Incidents: September–October 2025 near Elmira, Michigan. Public record: NUFORC 193270, submitted 15 October and posted 18 October, with later investigation material; overview published 28 November. The sightings are grouped as one continuing case at one private property.
A couple described repeated, silent, changing lights close to their wooded home. Their accounts include a multicoloured appearance on 1 September, several lights on 28 September and a bright white appearance during the night of 2–3 October. They supplied photographs and video to NUFORC.
.jpg){kind=link}
A record that needs its chronology preserved
The database header and narrative differ by ten minutes for the first event. The third episode also has an ambiguous date label around midnight. UKPX retains those discrepancies rather than creating a falsely precise timeline. The late-September account includes an observation by the husband that the wife said she missed, so that detail is not presented as filmed.
What the investigation added
Investigator John Budrys interviewed the witnesses and visited on 9 November. His published analysis moved from an initial paranormal-orb interpretation towards a possibly structured object, without identifying it. The overview compresses that discussion; the fuller record is more useful for seeing what remained uncertain.
The investigator did not see the reported object during his visit. The report also acknowledges that an apparent V-shaped departure on video could involve camera movement. That distinction matters: an investigator examining a location later is not an independent witness to the earlier event, and motion in a handheld recording is not automatically object motion.
The flashlight test and AI
The witness described using an AI assistant while considering explanations. The proposed flashlight test for drones did not produce a reaction. Budrys questioned the usefulness of that advice. It should not be treated as a validated experimental exclusion of drones.
A list of possibilities discussed in a conversation, whether with a person or software, is not equivalent to testing each one against calibrated evidence. In this case, strong conclusions would need original recordings, exposure and focus information, stable reference points, and a way to establish the distance of the lights.
Unidentified does not mean identified as extraordinary
The investigator’s speculation about covert technology or non-human origins is an interpretation. The record does not establish either. The reported low heights and short distances are witness estimates, which remain separate from a measured three-dimensional location.
The fuller account also discusses another illuminated area for which ordinary roadside activity was considered possible. That should not be merged into proof of the principal appearances. Keeping episodes, footage and later interpretation distinct makes this a useful record without turning every light around the property into the same phenomenon.
Privacy and map
The map is a broad Elmira-area reference only. The private property is not identified as a visitor destination, and no directions to the witnesses’ home are provided.
Sources and record status
Primary case source · NUFORC investigation overview, 28 November 2025. Sources checked 10 November 2026. Reported observations, source assessments and UKPX’s discussion of evidence limits are distinguished above. No independent identification is claimed beyond the stated source finding.
Map reference: Broad Elmira-area reference from the USGS geographic names record; rounded to protect a private home. Not the property, an approach route or a precise light position. Coordinate source. Incident year: 2025.