Quantum Gates Explained: What Encryption's Future

Every few months I get a reader email that goes something like this: "Should I care about quantum computing yet?" My honest answer has always been "not urgently, but you should understand what you're waiting on." A new IBM Technology video by Olivia Lanes — a clean, technically careful 10-minute explainer on quantum gates — is exactly the kind of thing I'd point that reader toward, with a few footnotes of my own.

Here's why this matters to people who read security coverage rather than physics journals: the encryption protecting your emails, your banking sessions, your VPN traffic — RSA, elliptic curve cryptography — depends on computational problems that are hard for classical computers and easy for sufficiently powerful quantum ones. The gates Lanes walks through in this video aren't abstract curiosities. They're the mechanical vocabulary of the thing that will eventually (with real caveats on the "eventually") require us to rebuild much of the internet's cryptographic foundation. You don't need to be a physicist to have a stake in understanding them. You just need to care about what replaces AES-256 and elliptic curve Diffie-Hellman when the time comes.

I'll flag upfront: this video comes from IBM, which has significant commercial interests in quantum computing adoption. Lanes is a researcher there. That doesn't compromise the physics she's explaining — the math checks out — but IBM's framing around quantum readiness tends to lean optimistic on timelines. Keep that in your back pocket as you read.

Gates are just instructions. Quantum gates are weirder instructions.

Lanes starts with a useful grounding: a gate, in any computing context, is simply an instruction that changes the state of information. A classical NOT gate flips 0 to 1 and 1 to 0. Simple enough. Quantum gates do the same job for qubits, but qubits aren't bits — they can exist in superposition, a combination of 0 and 1 simultaneously, until you measure them and the system resolves into one or the other.

The gate that creates superposition is called the Hadamard gate. Mathematically, it's a 2x2 matrix. Practically, it takes a qubit sitting firmly in state 0 and puts it into an equal-probability superposition — 50% chance of measuring 0, 50% chance of measuring 1. That's not indecision or noise. That's the system genuinely occupying both states at once until observation forces a resolution.

If that sounds like philosophy, it has experimental teeth. The behavior is real, repeatable, and measurable. The Hadamard gate is the entry point.

Entanglement: the part that breaks local intuition

The second gate Lanes covers is the CNOT — Controlled-NOT — which operates on two qubits simultaneously. Where the Hadamard is a 2x2 matrix acting on one qubit, the CNOT is a 4x4 matrix. It flips the "target" qubit only when the "control" qubit is in state 1. Run both gates in sequence on two qubits that start in the ground state, and you produce what's called a Bell state — a specific entangled configuration.

What makes Bell states significant isn't just that they're mathematically elegant. It's what John Bell's work in the 1960s established about what entanglement implies: no local hidden variable theory can reproduce the predictions of quantum mechanics. To put that less technically — the correlations between entangled qubits cannot be explained by assuming the particles always had definite hidden properties that we simply hadn't measured yet. Bell's theorem doesn't resolve every interpretive debate about what entanglement fundamentally is, but it does establish that you can't explain away the correlations as a measurement artifact or a bookkeeping issue. Subsequent experiments — including rigorously controlled loophole-free tests in 2015 and after — have consistently borne that out.

Lanes puts it this way: "Measure one and you instantly know the state of the other, no matter the distance." That framing is clean, maybe a touch too clean, but the underlying point holds.

For encryption, entanglement is relevant because quantum algorithms like Shor's — the one that can factor large numbers exponentially faster than classical methods — depend on it. That's the algorithm sitting somewhere in the distance that makes every security professional paying attention to quantum timelines vaguely uneasy.

The part IBM doesn't lead with: Clifford gates alone aren't enough

Lanes gets to what I find genuinely interesting here, and it's the piece that most quantum explainers either gloss over or omit entirely.

Both the Hadamard gate and the CNOT gate belong to something called the Clifford group. They produce superposition. They produce entanglement. They feel like the full story of what makes quantum computing quantum. But they're not sufficient.

The Gottesman-Knill theorem establishes that circuits built exclusively from Clifford group gates — no matter how many qubits, no matter how deep the circuit — can actually be efficiently simulated by a classical computer. "Clifford-only circuits are actually classically tractable," as Lanes puts it directly. If quantum computing stopped at Hadamard and CNOT, it would be a fascinating piece of physics with no practical advantage over the computers we already have.

What closes that gap is the T gate. It's almost comically modest-looking: a 2x2 matrix that does exactly one thing — rotates the |1⟩ state by a phase of e^(iπ/4), leaving the |0⟩ state alone. It doesn't change measurement probabilities directly. It doesn't obviously do anything dramatic. But that tiny phase rotation is, as Lanes describes it, "the crack in the classical world."

Add T gates to a Clifford circuit and the simulation cost for a classical computer explodes — exponentially, in the worst cases. The combined Clifford+T gate set achieves what's called approximate universality: the ability to approximate, to arbitrary precision, any quantum operation that physics allows. Not exact universality — the "approximate" qualifier matters, and it's worth sitting with — but close enough that the gap between "approximation" and "exact" is not the practical bottleneck. The bottleneck is building hardware stable enough to run these circuits without decoherence destroying your computation first.

That gap between theoretical universality and working hardware is where most of the honest uncertainty about quantum timelines lives.

What you should actually take from this

The current consensus among people I trust in post-quantum cryptography is that a cryptographically relevant quantum computer — one capable of running Shor's algorithm against real-world key sizes — is probably still ten to twenty years out, and that estimate has meaningful error bars in both directions. NIST finalized its first set of post-quantum cryptographic standards in 2024. Your organization's security team should be looking at those now, not because the threat is imminent, but because cryptographic migrations take years and the organizations that wait for the threat to materialize are the ones that get caught mid-transition.

Understanding the gate architecture Lanes explains — Hadamard for superposition, CNOT for entanglement, T gate for the computational escape from classical tractability — gives you a more calibrated threat model than "quantum computers will break everything eventually, somehow." It helps you ask better questions: What's the actual computational complexity? What error rates make Shor's algorithm viable? How much does T gate overhead affect real-world circuit depth? Those questions have specific answers, and the answers matter for assessing how seriously to treat any given quantum timeline claim, including IBM's.

Lanes closes by observing that "the universe is more than just zeros and ones." True, and a nice line. The more pressing version of that observation, for my readers, is this: the standards that protect our zeros and ones were built for a classical world, and the architecture to challenge them is now well-understood at the gate level. The question isn't whether the cryptographic migration needs to happen. It's whether you're starting early enough to do it deliberately rather than in a panic.

Rachel "Rach" Kovacs is Buzzrag's cybersecurity and privacy correspondent.