Pyramids in the Archive: What the CIA’s Declassified Files Say About Ancient Egypt

Twelve Million Pages, Free to Search

Search the phrase “declassified CIA files on ancient Egypt” and you will find hundreds of pages citing document numbers. Reports on subterranean chambers beneath the Giza plateau. Interagency memoranda on anomalous materials. Technical surveys of the Sphinx. The numbers look authentic, because they are built to look authentic: a prefix, a series code, a sequence number, a hyphenated check digit.

Almost none of them resolve to anything.

This is unusually easy to establish, because the archive in question is not hidden. In January 2017 the Central Intelligence Agency moved its CREST collection online: roughly 930,000 documents, more than twelve million pages, previously available only on four terminals at the National Archives facility in College Park, Maryland. The collection is full-text searchable from a browser, free, with no account required. Anyone claiming a CREST document number is making a claim that can be checked in under two minutes.

What follows is what the archive actually contains on the subject of ancient Egypt and its monuments. It is stranger than the fabrications, and it is real.

The Program Had Six Names, and the Dates Matter

Nearly every genuine reference to pyramids in the CIA’s declassified holdings comes from a single body of material: the STARGATE collection, some twelve thousand documents covering the United States government’s twenty-three-year investigation into remote viewing, the claimed ability to describe a distant location using no ordinary sensory channel.

The program’s naming history is worth learning, because it is the single most reliable test for a fabricated citation.

The Central Intelligence Agency’s own involvement began in 1972 and was discontinued in 1977; the early CIA-funded work at Stanford Research Institute ran under the designation SCANATE. In 1977 the Army Assistant Chief of Staff for Intelligence started GONDOLA WISH to assess whether adversaries might be developing such techniques. Army Intelligence formalised this in mid-1978 as an operational program named GRILL FLAME, housed in buildings 2560 and 2561 at Fort Meade, Maryland, under the INSCOM designation “Detachment G”. The SRI research effort was folded into GRILL FLAME in early 1979. In 1983 it was redesignated the INSCOM CENTER LANE Project. In April 1984 the columnist Jack Anderson reported the program’s existence publicly, and in the same year the National Research Council of the National Academy of Sciences assessed it unfavourably. Army funding ended in late 1985, at which point the Defense Intelligence Agency picked it up under the name SUN STREAK.

The name STARGATE was not applied until 1991, when the surviving efforts were consolidated under a single designation.

What this establishes

Any document dated before 1991 that refers to itself as part of “Project Stargate” is, on its face, impossible. A 1978 memorandum cannot cite a code name that did not exist for another thirteen years. This one test disposes of a large fraction of the fabricated citations in circulation, and it requires no archive access at all.

The Mars Session: A Pyramid in a CIA File

On 22 May 1984, an intelligence officer seated a trained viewer in a room at Fort Meade and handed over a sealed envelope with instructions not to open it. The officer then read a set of geographic coordinates aloud. The session ran approximately one hour. Inside the envelope, revealed only afterwards, was an index card reading: the planet Mars, time of interest approximately one million years B.C.

The nine-page transcript is filed as CIA-RDP96-00788R001900760001-9 in the STARGATE collection and was released with the CREST archive in January 2017.

What the document says

The viewer’s opening impressions are of a yellowish, ochre-coloured landscape and an oblique view of something pyramidal. There are very large, very tall figures. There is a basin of enormous size, and structures with channels cut deep into the ground, described as shelters built against storms. The viewer uses the word obelisk at one point, and describes a monument comparable to those known from ancient Egypt. Asked about the inhabitants, the viewer describes an ancient people confronting the rapid collapse of their environment, and a group that separated from the main population and travelled a long distance seeking somewhere else to live. Pressed on the cause, the viewer produces an image of a globe passing through something resembling a comet’s tail. The transcript ends when the monitor terminates the session; it reaches no conclusion.

What the document does not say

It does not say that anything described was observed on Earth. The Egyptian comparison is the viewer’s own analogy, offered to describe an impression, and it points from Mars towards Egypt rather than the reverse. The file contains no assessment, no analyst’s evaluation, no cover memorandum, and no indication of who requested the session or what was done with the result. The document establishes exactly one fact: the CIA tasked a viewer against ancient Mars, transcribed the result, classified it, and kept it. It establishes nothing whatsoever about the pyramids at Giza, which the session never addresses.

The design of the session is itself significant. Operational remote viewing targets were, at least in principle, checkable — a building at known coordinates, a missing aircraft, a hostage location. A target set one million years in the past on another planet cannot be checked against anything. Whatever the transcript contains, no procedure existed by which it could be confirmed or refuted.

The Ark of the Covenant, Filed as Training

The second genuine document that circulates in this space concerns an object with a direct connection to Egypt in the biblical narrative.

On 5 December 1988, under the SUN STREAK designation, a viewer identified only as 032 was given a coordinate remote viewing task designated Project 0209. The viewer was not told the target. The resulting session is filed as CIA-RDP96-00789R001300180002-7 and was released in August 2000. The viewer described a container holding a second container, made of wood, gold and silver, shaped somewhat like a coffin and ornamented with winged figures, located underground in a damp, dark setting somewhere in the Middle East, with domed structures visible in the distance and Arabic being spoken nearby. The viewer added that the object was protected, and could not be approached by the unworthy.

The stated target was the Ark of the Covenant.

What the document says

It says that on that date, a viewer working blind produced a description broadly consistent with the account in Exodus, including the pairing of one container inside another and the winged figures on the lid. It records the session, the viewer number, the project number, and the stage of the protocol used.

What the document does not say

It does not say the Ark was located. The single most important word in the file is one that almost never survives into the retellings: the session is logged as training. The same target, Project 0209, appears in a separate session run a year earlier on 9 December 1987 with viewer 095, filed as CIA-RDP96-00789R001601010001-2 and titled as a training exercise at Stage IV. Targets in training sessions are chosen for reasons of instruction — to give a viewer something rich in describable structure — not because an agency has an operational interest in recovering a biblical artefact. And the Ark is among the most thoroughly described objects in Western literature. Exodus specifies the dimensions, the acacia wood, the gold overlay, the two golden figures on the cover. A description that matches the text is a description that matches a text almost anyone raised in a Judaeo-Christian culture has encountered, which is precisely why it makes an unsatisfying test of anything.

What the Program’s Own Reviewers Concluded

The most useful document in the entire collection is the one that ended it.

In June 1995 the CIA’s Office of Research and Development commissioned the American Institutes for Research to conduct an external evaluation. The resulting report — <em>An Evaluation of Remote Viewing: Research and Applications</em>, by Michael D. Mumford, Andrew M. Rose and David A. Goslin, dated 29 September 1995 — is held by the National Security Archive at George Washington University and in the CIA’s own reading room.

The review panel was deliberately balanced. Jessica Utts, a professor of statistics at the University of California, Davis, had published work sympathetic to paranormal interpretations. Raymond Hyman, a professor of psychology at the University of Oregon, was among the field’s most prominent sceptics. They reviewed roughly eighty publications and wrote separately.

Utts concluded that, judged by the standards applied to any other area of science, psychic functioning had been established, the statistical effects being too large and too consistent across laboratories to be dismissed. Hyman agreed the better experiments were free of the obvious flaws, and disagreed that anomalous statistics amounted to a demonstrated paranormal cause. That disagreement was never resolved and is printed in the report.

On the operational question, however, the panel was unanimous, and its finding is the sentence that matters most for anyone citing these files as evidence of anything: <q>In no case had the information provided ever been used to guide intelligence operations.</q> The reports were judged inconsistent, inaccurate as to specifics, and dependent on heavy subjective interpretation. The program was terminated.

What this means for the Egypt claims

The archive does contain descriptions of pyramids, obelisks, ancient peoples and buried containers. Those descriptions exist because a government program spent two decades asking people to describe things while blindfolded by protocol, and transcribed whatever came out. The existence of a transcript in a government file is a fact about the government’s filing practices. It is not a fact about Egypt. The program’s own commissioned reviewers, having examined the entire operational record, reported that none of it had ever guided a single operation.

How to Test a Document Number in Ninety Seconds

The fabricated citations succeed because most readers have never opened a real one. They are easy to expose once you know the shape of the genuine article.

Every CREST document carries an identifier beginning with CIA-RDP, followed by a two-digit series number, a container code, and a sequence. The RDP number is not decorative: it identifies a records series, and each series has a subject character. RDP96-00788 and RDP96-00789 are the STARGATE session files. RDP80-00810A, by contrast, is a run of Cold War information reports on foreign industry and commerce — the kind of material that lists rectifiers, marine diesel engines and chrome-nickel wire. A citation attaching an archaeological survey to that series is not merely unverified; it is attaching a title to a series that does not hold that kind of record at all.

Four checks will settle almost any claim of this type:

• Paste the number into the reading room search. Genuine identifiers return a document page with a page count, a creation date, a release date and a scanned PDF. Fabricated ones return nothing.
• Check the code name against its dates. STARGATE before 1991, CENTER LANE before 1983, SUN STREAK before late 1985 — each is an anachronism that invalidates the document on sight.
• Check the named author. Real files are signed by project officers whose names appear across dozens of documents, or are redacted entirely. A single, otherwise untraceable researcher with an impressive title, appearing in exactly one document, is a narrative device.
• Check the series against the subject. If an economics series is said to hold a geophysical survey, the citation has been assembled rather than found.

The reason this matters extends well beyond one topic. The declassified record is a genuine and underused resource, and the fabrications actively damage it: every invented file number gives a reader one more reason to assume that all such citations are noise. The Mars transcript is a peculiar and revealing artefact of Cold War institutional thinking, and it is diminished, not helped, by being surrounded by inventions.

The same discipline applies to the physical evidence. The construction methods behind Egypt’s monuments are addressed at length in the excavation record, which is published, dated and citable from Petrie onwards; the best-known “anomalous technology” relief in Egypt has documented publication history going back to the 1870s. And the pattern of a real fragment of the record being inflated into a hidden history is not confined to Egypt: it is exactly what happened to a label on an eighteenth-century map.

The archive is open. The claims that avoid it are avoiding it for a reason.