An insulated bottle keeps a drink hot for hours without a heater and without any thick layer of insulation. Put a hand on its side and it stays cool to the touch even when the contents are near boiling. The reason is not that the wall is a good insulator. The wall is two very thin shells of stainless steel, and between them there is almost nothing at all. The empty space is the working part of the object.
Heat has three ways to cross a gap: conduction through material, convection in a gas that can move, and radiation. Evacuating the space between the walls removes the first two almost entirely. With very little air left, there are too few molecules to carry energy from the hot inner shell to the cool outer one, and no gas dense enough to set up a circulating current. What a thick wool or foam layer does by trapping air in tiny pockets, this does by taking the air away.
Radiation does not need air, so a vacuum alone is not enough. The inner shell is given a bright, low-emissivity surface — often copper-plated or polished to a mirror — so it is poor at both emitting and absorbing infrared. Heat that would otherwise cross the gap as radiation is mostly reflected back toward the liquid. This layer is invisible to the user and easy to miss when the bottle is cut open, but without it the same vacuum would perform noticeably worse.
Where the two walls have to meet
The shells cannot float apart. They have to be joined somewhere, and the only practical place is at the mouth. So a thin ring of solid steel runs from the inner wall to the outer wall at the rim, and that ring is a thermal bridge: heat travels along the metal from the hot side to the cold side, bypassing the vacuum completely. Designers make this neck as narrow and as thin as strength and the thread for the cap will allow. This is why the mouth of an insulated bottle is smaller than its body, and why the rim is the first place the outside begins to feel warm.
The cap is the larger leak. It closes the one opening that leads straight to the liquid, and it does so with plastic, silicone and an air gap — but every time it is opened, the hot air above the drink is replaced by cool room air and a little steam leaves with it. A bottle drunk from repeatedly loses far more heat this way than through its walls. Fill level matters for the same reason: a half-empty bottle carries a large pocket of air that is heated, vented and reheated at every sip. Keeping it full and opening it rarely is worth more than an extra millimetre of steel.
At the bottom of the vacuum space there is usually a small getter, a patch of reactive material that absorbs the tiny amounts of gas slowly released by the metal over years. Seals are not perfect at this scale, and the vacuum degrades gradually. A dropped flask can fail all at once instead: a dent or a cracked weld lets air in, and the outer wall then warms within minutes of filling. That symptom is diagnostic — the steel is fine, the threads are fine, but the emptiness is gone.
Why thin walls are better than thick ones
It is tempting to imagine that a heavier bottle insulates better. It does not. Thick steel stores heat and conducts it along the shell toward the rim, and it adds weight that has to be carried all day. Thin shells reduce that path, keep the mass low, and leave the vacuum to do the work. The engineering effort goes into holding two thin shells in a stable shape against an external pressure that is trying to squeeze them together, and into the weld at the mouth, not into bulk.
This is the same question asked by where the heat of an LED bulb goes: heat is not stopped, it is given a route, and the design is about which route is available. An LED wants a route out to the air; a flask wants every route out closed. A hollow mug handle is a smaller version of the same thinking — reduce the solid cross-section so there is less material for heat to travel through on its way to a hand.
Failure follows the same map. A switch wears out first because the load concentrates at one small contact; here the load concentrates at the rim and the cap, which is why a gasket that has gone hard or a thread that no longer seats squarely can shorten the holding time by hours. The vacuum in the middle is the part that almost never fails slowly.
So the two walls are not a container with insulation added. They are a deliberate absence with three carefully managed exceptions: a narrow welded ring, a reflective inner surface, and a cap that has to be opened sometimes. Everything else in the design exists to keep that absence intact. The bottle works not because of what it is made of, but because of what was removed from between its walls and then kept out.