The Department of War published its fifth PURSUE tranche on Friday, August 7, 2026 — four weeks after the July 10 release, holding the Friday cadence the program has kept since May. The batch comprises 41 records, and it does something none of the first four did: it leaves the Pentagon. For the first time the archive carries files from the Department of State and from the Executive Office of the President, which means the paper trail now runs through the diplomatic corps and the White House as well as the services and the Bureau. Every record is indexed in the full PURSUE archive, and the running corpus stands at 375 released records across five tranches.
Three agencies and a sphere that was not there
The centrepiece is small, strange, and instructive. On the evening of November 8, 1963, Radio Tupi in Rio de Janeiro broadcast a report that a large metal sphere had fallen in the centre of Conde, in the state of Bahia, gouging a hole four metres deep. The sphere, the broadcast said, had window-shaped openings covered with thick glass, and inside was a human body in heavy clothing bearing inscriptions that could not be deciphered. The CIA’s Foreign Broadcast Information Service picked the broadcast up the following day, and the wire ended up in the files of Edward C. Welsh, executive secretary of the National Aeronautics and Space Council — a cabinet-level advisory body inside the Executive Office of the President that coordinated U.S. space policy from 1958 to 1973. Handwritten annotations on the file, dated November 13, record a conversation about it between offices.
What follows is the part worth reading closely. Six days after the broadcast, the U.S. Embassy in Rio cabled the Secretary of State to confirm that Washington had assessed the newspaper version of the story as inauthentic, quoting the press language back — a “large metal manned balloon,” a “dead crewman aboard” — to be certain both ends of the conversation were discussing the same claim. The cable exists because the previous night’s telephone connection had been poor. Then, on November 20, the Embassy reported the findings of the scientific attaché at the American consulate in Salvador: the incident was unsubstantiated and the reports were a fabrication originating in Rio. There was no contemporaneous local reporting, and what local coverage existed contradicted itself. One of the Brazilian articles the attaché cites carries the headline “The Airship that wasn’t there.” Brazil’s Ministry of Aeronautics, the cable notes, categorically denied that any strange object had been in Conde at all.
Three arms of the United States government spent twelve days establishing that a story was false. That is the honest shape of a great deal of the UAP record, and it is why a government file is not the same thing as a government finding. The archive preserves what officials chased, not what happened.
The films, and the limits of 1953
The strongest historical document in the tranche is the U.S. Naval Photographic Interpretation Center’s frame-by-frame analysis of two motion picture films — one shot over Montana in 1950, the other over Utah in 1952. These are the Great Falls and Tremonton films, the two pieces of footage that dominated official American thinking about UAP in the early 1950s, and the Center went at them with densitometry and photogrammetry, comparing luminosity against apparent angular size.
Its conclusion, in the document’s own words, is that the analysts “presented certain natural phenomena theories to explain the cause of the objects under certain conditions” but “were not able to do so under the conditions during which the photography was taken, nor were they able to recognize these objects as man made.” That is a genuinely negative result rather than a positive one, and the difference matters: the Center could not identify the objects, which is not the same as establishing that they were unidentifiable. The file is candid about why. The analysts flag the assumptions that constrained their confidence, note that they were working from a duplicate of a copy rather than original film, and close by recommending a better-resourced investigation — citing a lack of funding and personnel — that they never got. Seventy-three years later the recommendation is still the most quotable thing in the file.
Ghost rockets, and a war scare that dissolved
The intelligence review of the Scandinavian “ghost rockets” is a Department of War General Staff assessment written in January 1947, looking back at the wave of unidentified missiles reported over Sweden from May 1946. The numbers are the striking part: by the end of July, nearly a thousand reports had come in. The two most common descriptions were “a ball of fire with a tail” and a “shiny cigar shaped object,” and reported altitudes ran from treetop height to 160,000 feet, with the highest figures coming almost exclusively from Finland — which is itself a clue about how these estimates were produced.
The review is unusually frank about the investigation it is reviewing. It records that the Swedish Defense Staff handled the matter oddly, with key personnel on summer leave and not recalled to deal with it, and it documents a divergence between what the Swedish government said publicly and what it was willing to say privately. The suspicion at the time was Soviet — captured German rocket expertise, test-fired across a neutral neighbour — and the review finds the contemporary evidence insufficient to substantiate that or anything else. The wave stopped, no wreckage was ever produced, and the episode survives mainly as the first modern demonstration that a nation can generate a thousand sightings of something that leaves nothing behind.
The Air Force weighs the interplanetary option
Alongside it sits a collection of Air Materiel Command memoranda and correspondence on Project Sign, the 1948 programme whose progress report anchored Release 04. The correspondence characterises some reports as consistent with real objects rather than fabrications or perceptual error, and grounds the national-security argument in performance: speeds and manoeuvres beyond the envelope of known U.S. aircraft, which would matter a great deal if the reports were validated.
It also contains the sentence that every subsequent official document has been circling for seventy-eight years. “The possibility that the reported objects are vehicles from another planet has not been ignored,” the file states. “However, tangible evidence to support conclusions about such a possibility are completely lacking.” Both halves are load-bearing. The hypothesis was considered by serious people at the time, and it was set aside for want of evidence, not for want of imagination.
The CIA checks whether it is chasing itself
Two CIA files concern an incident near Puerto Rico in November 1964. The first is an exchange of memoranda between the Assistant Director for Special Activities — who ran the Agency’s overhead reconnaissance, including the U-2 and the A-12 OXCART — and the Assistant Director for Scientific Intelligence, responsible for foreign technical threats. The correspondence records that the Office of Naval Intelligence was, in the writer’s phrase, “finally convinced, through our efforts, that UFO in fact not YF-12A or similar category Lockheed Aircraft,” and requests that the original radar film be evaluated scientifically for what more it might yield about a high-speed, high-altitude object.
The accompanying briefing notes, prepared for Walter N. Elder, executive assistant to the Director of Central Intelligence, add the detail that a radar aboard the guided missile destroyer USS Gyatt held a contact in the area. Read together with Release 04’s file on the Avro VTOL programme, a pattern emerges across the tranches: a meaningful share of the mid-century official effort went into determining whether the unidentified object was one of America’s own secrets. Sometimes it was. Here, the Agency concluded it was not, and asked for better analysis of the radar — a request whose answer the release does not contain.
Cold orbs, and a camera pointed at a screen
The modern half of the tranche is thinner but more specific. On September 8, 2021, the crew of a U.S. Special Operations AC-130J gunship on a night live-fire exercise over the Gulf of Oman reported two objects appearing after they dropped a water-activated flare: stationary, roughly four feet across, hovering between zero and twenty feet above the water. Over the sortie the crew logged around twenty-five instances of similar phenomena on infrared, described in the intelligence information report as “cold orbs,” with visual speed estimates ranging from 250 to 1,300 miles per hour, formations, aggressive manoeuvring, and apparent reactions to the aircraft firing its main cannon.
Six video clips accompany the report, and the caveat attached to them is the most important sentence in this part of the release. The aircraft’s digital video recorder was not working. What survives is a phone held up to a sensor display — a secondary capture, with the blurring and flickering that filming a screen produces, and none of the fidelity of native sensor data. The report itself is an Intelligence Information Report, which by its own definition disseminates unevaluated information rather than a finished assessment, and it states plainly that the crew’s descriptive and estimative language reflects their subjective interpretation at the time. A speed estimate of 250 to 1,300 miles per hour, made by eye at night against a featureless sea, is a range that tells you about the difficulty of the estimate rather than about the object. Five further clips from a 2019 Pacific Ocean encounter and four from the Middle East arrive with the same structural limitation.
Seventeen FBI files, and eight drawings
The largest single block in the tranche is FBI material: eight FD-302 interview write-ups, eight digital renderings, and one video. The renderings are new to the archive and need to be understood correctly, because they are the most arresting images in the release and the least evidential thing in it. The Bureau commissioned them in 2026 as visual aids, drawing each one from a witness’s spoken description of something seen years earlier. A rendering cannot corroborate the account it was drawn from, and it does something testimony alone does not: it resolves every detail the witness left vague, so a half-remembered shape acquires a clean geometry it may never have had. Each rendering is published here beside the interview it illustrates, which is the only way to read it.
The interviews themselves are more useful. One former military pilot describes standing outside at Bagram in 2002 and noticing the stars being blacked out overhead, as if something huge passed over — then, separately, a run of lights seen on transatlantic flights through 2023 and 2024, which he filmed from a darkened cockpit. Another pair of witnesses to the same 2026 events in the western United States watched multiple red lights through night optical devices; one had twenty-five years of experience with the equipment, and also disclosed that he has astigmatism, and that while the objects looked “dash-shaped” to him they appeared triangular or diamond-shaped to his companion. That is the sort of detail that makes a witness more credible and their observation less conclusive at the same time, and the Bureau recorded it either way.
The one place the FBI material approaches an instrument is a 2026 interview with a government special agent who recorded the event himself on a handheld thermal imager, capturing two slow-moving “black-hot” areas of contrast. A second agent’s account from the same period describes a thermally elevated object over the same western sector on a following night. These are careful observers using equipment they know, filing statements under their own signatures, and what they produce is still two smudges of contrast on an infrared display at unknown range. That is not a criticism of the witnesses. It is the ceiling of what this kind of evidence can reach, and Release 05 sits under it as squarely as the four before it.