imaging would have tracked differential subsidence of the pool deck slab as reinforcing steel corroded and structural connections weakened. GeoSIG’s on-premise sensor network would have registered anomalous micro-vibrations and load redistribution across garage columns. GeoSMART’s AI trend analysis would have flagged both data streams as anomalous and triggered automated early-warning alerts, giving building managers and residents days or weeks to act.
Estructura emphasizes that the structural design failures documented at Surfside represent only one of four categories of risk that can push any building toward catastrophic failure: design flaws and construction deficiencies; wear, aging, and deferred maintenance; seismic events; and extreme climate events. The company's monitoring platform is designed to detect structural signatures of all four risk categories before small deviations become irreversible failures.
Founded as a division of Dorado Services, a U.S. engineering firm and federal contractor to the Army Corps of Engineers and FEMA since 1999, Estructura says its heritage in emergency response distinguishes it from pure-technology vendors. "We understand what happens when structures fail," said Julio Miranda, Estructura Vice President and co-founder. "We’ve built our company around preventing it." The company integrates GeoSIG precision sensors and the GeoSMART AI platform with TerraIntel satellite InSAR imaging into a single monitoring solution deployable in any structure type worldwide, with a geographic focus on the Americas.
The NIST findings arrive five years after the Surfside collapse prompted Florida to pass landmark legislation requiring condominium associations to maintain adequate reserves for major structural repairs. But Estructura notes that regulation alone is insufficient without the means to continuously verify structural condition. "A reserve fund is only useful if you know what you need to repair, and when," Miranda added. "The Surfside building gave weeks of warning that no one had the technology to read. Every building owner and manager in a coastal city, a seismic zone, or a hurricane corridor should ask themselves the same question: if my building were failing right now, would I know?"
