Why do they need to be so cold, anyway?
One of the coldest places in the universe isn’t on some distant planet or at an empty point between galaxies. It’s here on Earth, inside a lab. Deep space sits at about 2.7 Kelvin (454.8 degrees below zero in Fahrenheit), in the leftover glow of the Big Bang, according to Space.com. IBM’s latest modular cryogenic system is cooled below 15 millikelvin, which the company describes as more than 180 times colder than deep space. That’s a hair above absolute zero.
So how does IBM get a machine colder than deep space, and perhaps more importantly, why bother? How will it benefit us warm-blooded humans? The answer starts with what’s being kept cold: the qubit.
What is a qubit, and how does it differ from a bit?
Regular computers calculate using bits: tiny switches that are either 1 or 0, on or off. A qubit is closer to a dial than to a switch. It can hold many different positions between on and off. You can imagine the difference between the two like the difference between a light switch and a dimmer. In addition to that, each qubit also has a “direction” that carries information. So while a bit is either on or off, a qubit can hold multiple values in two different directions.
Bits also work alone: Each one switches on or off separately. Qubits can be linked together, so that what you learn about one tells you about the others. Picture a console of connected dials, where turning one shifts the rest along with it.
A quantum computer runs algorithms built so the wrong answers cancel out and the right one is left standing, much like waves on a lake that cancel in some spots and stack up in others. The catch is that the dials are extremely delicate. Any outside disturbance spreads across all of them and raises the odds of a wrong answer. That is where the cold comes in.
How does heat ruin a qubit?
To understand why a quantum computer needs the coldest conditions possible, start with a simple question: What is heat? The answer is straightforward. Heat is energy in the form of motion. When you stand in the sun, the molecules in your body absorb some of its energy and start vibrating faster as a result. The warmth you feel is just those tiny vibrations ramping up.
Qubits are so delicate that even a tiny amount of energy can knock one out of its state, which disturbs the way qubits interact with each other. This loss is called decoherence, and it raises the odds of a wrong answer. The more energy a quantum computer is exposed to, the less accurate its calculations get. That is why IBM and other quantum computing companies are trying to get as close to absolute zero as they can, so there is as little heat as possible to scramble the qubits.
How do dilution refrigerators work?
If you have ever seen a quantum computer, you probably noticed that it looks like a sci-fi chandelier, hanging down and getting narrower toward the bottom. That is the dilution refrigerator, which is a stack of metal plates, each one colder than the one above it. The top sits at room temperature. The bottom is the coldest spot in the machine, and the processor hangs right there. But there is a catch. The wires that control the qubits come from room temperature, and wires conduct heat, so every one of them carries warmth down toward the very thing the fridge is trying to protect. That makes for a paradox: The more qubits a machine has, the more wires it needs, and the warmer it gets.
Does every quantum computer need a giant fridge?
No. Not every quantum computer is built around a giant fridge, though. Neutral-atom machines, like the ones QuEra builds, trap individual atoms in a small vacuum chamber and use lasers to cool them to microkelvin temperatures, colder than any refrigerator can reach. The chamber walls and the equipment around it stay at room temperature. The wiring problem mostly goes away, because there is no cold stack for the wires to warm up. Photonic machines, which use particles of light as qubits, are similar: Many of their parts run at room temperature, but the single-photon detectors typically still need cooling. Each approach trades the cooling problem for a different one, such as trapping thousands of atoms or detecting single photons reliably.
What does quantum computing mean for passwords and bitcoin?
So what happens once we solve the cooling problem and have a working, trustworthy quantum computer? Like most technological leaps, it brings real benefits and real danger. Quantum computers aren’t faster at everything, but for certain problems, such as simulating molecules for a new drug, they could be dramatically faster. They could also break the public-key encryption that protects your logins and bitcoin. A 2025 Google Quantum AI estimate put the cost of cracking a standard RSA-2048 key at under 1 million noisy qubits, down from 20 million in 2019. A 2026 Google paper put bitcoin’s signature scheme at fewer than 500,000. Today’s machines have hundreds to low thousands, so that’s not happening tomorrow. And the fix already exists: The National Institute of Standards and Technology (NIST) finalized its first post-quantum encryption standards in August 2024 and called them ready for immediate use.
For now, the hard part isn’t what quantum computers will do once they work. It’s keeping them cold enough to work at all.
