Drones Mapped a Typhoon's Damage in 10 Days. The Red Cross Used the Data to Write Checks.

The bottleneck in disaster aid is not flying drones. It is closing the loop between aerial imagery and a check being written. Ten days after Super Typhoon Sinlaku hit Saipan, the American Red Cross had enough verified damage data to issue homeowner grants.

Super Typhoon Sinlaku made landfall in the Northern Mariana Islands as a Category 5 storm, with sustained winds above 180 mph that tore roofs off homes on both Saipan and Tinian. The traditional damage assessment that follows a disaster of that scale usually takes one to two months, long enough that the first round of federal aid still has not reached most survivors when shelter costs stack up. According to a joint announcement from FlyGuys and TerraFort, the two companies cut that window to 10 days by chaining drone capture directly to an AI assessment pipeline.

A network of more than 20,000 drone pilots and an AI assessment platform did the work in tandem. FlyGuys dispatched pilots to the Marianas in the days after landfall, where they collected roughly one terabyte of aerial imagery. TerraFort's platform, which the company says was built by former FEMA officials, ingested those images and produced a structure-by-structure damage classification. The two companies report that 4,135 residential structures were assessed in that 10-day window, with 991 destroyed, 2,267 suffering major damage, 739 with minor damage, and 138 affected.

The same dataset reached the Red Cross. Supplemental assistance grants went to homeowners whose properties were classified as destroyed or with major damage, the same threshold the Red Cross uses to target its largest checks. The data also supported local infrastructure planning that fed into a FEMA disaster declaration. A Red Cross check issued inside the first two weeks after a storm is not the federal Individual Assistance program, but it can cover the gap between a tarp on the roof and a federal check that often takes longer to arrive.

The vendors also published a more uncomfortable number in the same release. Both companies note that the actual extent of damage is likely higher than what their pipeline captured, because aerial imagery misses interiors and some structures the flyovers could not safely approach. That is a real limitation of any imagery-only damage assessment, and it is worth keeping in mind when reading the 4,135 figure as a count of damage rather than a count of all damage.

The follow-on change is operational. The two companies have integrated their workflow so TerraFort users can select FlyGuys pilots directly from inside TerraFort, and the system includes an auto-trigger on qualifying weather events. The next time a storm of Sinlaku's size makes landfall in a market TerraFort serves, the data pipeline is supposed to start itself before an emergency manager picks up the phone.

TerraFort says it has active deployments with U.S. county and state agencies and is in talks to expand to additional states, though the company did not name them. Grady Joseph and Joe Stough are both quoted in the release framing the result as a working example of what the two platforms can produce together.

The next test of whether 10 days is a real benchmark, or a vendor-reported one, will come the next time a qualifying storm makes landfall inside a market the integrated workflow covers.