Scientists have shown off what can be described as the world's smallest electric car - made of a single, carefully designed molecule.
The molecule has four branches that act as wheels, rotating when a tiny metal tip applied a small current to them.
With 10 electric bursts, the car was made to move six billionths of a metre.
The approach, published in Nature, joins recent single-molecule efforts, and seems to overcome the forces that often dominate at such tiny scales.
The "batteries" of the electric car come by way of the tip of what is called a scanning tunnelling microscope - an extraordinarily fine point of metal that ends in just an atom or two. As the tip draws near the molecule, electrons jump into it.
The motor of the approach lies with the four "molecular rotors" that act as the car's wheels; they undergo a change in shape when they absorb the electrons.
The demonstration is a tour de force in what is called "bottom-up" nanotechnology. A wide array of machines has been demonstrated in recent years, incorporating parts etched to minuscule sizes from chunks of metals or semiconductors - a small version of traditional, "top-down" manufacturing.
Building up from single, designed molecules is another matter, said Tibor Kudernac, a chemist now at the University of Twente, the Netherlands, and lead author of the paper.
"If you look around, in all biological systems are a vast number of molecular machines or rotors based on proteins that do important things very well; muscle contraction is based on protein motors," he told BBC News.
"This is a simple demonstration that we can achieve anything like that. It's an important observation and I think it will motivate people to think about it perhaps a bit more from an application point of view."
Dr Kudernac concedes that applications for molecular machines like the car are probably far in the future. The first task, he said, was to make it work under normal conditions; the current work has been done at a blisteringly cold -266C and in a high vacuum.
And although each potential application will require a newly designed molecular machine, Dr Kudernac remains confident.
"There are ways to play around," he said. "That's what we chemists do - we try to design molecules for particular purposes, and I don't see any fundamental limitations."