On March 3, 2026, a team at the University of Tokyo watched an electron flip its spin in 140 picoseconds. Not simulated. Not inferred after the fact. Observed, directly, using ultrafast electrical pulses and precisely timed light flashes in a material called Mn₃Sn, an antiferromagnetic Weyl semimetal. That sentence deserves a moment. Scientists engineered a material, fired a current through it, and caught the magnetization switching in real time at speeds that make your laptop's transistors look geological.
This is not incremental. Silicon transistors operate in nanoseconds. Mn₃Sn switches in picoseconds. That is three orders of magnitude. The engineers who built that experiment did not just measure something interesting; they demonstrated that the physical ceiling for computing speed is somewhere we have not reached yet.
Two Switches, One Material, One Giant Implication
What makes the Tokyo result sharp is the discovery of two distinct switching mechanisms inside the same material. Strong currents flip the spin through heat. Weak currents do it non-thermally, without the energy waste. Ryo Shimano, who led the research, put it plainly: the material itself could switch even faster under appropriate conditions. The 140-picosecond figure is a measurement limit tied to current pulse generation technology, not a hard wall in the physics.
That distinction matters enormously for anyone building memory or logic devices. Non-thermal switching means low energy cost at extreme speed. That combination is exactly what the next generation of computing needs, whether you are running an AI inference engine, a quantum timing system, or a satellite navigation array that cannot afford a single dropped signal.
Fair point to skeptics: Mn₃Sn is a research material, not a wafer rolling off a fab line. Scaling exotic antiferromagnets into commercial devices is genuinely hard, and nobody should pretend otherwise. But the history of semiconductor engineering is a history of "we cannot fabricate that at scale" becoming routine within a decade. The physics here is sound. The engineering challenge is the known kind.
Electrons as Engineering Components
Meanwhile, at Cambridge, a separate team revealed something equally precise. They showed that electrons in engineered solar materials move at near-maximum speeds through what they call a molecular catapult: vibrational bursts within molecules that kick electrons across the structure before energy bleeds away as heat. Sub-picosecond timescales. The mechanism essentially mimics what photosynthesis has been doing for billions of years, except now scientists understand it well enough to build it deliberately.
The tension I will admit: these two results, spintronics and molecular electron transport, come from very different corners of physics. Connecting them into a single narrative about "controlling electron flow" risks overselling a unified field that does not quite exist yet. They are separate bets on the same underlying idea: that electrons are engineering components, not just passengers in a circuit.
I think that idea is correct, and I think the industry should act on it now. The semiconductor roadmap has been running on fumes for years, squeezing geometry and chasing clock speeds with diminishing returns. Spintronics and molecular electronics are not future options. They are the actual next chapter, and the labs doing this work deserve funding at the scale we give to silicon optimization.
Shimano's team measured 140 picoseconds. The next team, with better pulses and optimized device structures, will measure less. That is how physics works. The electron does not care about your product roadmap; it just flips when you tell it to.
