A research team at the University of Queensland (UQ) in Australia has successfully sent the word "Gravity" through a brick wall by moving an elevator counterweight to change the local gravitational field. Electromagnetic waves used in radio communication can be blocked by materials such as metal plates, but gravitational fields cannot be passively shielded by ordinary matter. The team exploited this property to show that information can be sent and received using a 1950s-era elevator and a commercial gravimeter. However, the data rate was just 1 bit per minute, and the transmission distance was under 1 m.
Why gravity passes through walls, unlike electromagnetic waves
Wi-Fi and mobile phone signals are electromagnetic waves, propagating changes in electric and magnetic fields. Because they interact with electrons in matter, they can be blocked by a metal enclosure (a Faraday cage), and they weaken when passing through thick concrete or brick walls.
Gravitational fields behave differently. Electric charges come in positive and negative forms, which can be arranged so that their fields cancel. For mass, the source of gravity, only positive mass has ever been observed. As a result, no passive shielding method is known that cancels a gravitational field simply by placing metal or rock in between.
Using this property, a UQ team including Dr. Andrew J. Groszek, Dr. Charles W. Woffinden, Michael D. Harvey, Professor Andrew G. White and Professor Matthew J. Davis set out to build a proof of concept for transmitting information with gravity. The results were released on the preprint server arXiv on November 20, 2025, and later published in the peer-reviewed journal Science Advances (vol. 12, eaee0802). UQ announced the findings on November 19, 2026.
What the team transmitted was 49 bits of data: the seven letters of "Gravity" encoded as 7-bit ASCII. The transmission rate was 1 bit per minute, and the shortest distance between the transmitting mass and the receiving sensor was 0.68 (±0.04) m. A brick wall stood between them, blocking communication by sound, visible light and near-infrared light.
Importantly, this communication does not use the "gravitational waves" predicted by Einstein's general theory of relativity. The team used quasi-static changes in the Newtonian gravitational field produced by moving a nearby mass. Communication schemes using gravitational waves or neutrinos, which are also hard for ordinary matter to block, have been proposed, but they face challenges such as requiring enormous sources or detectors. The distinguishing feature of this study is that it demonstrates information transfer using only a moving mass and a commercial gravimeter.
49 bits sent with a 1.8-tonne counterweight and a commercial gravimeter
The experiment took place in a 1950s-era elevator in UQ's Parnell Building and the adjacent restroom. The transmitter was the elevator's counterweight, with an estimated mass of 1,800 (±200) kg. The team moved the weight up and down and measured changes in gravitational pull with a gravimeter placed on the other side of the wall.
The receiver was a commercial spring-based gravimeter, the LaCoste and Romberg Model D. The device measures relative changes in the vertical component of gravitational acceleration, and with about 30 seconds of averaging it achieved a sensitivity of roughly 2 nm/s² at the experimental site.
Signals were encoded using the position of the weight itself. The horizontal distance between the center of mass of the weight in the elevator shaft and the gravimeter was 0.56 (±0.03) m. A "0" was defined as the weight's center of mass sitting 0.39 (±0.06) m above the sensor, and a "1" as it sitting 5.80 (±0.06) m above. Each bit was allotted about 25 seconds for moving the weight, 10 seconds of waiting for vibrations to settle, and 25 seconds for integrating the measured signal, giving a rate of one bit per minute.
To correct for environmental variation and instrument drift in the measurements, the team recorded about 25 minutes of background data before the message and about 20 minutes after. During this time the weight was held at the low "0" position. The message was also bracketed by a 7-bit control sequence, "1010101," which marked the start and end of the signal and told the receiver what bit rate to expect.
From the measurement data, the team subtracted the effects of Earth tides caused by the gravitational pull of the Moon and Sun, as well as the gravimeter's drift over time. They found that each time the weight moved between "0" and "1," the vertical component of gravitational acceleration changed by about 50 nm/s². This change matched the predictions of the team's physical model.
In the end, the 49 bits decoded by the receiver matched the transmitted bit string for "Gravity" exactly. The bit error rate for individual unaveraged measurements was about 0.01 (1%), but after 25 seconds of integration it was estimated to fall to about (one in a billion), assuming the measurement fluctuations follow a Gaussian distribution. The team evaluates this estimate as comparable to the bit error rate of optical fiber communication.
Speed of 1 bit per minute; models suggest range could extend to about 500 m
Although the experiment succeeded in sending and receiving information, the transmission speed is extremely slow. According to the paper, the Shannon channel capacity calculated from the experimental conditions is about 2.4 bits per minute, and the actual rate of 1 bit per minute falls below it. The main factors limiting speed are the old elevator, which takes about 25 seconds to move the weight, and the 40-second low-pass filter built into the gravimeter.
The team also used modeling to estimate how far performance could be improved by upgrading the equipment. The paper sets different conditions for increasing speed and for extending range.
The first scenario is a speed increase using existing technology. Combining a modern elevator that moves the weight in 5 seconds with a newer gravimeter, the Scintrex CG-6, and assuming an integration time of 0.1 seconds and a signal-to-noise ratio (SNR) of about 5, the channel capacity was calculated at about 15 bits per minute. Even in this configuration, however, the time required to move the mass remains a constraint on speed.
The other scenario is range extension. The team estimates that even with the current equipment, lowering the SNR could extend the range to about 1.7 m. Further, Table 1 of the paper considers changing the weight to a nearly spherical shape with an optimized arrangement, increasing its mass to about 75,000 kg, and raising the gravimeter's sensitivity to about 0.01 nm/s². Combining all these improvements, including accepting a low SNR, the team predicts the range could reach about 500 m, roughly 700 times the current distance.
These two estimates need to be kept distinct. In UQ's official announcement, Professor Davis explained that, in contrast to this experiment using classical gravity sensors, model calculations suggest that under fairly ambitious assumptions, using quantum sensors, a range of up to about 500 m and a data rate of about 15 bits per minute might be achievable.
However, in the paper the figure of about 15 bits per minute was calculated for a modern elevator combined with the classical Scintrex CG-6. The roughly 500 m distance, by contrast, is a prediction for a case with major improvements to the weight's shape and mass and to sensor sensitivity. Both are values based on model calculations, and neither performance level has been achieved simultaneously with the same equipment configuration.
As a point of comparison, the paper cites a radio uplink to the Voyager 1 space probe. Even at the distance the paper gives, about m, radio communication reaches about 960 bits per minute. Gravity communication is far inferior to radio in both speed and range, but it has a different characteristic: it is not subject to passive shielding by ordinary matter.
The team envisions uses for gravity communication in harsh or constrained environments where conventional means are unavailable, sending short, simple messages over short distances. Professor White pointed out that it took more than 100 years to go from the first telegraph systems to mobile phones, and said that gravity communication, impractical at present, may become practical in the future.
From contactless vehicle weighing to a comparison with neutrino communication
The experiment was not originally aimed at developing communication technology. According to UQ's official announcement, the research began with the development of a gravity-based weighbridge that measures the mass of large vehicles such as trucks without contact, using the faint gravitational field the vehicle creates. Dr. Groszek explained that when they showed this measurement method worked, they realized the same physics could be used to encode and send information.
Another communication method that ordinary matter has difficulty blocking is the neutrino communication demonstration carried out in 2012 by a team including the US Fermi National Accelerator Laboratory (Fermilab), published in Modern Physics Letters A, vol. 27, 1250077. Comparing the two reveals large differences in the scale of equipment required, the communication distance and the transmission speed.
| Item | Gravity communication (2026, UQ) | Neutrino communication (2012, Fermilab) |
|---|---|---|
| Message sent | Gravity (49 bits) | neutrino |
| Distance | 0.68 (±0.04) m | 1.035 km |
| Barrier traversed | Brick wall | 240 m of bedrock underground |
| Data rate | 1 bit/min | About 6 bits/min (0.1 bit/s) |
| Bit error rate | About 0.01 (raw measurements), about (estimate after averaging) | 1% |
| Transmit/receive equipment | 1,800 kg elevator counterweight, commercial spring gravimeter | Particle accelerator (NuMI beamline), large detector (MINERvA) |
Note: The data rate and bit error rate for neutrino communication are based on decoding results using two frames. The gravity communication figure of about is an estimate calculated after 25 seconds of averaging, assuming Gaussian fluctuations.
In the 2012 neutrino communication experiment, a data rate of about 6 bits per minute was achieved over a distance of more than 1 km, including 240 m of bedrock. That, however, required a huge particle accelerator and detector. UQ's gravity communication, by contrast, covers less than 1 m, but it sends and receives information using existing building equipment and a portable gravimeter.
This result is a proof of concept carried out at a single facility, and no independent replication by other research institutions, outdoor experiments or development of practical communication devices has been reported. Dr. Woffinden also acknowledged that the current system is far from practical, while saying it shows the possibility of communicating by a method that ordinary physical barriers cannot block.
