The GEM Effect: New Evidence for Electromagnetic Gravity Modification?

March 1, 2026

Dr. John Brandenburg reports that an electromagnetic device—a roughly 170‑gram coil taken from a miniature electric motor and driven with Tesla-style three‑phase AC—shows an apparent drop in weight of about 0.15 grams (150 milligrams) when energized, then returns to its prior weight when the power is turned off. He emphasizes that the coil was suspended from a load cell by nylon filaments specifically to reduce the chance that electromagnetic fields could couple into the sensor and fake the effect. In both the paper and the talk, he treats reversibility—on with power, off without it—as a key signature that the signal is not simple drift.

GEM in brief: a unification framework aimed at the “long-range” forces

Brandenburg’s starting point is a philosophical and historical one: if gravity can be modified, he argues, it will not be by tinkering with Newton’s law in isolation but by finding a deeper connection between gravity and electromagnetism. He frames GEM as a unified-field model for the two long-range forces—gravity and EM—explicitly separate from the short-range nuclear forces. In the talk he leans on Einstein’s long pursuit of a unified field theory as a signpost, while also insisting he is not trying to deliver a full “theory of everything.”

“People have been trying to do string theory to unify the strong, weak, and electromagnetic forces – and occasionally gravity – all in one theory. I’s been like a quest for the Holy Grail, and it’s produced nothing. So on we go. Let’s be more humble. Rather than trying to make a theory for everything, let’s make a theory for what we see in front of us.” – John Brandenburg

The formal anchor in his slides is Kaluza–Klein theory: a mathematical approach in which gravity and electromagnetism can be written in a common geometry by adding a compact “fifth dimension.” Brandenburg gives the idea a concrete cosmological twist. He describes an early universe that begins as a Planck-scale vacuum and then becomes an expanding plasma in which electrons and protons dominate; the hidden dimension, in his telling, “deploys” from Planck scale to subatomic scale, shaping the way gravity and electromagnetism appear as distinct fields.

Quantum vacuum physics is the other pillar. Brandenburg portrays the vacuum as filled with zero-point fluctuations (ZPF) and uses Stephen Hawking’s black-hole evaporation as an intuition pump: a black hole begins as “pure gravity” and ends by radiating gamma rays—“pure electromagnetism”—which he takes as evidence that the two are not fundamentally separate. The point is not that black holes are laboratory devices, but that quantum theory already allows gravity and EM to be connected in extreme settings, so a unification model might plausibly extend that connection in weaker fields.

To make the link feel tangible, Brandenburg repeatedly returns to the Poynting vector, the standard measure of electromagnetic energy flow and radiation pressure. In a talk-friendly thought experiment, he describes bright objects in a dark box pushing apart, while dark objects in a bright box can appear to “attract” by shadowing each other from ambient radiation pressure. He then states two postulates on a slide: that gravity fields are equivalent to arrays of E×B drifts or Poynting vectors associated with Planck-scale ZPF, and that gravity and electromagnetism appear separate because of the hidden fifth dimension. He also highlights internal numerical “checkpoints”—including a computed value of Newton’s constant G and a repeated appearance of 42.85—that he treats as signs the framework is more than metaphor. Even readers unconvinced by the derivation can treat these as a window into what the theory is trying to explain: not just a force anomaly, but a connection between constants and scales.

The Vacuum Bernoulli Equation: how he connects theory to a measurable effect

The bridge between Brandenburg’s theory story and his experimental claim is something he calls the Vacuum Bernoulli Equation (VBE). In the transcript he introduces it as “analogous to the Bernoulli equation that allows us flight in air,” and in the slides he presents it as derived from a Kaluza–Klein action formulation in a weak-field approximation. The pitch is clear: if ordinary Bernoulli tells you how pressure differences produce lift in a fluid, VBE is meant to tell you how structured electromagnetic fields can produce an effective “lift” against gravity.

In ordinary aerodynamics, Bernoulli’s relation can be taught as a tradeoff between static pressure and dynamic pressure along a streamline. Brandenburg’s VBE borrows that intuition but changes what the “flow” is. Here, the relevant quantity is electromagnetic energy flow—conceptually tied to the Poynting vector—and the competing term is associated with gravitational energy density. He stresses a sign difference compared to the standard Bernoulli form and argues that this leads to a qualitatively different “lift” mechanism that depends on how electromagnetic power is circulated through space rather than on air moving over a wing.

From that equation he draws a specific prediction: a coil energized with alternating current, especially in a Tesla-style rotating three‑phase configuration, should exhibit a small reduction in apparent weight. In the talk he describes this as electromagnetic fields bending or redirecting gravitational “lines of force” around the device. He states the claim in blunt terms—“it is possible to make anti-gravity,” as he puts it—while also acknowledging there is a large gap between a milligram-scale laboratory signal and any practical propulsion system.

This is also where the burden on measurement becomes sharpest. If the claim is “gravity modification,” then a load cell reading is not enough by itself; the experiment has to convincingly exclude other vertical forces that could change the reading. Brandenburg’s materials treat VBE as the mechanism that makes the signal meaningful, but the credibility of that mechanism, for most readers, will rise or fall on replication and on whether the observed effect scales with field configuration the way the theory says it should.

Where the idea comes from: historical anomalies and “cookbook” demonstrations

Brandenburg does not present his work as emerging from nowhere. He cites earlier reports—especially a 1989 paper by Hayasaka and Takeuchi claiming a direction-dependent weight reduction in a spinning gyroscope—and he treats those reports as hints that gravity might couple to rotating electromagnetic field patterns. He also cites a 1990 study that reported no effect, but argues in his paper that the attempted replication differs in a crucial way: the gyroscope was spun using compressed air, not by an electromagnetic drive, which he views as changing the relevant physics.

In his slides he also points to a 1968 report attributed to N. A. Kozyrev, presented as a translated government document, and labels such sources “chance reports.” In the talk he says he tracked that report down and that his group reproduced the effect when they repeated the gyroscope-style experiment using Tesla fields. These claims sit in contested territory—anomalies that have circulated for decades without broad consensus—so Brandenburg’s strategy is to treat them as motivation rather than as definitive proof.

A turning point in his narrative is practical rather than theoretical. He describes a bearing failure that led his team to remove the rotor from a brushless motor and run only the stator coils. In his telling, the “rotorless” configuration produced a larger signal, suggesting that the key ingredient is not mechanical rotation but the electromagnetic field geometry itself. That pivot is important because it shifts the hypothesis toward something simpler and more reproducible: powered coils rather than precision spinning hardware.

“One of the bearings on one of the gyroscopes we were using were old and quit working. So, we took out the rotor and just to see what would happen, we ran the coils with without the rotor. It lost even more weight. We found out that just about any coil, if you activate it with AC, you will get a weakening of local gravity.” – John Brandenburg

To reinforce that simplicity, the slide deck includes a “Gravity Modification Demonstration ‘Cookbook’,” describing balloon-based demonstrations in a still-air enclosure and citing a frequency threshold on the order of hundreds of hertz. Brandenburg emphasizes the fragility of such setups—air currents and heating can overwhelm milligram forces—and he portrays them as illustrative rather than final. In effect, the “cookbook” is a recruitment tool: it says the phenomenon is accessible, but it also sets up why more controlled measurements, such as a load cell experiment, are necessary.

The load-cell experiment: what he built, what he reports seeing, and how he tries to control artifacts

The most concrete evidence in Brandenburg’s materials is a load-cell measurement of a rotorless three‑phase stator coil. In the paper’s abstract, he describes a 170‑gram coil from a miniature race car motor energized with Tesla three‑phase power at frequencies around 400 hertz. The coil is suspended from a load cell by nylon filaments, explicitly “to avoid EM interference,” and the reported effect is a loss of approximately 0.15 grams when energized, returning to normal when the power is terminated.

The slide deck adds implementation specifics: the stator coil is driven by a standard three‑phase controller and supported in a Plexiglas structure. The coil is hung beneath a load cell—identified in the slides as a Loadstar RAPG‑300G‑A with a 300‑gram capacity—and the signal is read out through a small converter module. To reduce mechanical coupling through wiring, the paper describes liquid-metal (Galinstan) cup electrodes used to feed current through thin copper conductors while keeping the assembly mechanically compliant.

Brandenburg reports the results as a clear step-like change in the measured force. One slide is labeled “GEM Effect Data” and states “150 mg weight loss,” and in the transcript he spells it out: “150 milligrams weight loss … that’s 0.15 grams,” emphasizing that the signal-to-noise ratio was strong. The claim is not that the weight drifts slowly over time; it is that the force reading changes with power application and then relaxes back when power is removed, consistent with his interpretation of a field-driven effect.

He also discusses what can go wrong. In the paper he identifies a thermal artifact: heating causes Galinstan to expand, which can alter buoyancy forces on immersed electrodes and distort the measurement; he therefore limits power pulses to under about ten seconds. He warns that the controller can ramp power over several seconds by design, so “power on” is not the same as “steady electromagnetic field applied,” and he recommends monitoring the waveform with an oscilloscope. Finally, he highlights a null test: by partially supporting the coil and toggling power, he reports that the load cell shows only small random noise—on the order of a few milligrams—rather than a systematic 150‑milligram shift, which he argues rules out simple electromagnetic pickup in the sensor. In his view, that pushes the debate away from “the scale is being fooled” and toward the harder question of whether some overlooked physical force remains.

How to read the claim: what would convince skeptics, and what would be learned either way

A load cell does not measure gravity directly; it measures net vertical force. In most everyday situations that force is simply weight, but any additional vertical forces—magnetic attraction to nearby materials, forces transmitted through leads, induced currents in surrounding conductors, buoyancy changes from heating, vibration, or subtle shifts in how stress flows through the mounting hardware—can move the reading. That is why the phrase “weight loss” needs careful handling: the measurement is real as a sensor output, but its interpretation depends on whether competing force paths have been convincingly excluded.

Brandenburg’s materials show that he is thinking about these pitfalls. The nylon suspension, the dome to suppress air currents, the use of compliant liquid-metal electrodes to reduce wire forces, the short pulse duration to limit heating, and the supported-coil null test are all aimed at specific failure modes. At the same time, the remaining vulnerability in any such setup is that time-varying electromagnetic fields can exert forces at a distance in ways that are easy to underestimate, especially if nearby structures contain conductive loops or magnetic materials. A replication effort would likely want comprehensive mapping of nearby materials, alternative mounting geometries, and parallel sensors—ideally with synchronized logging of temperature, vibration, and electrical waveforms—to determine whether the force change tracks the coil’s field in the way VBE predicts or tracks some environmental coupling.

There is also a mismatch of scale that makes physicists cautious. In conventional physics, adding energy to a device increases its total mass-energy, but for ordinary laboratory power levels the equivalent “mass” is unimaginably small—far too small to produce a 0.15‑gram change in weight. Brandenburg explicitly states that his reported effect is “at odds” with Standard Model expectations, which is exactly why independent replication and parameter sweeps matter. If the effect is real and gravitational in nature, it should show consistent scaling with frequency, field configuration, and applied power, and it should survive different measurement technologies.

“Don’t pack your bags for Mars yet – there’s a lot of engineering that must be done to get this thing to work, and to fly human beings anyplace. But this effect has been seen in the lab on a small scale, and it works.” – John Brandenburg

The cleanest way to read Brandenburg’s work is as an empirical proposal with a clear test: build the rotorless three‑phase coil apparatus, document waveforms, temperatures, and geometry, and see whether the 150‑milligram signal appears under controlled conditions and disappears under null configurations. If multiple independent groups reproduce the effect and map its scaling, the theoretical questions become urgent. If they do not, the exercise still teaches something valuable: milligram-level force measurements around dynamic electromagnetic devices are harder than they look, and careful controls are not optional. This summary is based on Brandenburg’s 2026 paper on the GEM effect, his APEC slide deck, and a transcript of his APEC virtual presentation.