2.0 – 2.6 Hz

We know what the pyramids look like. But what did they do?

A data-driven investigation into the geometry, chambers, shafts, water, stone and frequencies of Giza - and a hypothesis that treats the pyramids as engineered systems rather than monuments alone.

We aren't starting with the assumption that they're machines. We're asking whether the physics gives us a reason to consider it.

Roger Graham · Independent Researcher · 2026

In March 2025, a research team announced spiral shafts descending 648 metres beneath the Pyramid of Khafre. Their data was never released, and the claim is probably wrong. Mainstream archaeology rejected it within days.

Instead of arguing about the scan, I asked a different question: if it were true, what machine would that be? What physics would it need, what would it do, and - most importantly - what measurements would prove or kill it?

One real measurement anchors everything here: a 2026 peer-reviewed seismic study found the Great Pyramid vibrating preferentially at 2.0–2.6 Hz. The boring explanation - any stone mountain that shape rings near that frequency - is fully adequate. This site maps the interesting explanation, states exactly what it would take to be true - and follows the question wherever it leads. It has led far: through 40,000 years of resonant architecture, into the primary surveys of seven pyramid interiors and 107 ordinary tombs, and to a single pre-registered test now waiting to run: when we remove everything ordinary Egyptian tombs already explain, is there any unusual physics left?

700+ primary-sourced dimensions 15 verified acoustic studies 7 famous numbers failed at their sources 1 locked, falsifiable prediction 1 pre-registered test, pending
01 · SEE IT

The Machine

An interactive 3D model. Fly over the plateau, walk the resonance step by step, enter the Grand Gallery, and descend into the claimed tube network.

Open the model → 02 · UNDERSTAND IT

The Giza Test

The full argument in plain language - from taxi driver to geologist. Terminology, diagrams, assumptions, and every source, layer by layer.

Read the explainer → 03 · CHECK IT

The Paper

The 16-page preprint: nine falsifiable tests, a per-reference source-verification ledger, and every place the hypothesis can lose.

Read the preprint →
04 · WEIGH IT

Explorations

The evidence map in 3D: every claim, model and failed idea pinned to its place on the plateau, filterable by discipline and trust level, with sources.

Open the map → 05 · TRACE IT

The Lineage

40,000 years from resonant caves to the Giza question - the honest timeline, with what's established, what's measured, and the famous numbers that died at their sources.

Walk the timeline →

EVERY SOURCE AND DATASET, ONE TABLE, ALL DOWNLOADABLE → SOURCES & DATA

CONDITIONAL HYPOTHESIS
IF IT'S TRUE

If the underground structures are real and the physics holds, the pyramids may have been engineered to place people inside controlled fields of sound, pressure and vibration.

EXPLORE THE HYPOTHESIS →

THE HONEST LEDGER

MEASURED

Khufu vibrates at 2.0–2.6 Hz. Published 2026, 37 seismic measurement points, peer-reviewed. This is the only hard number on this site - everything else is built on, or against, it.

CLAIMED · DISPUTED

648 m shafts and descending spirals under Khafre. Announced March 2025; the tomography data was never released; experts dispute that radar can see that deep. Treated here as an unproven premise, never as fact.

MISSING

Two parts no one has found. A confirmed pressurized water source under the plateau, and an oscillator to turn steady pressure into a beat. Without both, the machine story stays a hypothesis - and we say so on every page.

Nine tests. Any one of them can kill this idea.
That is the point.

The hypothesis is written to lose informatively. The most likely outcome is that the null hypothesis wins: the pyramids ring at these frequencies simply because any limestone mountain of that shape would. But the tests - array seismology across all three pyramids, borehole checks, chamber transmissibility - produce real data either way. That's the wager this site makes.

HOW THIS WAS MADE

This investigation was directed by one person and built in extended collaboration with AI systems. The rules of evidence - primary sources only, pre-registered tests, kill rules stated before results, failures published - were set by the author, who made every judgment call about what to claim and what to discard. The mathematics and modelling were AI-assisted throughout, and several of the project's errors were caught by the author's sourcing discipline, not by the machine. This is disclosed because the project's only asset is its honesty, and because it changes nothing about how to evaluate the work: every model is published, every dataset is downloadable, and every claim stands or falls on its artifacts - not on anyone's authority.