Types of Quantum Computers and Differences Between Them
Chapter 3 of Quantum Computing Series. Discussion about types of quantum computers, and which are the most viable now, with Joel Pendleton of Conductor Quantum
About the Quantum Computing Series
Understanding quantum computing requires understanding multiple disciplines. That’s why I’m so drawn to write about this topic, but also why I like to invite experts to the discussion. In this chapter, I talk to Joel Pendleton, who withdrew his Oxford PhD to build Conductor Quantum - a company creating quantum computers on silicon chips.
Enjoy the reading & see the complete Quantum Computing Series listed at the conclusion of this post.
Types of Quantum Computers
Quantum computing has multiple paths, each with its own challenges and benefits. The main challenges are decoherence (how long the quantum states remain stable before environmental noise destroys them), scalability (how many qubits can be packed together), and error rates (how often operations fail). The ideal quantum computer would have long coherence times, low error rates, and an easy path to scaling up to millions of qubits - but no current approach hits all three targets.
Superconducting Qubits (Google, IBM): These are the workhorses of the quantum world right now. They're essentially tiny circuits cooled to almost absolute zero that use Josephson junctions. They're relatively mature and easier to control, but they require exotic materials and extreme cooling across larger spaces as you add more qubits.
Trapped Ions (IonQ, Honeywell): Here, individual ions are suspended in electromagnetic fields and manipulated with lasers. Ion trap systems have absurdly long coherence times (qubits stay quantum for longer) and great fidelity, but they're painfully slow and struggle with precisely controlling large numbers of ions.
Neutral Atoms: Similar concept to trapped ions but using neutral atoms in optical lattices instead. They have decent coherence times and can scale to hundreds of qubits, but controlling individual atoms remains tricky. They're the middle-ground approach - not the best at anything but solid across the board.
Semiconductor/Spin Qubits: These use the spin states of electrons in semiconductors (like silicon). They're potentially the most scalable since they piggyback on existing chip fabrication tech - you could theoretically pack billions on a chip. They're also more energy-efficient and operate at higher (though still very cold) temperatures. The catch is that they're currently hard to control precisely and have higher error rates.
Topological Qubits (Microsoft): Instead of particles, these use exotic quasiparticles that form braids, which are theoretically immune to local noise - potentially solving the error correction problem that plagues other approaches. The massive downside is they're still largely theoretical.
Reading
Quantum Decoherence on Wikipedia
DiVincenzo's criteria on Wikipedia
Let's talk about what types of quantum computing exist today. What approach is Conductor Quantum using, and why?
There are several major approaches to building quantum computers. The most common is superconducting quantum computing, used by Google and IBM. Other popular methods include trapped ions and neutral atoms.
At Conductor Quantum, we're focused on semiconductor quantum computation—specifically, spin qubits. This approach uses the spin states of electrons in semiconductor devices as qubits. Despite its potential, semiconductor quantum computing hasn't received as much attention as other methods until recently.
What advantages does semiconductor quantum computing offer?
The semiconductor quantum computing approach is good in three critical areas:
First is scalability. Since we can leverage modified versions of the existing technology of chip fabrication, we could theoretically print billions of transistors on a single chip, each potentially hosting a qubit. This gives us a clear path to millions of qubits.
Second is energy efficiency. Our approach requires an order of magnitude less power than superconducting qubits. The physical components are also much smaller, eliminating the need for football field-sized refrigeration systems.
Third, coherence time. One of the biggest challenges in quantum computing is decoherence—how quickly quantum information degrades. Spin qubits in semiconductors can maintain their quantum states significantly longer than competing architectures, allowing for more operations before information is lost.
What materials are critical for quantum computing, and what are the biggest obstacles to scaling up qubits?
One claim Microsoft made in their press release is that their topological approach could eventually enable "millions of qubits in the palm of your hand." This is a bold statement, considering how challenging it is to produce even hundreds of stable qubits today. The scalability challenge is precisely where we at Conductor Quantum focus our efforts. Our approach with semiconductors is built on silicon-based quantum computing. It’s using silicon chips that are made in a similar way to the chips that you use in your smartphone.
Silicon is abundant and we already have decades of manufacturing expertise with it. The challenge isn't the raw material itself, but rather the extreme purity requirements and precision engineering needed for quantum applications. While a regular computer chip can tolerate some minor defects, quantum computing requires near-perfect environments to maintain fragile quantum states.
Can we also discuss superconducting quantum computing, which is really popular now? What makes it good?
Google and IBM are using superconducting qubits built around Josephson junctions—two superconductors separated by an extremely thin insulating barrier (2-3 nanometers). This barrier is so thin that electrons can tunnel through it—a quantum phenomenon impossible in classical physics.
When cooled to near absolute zero, these junctions develop unique quantum properties ideal for computing: they maintain coherence, provide well-isolated energy states, and allow precise control of superpositions.
The advantages include faster gate operations, relatively mature technology with 50-100+ qubit systems, and strong coupling between qubits. Challenges include requiring extreme cooling, shorter coherence times than some alternatives, and physical scaling limitations that may eventually constrain system size—explaining why companies continue pursuing multiple quantum computing approaches.
Let’s talk more about the Josephson junction. Josephson junctions involve superconducting materials separated by a thin insulator, but how does this method actually create a qubit?
Superconducting and spin qubits (semiconducting approach) create the necessary two-level system in fundamentally different ways:
In superconducting circuits, the Josephson junction creates "anharmonicity" in the energy levels. Without this junction, energy levels would be equally spaced like a ladder with uniform steps. The problem with evenly spaced energy levels is that when you try to manipulate a qubit (say, moving between levels 0 and 1), you might accidentally provide enough energy to jump multiple levels (to level 2 or higher).
The Josephson junction - that “sandwich” of superconductors with an insulating barrier - changes this behavior. When incorporated into a superconducting circuit, it creates an "anharmonic oscillator" where:
The energy gap between levels 0 and 1 is larger
The gap between levels 1 and 2 is smaller
The gap between levels 2 and 3 is even smaller, and so on
This uneven spacing is crucial because:
It lets us target just the 0→1 transition with a specific energy pulse
The energy needed for 0→1 won't accidentally cause 1→2 because that requires a different energy
We can essentially ignore all higher energy levels and treat the system as a true two-level qubit.
Is the problem of equally spaced energy levels the same for semiconducting quantum computers that use spin qubits? Or how is the approach to making a two-level system different there?
Spin qubits have a natural advantage - electrons intrinsically exist in only two spin states (up or down). This provides a ready-made two-level system. We just need to ensure these states have different energies when measured, usually by applying a magnetic field. There's no need to engineer complex energy level spacing because nature has already limited our options to exactly two states.
Both approaches achieve the same goal: a well-defined two-level system that can represent 0 and 1 simultaneously through quantum superposition. Once you have this, quantum computing requires just two more capabilities: initializing qubits to a known state and performing operations that transform these states according to quantum algorithms.
Let’s also talk about topological quantum computing. We already discussed the advantages of it, but I've been trying to get some intuition behind these Majorana fermions Microsoft is using. They seem different from the electron spin states in your semiconducting approach. Practically, how exactly are quantum states encoded with these exotic quasiparticles?
In every quantum computing system, information is ultimately encoded in the quantum states of physical systems. While the physical implementation varies, we always rely on distinguishing between different quantum states, that’s important to understand.
For Microsoft's topological qubits using Majorana fermions, the specifics are complex, but this core principle remains the same. These quasiparticles will have different quantum states that can be addressed and read. Their recent paper claims they can detect these distinct quantum states of Majorana fermions, which is the first step toward using them for computation in the future.
An energy-based approach is commonly seen in quantum physics. Even in our spin-based qubits, we're not directly measuring the spin itself. Instead, we apply a magnetic field that creates an energy difference between spin-up and spin-down states. This energy gap is what allows us to distinguish between the states. Without that applied field creating an energy difference, we couldn't tell the states apart.
So regardless of the physical system—whether exotic Majorana fermions or electron spins—quantum computing always comes down to detecting and manipulating different quantum states.
Read more of the Quantum Computing Series:
Chapter 1: Microsoft And Google's Quantum Launches
Chapter 2: Building a Quantum Computer
Chapter 3: Types of Quantum Computers
Chapter 4: Quantum Computers Market Landscape
Chapter 5: Topological Quantum Computers Explained
Resources to start with quantum computing
What Quantum Computing Isn't by Scott Aaronson
Learn the Algorithms Behind Quantum Computing by Beau Carnes
Landscape of Quantum Computing in 2024 by Samuel Jaques
Quantum Computing on Wikipedia