The M3BOLT chip: magnetic and sound waves instead of transistors. A promise, not a finished processor

Headlines about “the end of silicon” and a chip that’s supposed to be a thousand times more powerful are grabbing attention. However, the actual news from October 20 and 21, 2026, is more modest and more interesting: An American research team presented a concept for a device called M3BOLT, which would transmit information using magnetic and sound waves instead of electrons. This is not a finished chip, and certainly not a replacement for transistors. It is a scientific project that aims to verify whether it is possible to circumvent the physical limit of current transistors.

Who's behind it

The device is being designed by a team from the University of Michigan in collaboration with the University of Wisconsin-Madison. The research is funded by the U.S. agency DARPA through the Fast and Curious program with up to $6 million. According to available descriptions, the research is in its early stages: the funding is intended to support the construction of devices, circuit design, and simulation.

Why Look for a Replacement Transistor

Researchers report that, as expected, modern transistors will reach their theoretical limit of energy efficiency by 2037, while demand for performance is growing, primarily due to artificial intelligence and edge computing. This is a technical assertion by the team, not a generally accepted term.

How M3BOLT Works

Two-tier structure

The device combines a ferroelectric crystal (PMN-PT) and a thin film of a magnetic iron-gallium alloy. When an electric field is applied, the atoms of the crystal shift, creating mechanical stress that causes the magnetic poles in the alloy to oscillate like a wave. Such a wave is called a magnon.

Magnetic waves and sound waves as 1s and 0s

According to the description, magnetic waves propagate through a material with very low resistance, so less energy is lost as heat. If the orientation of the magnetic poles is reversed, the waves stop and only ordinary sound waves remain. These two states could serve as one and zero instead of the two voltage levels in transistors.

Computations and memory in one

The orientation of the magnetic poles in the alloy is retained even after the power is turned off, so the same material could also serve as memory. Combining computation and memory would reduce the need to move data between different parts of the computer, saving energy and space.

What does “100 to 1,000 times” mean?

The team states that the device could be 100 to 1,000 times faster and more energy-efficient than current technologies. This is a projection by the researchers, not a measured result from a functional chip. According to the descriptions, the researchers are initially building devices for extremely low-power applications and refer to supercomputers as a distant possibility. Therefore, this figure should not be interpreted as a “thousand-times more powerful processor” that would appear in computers.

What is not confirmed

• A functional chip or processor. Sources mention the device’s design and a research program, not a finished processor.
• That it will replace silicon chips. We found no mention of a timeline for deployment.
• That it will “save Moore’s Law.” Researchers do not make such a claim. They speak of overcoming theoretical efficiency limits.
• Independent verification. The data comes from an announcement by the university and its partners, without independent testing.

Why is this still interesting?

The search for alternatives to transistors is important because the energy consumption of artificial intelligence computations is increasing. Methods for utilizing waves instead of electron currents fall within a broader field of research that combines magnetism, materials physics, and electronics. Even if M3BOLT never replaces the transistor, its findings can help us better understand how to process information efficiently.

Conclusion

M3BOLT is a proposal for a potential fundamental component of future computers—not a chip that readers could buy or that would speed up phones and servers anytime soon. The figure “a thousand times” reflects the scientists’ goal and estimate. Until functional prototypes and independent measurements are available, it should be viewed as a promise of a new direction, not as the end of the silicon era.