AARO's new annual report contains numbers that may seem more extraordinary when read in isolation than they really are. The agency received 319 reports of UAP in the analyzed set, resolved 114 of them and transferred 191 to its so-called active archive. But this last category does not mean that 191 cases exhibited unexplained characteristics or technology outside of the known. It means, first of all, that there is a lack of sufficient data to reach a conclusion.
It is also important to get the chronology right. The official AARO page records the FY2025 report as content added on July 20, 2026, while the document itself reports a data cutoff on May 30, 2025. Therefore, what is new about July 2026 is the publication of the report, not the recent occurrence of the 319 reports.
Report is new, but events date back
The main period covered is from 2 June 2024 to 30 May 2025. Of the 319 reports received by AARO, 284 refer to events that occurred within this window. Another 35 correspond to events between 2010 and 2024 that had not been included in previous reports. On May 30, 2025, the organization's total collection totaled 1,870 reports.
This distinction avoids two misreadings. The first would be to treat the 319 as events of 2026. The second would be to assume that they all occurred in FY2025. The report is partly a snapshot of the flow of new reports and partly an update on an accumulated stock.
There is another important difference between received reports and resolved cases. AARO resolved 114 of the 319 reports in this set, but also concluded 256 cases from previous periods. With this, the total number of resolutions carried out during the period covered reached 370.
What happened to the 319 reports
The report separates different analytical situations, and mixing them produces bad interpretations:
- 114 reports were resolved. AARO attributed them to conventional objects or phenomena. The official figure counts 60 balloons, 44 satellites, four aircraft, two UAS, two bird cases, one missile or rocket launch and one manned jetpack.
- 191 were transferred to the active archive, awaiting corroborative data to allow for conclusive analysis.
- Nine reports were selected for additional analysis by partners in the intelligence community and science and technology fields.
- Three cases received an analytical recommendation for closure, still pending review and authorization by the director. Preliminary assessment attributed them, with high confidence, to satellite flares.
These numbers should not be added mechanically as if they were 319 independent destinations. The document also informs two file merger actions: when reports from different sources correspond to the same event, AARO can combine them into a single incident.
Why “active archive” does not mean extraordinary evidence
The report's own glossary defines the active archive as a repository of cases that do not have sufficient data for comprehensive analysis. These records continue to be used in studies of patterns and trends and may be reopened if information of sufficient quality becomes available.
In the section dedicated to the 191 cases, AARO goes further: it states that the existing data does not allow assessing whether the observed performance characteristics are compatible with natural phenomena or whether they exceed the state of the art of known technological systems. This is a statement of uncertainty, not a confirmation of extraordinary performance.
The distinction works both ways. Insufficient data also does not allow us to automatically conclude that the phenomenon was a balloon, an aircraft or another conventional object. The correct result is simply: there is insufficient evidence for a reliable determination.
AARO itself currently uses this logic when explaining its materials classified as “unresolved”: a case remains unresolved when the government cannot definitively determine the nature of the phenomenon, and one of the possible reasons is precisely the lack of data.
NASA reaches a similar methodological limitation. On its UAP questions page, updated in May 2026, the agency states that the limited quantity of high-quality observations prevents firm scientific conclusions about the nature of these events.
Therefore, “unidentified” describes the status of the analysis. It is not, in itself, an alternative identification.
Satellite flares are major analytical change of FY2025
The most technically interesting point of the report may not be in the 191 cold cases, but in those that became explainable thanks to a better tool.
AARO claims to have incorporated a new capability based on three-dimensional modeling and simulation, which allowed it to resolve 238 reports of UAP as satellite flares during the period analyzed.
This number does not mean that 238 of the 319 newly received reports were identified as satellites. This would be incompatible with the other data in the report itself: only 114 of the 319 were resolved, and the figure of closed cases attributes 44 of them to satellites. Therefore, the total of 238 necessarily reflects the application of the new capacity also to cases that were in AARO's previous inventory.
The physical phenomenon is not new. A flare occurs when sunlight is reflected by a satellite's surfaces in a favorable geometry between Sun, satellite and observer. Specular reflections can produce very bright, short-lived spots; Multiple satellites can appear as lights that appear, disappear and appear to move in different directions. AARO's white paper on Starlink further notes that the same principles apply to similar constellations.
The FY2025 report provides an example of how this explanation affects ranking. 44 reports classified in the spatial domain were recorded. None came from sensors installed in space: 42 came from reports from civilian pilots forwarded by the FAA and two from ground-based sensors from U.S. Space Command. AARO says it used multi-source analysis and three-dimensional modeling to attribute all of these cases, with high confidence, to satellite flares.
This detail also helps avoid another confusion: “spatial domain” does not necessarily mean that an object was detected directly by a satellite or orbital sensor. In this set, the vast majority arose from observations made by pilots.
A method change can also change statistics
The effect even appears in the altitude distribution of the reports. According to AARO, the proportion of aerial events reported above 45,000 feet fell from 27% in the FY2024 report to 4% in FY2025. The agency attributes part of this difference to a collection bias in the previous period, which had received a batch of reports from civilian pilots about apparently very loud phenomena. Many of them were eventually identified as satellite flares.
This shows why UAP statistical series need to be interpreted with caution. The number of “anomalous cases” does not just depend on what happens in the sky. It also depends on who reports, what sensors are available, what information accompanies the report, and the tools used later to compare it with known hypotheses.
In this sense, progress in identifying flares is more informative than simply announcing that hundreds of cases have been “solved”. It shows that a previously difficult-to-separate class of observations can shrink as the ability to reconstruct the geometry of the observation improves.
What remains unanswered
FY2025 does not state that all reports have been explained. In addition to the 191 in the active archive, there are nine sent for specialized analysis. The only classified report in the maritime domain also remains under investigation: it was sent by US Navy assets operating off Virginia and described approximately 100 aerial UAPs and two likely unmanned surface systems. The report does not provide a conclusion about the nature of this episode.
There are also two reports that mention electronic or avionic interference in the vicinity of operational aircraft. AARO says it has not determined whether the reported effects were actually caused by the observed UAPs. At the same time, the document says that no reports received during the period indicated flight safety concerns.
In reports linked to national security installations or critical infrastructure, some narratives suggested performance characteristics above the known state of the art. The report makes a decisive caveat: there was no technical data accompanying these descriptions. Consequently, they have not been validated as demonstrations of extraordinary ability.
Among the cases AARO has actually resolved, none have indicated advanced foreign adversary capabilities or disruptive technology. According to the agency, by the cutoff date there was also no evidence that a government or private entity had captured or exploited materials derived from UAP.
These formulations need to be kept in the correct scope. They describe the findings and data available to AARO as of May 30, 2025; do not constitute a statement about any and all phenomena that may exist outside this set.
The report's most solid result is methodological
Read as a whole, FY2025 shows three different things. AARO is receiving hundreds of reports; can solve a portion using conventional explanations and increasingly specific analytical tools; and continues to find a large number of records whose quality does not allow for a conclusion.
The 191 cases of active archive are precisely an indicator of this third problem. They do not automatically increase the evidence in favor of an extraordinary hypothesis. They show the size of the gap between receiving an observation and having enough data to determine what happened. This interpretation is consistent with both AARO's own operational definition and NASA's emphasis on the need for better quality observations.
The report also indicates the direction chosen by AARO: multimodal sensor prototypes, new algorithms, expansion of existing sensor architectures and analytical techniques capable of detecting, tracking, characterizing and potentially attributing observations in closer to real time.
To follow the topic without turning uncertainty into conclusion, the most useful criterion continues to be to separate three questions: what was observed, what data exist and what these data allow to demonstrate. FY2025 shows that the answer to the third question can change as the quality of collection and analysis tools improves.
