Superconducting qubits in the millions: the potential and limitations of modularity

This paper presents an architectural model and resource estimation tool for modular superconducting fault-tolerant quantum computers to predict the physical scale, power consumption, and execution time required for practical algorithms, thereby quantifying the bottlenecks and trade-offs involved in scaling to millions of qubits.

S. N. Saadatmand, Tyler L. Wilson, Mark J. Hodson, Mark Field, Simon J. Devitt, Madhav Krishnan Vijayan, Alan Robertson, Thinh P. Le, Jannis Ruh, Alexandru Paler, Arshpreet Singh Maan, Ioana Moflic, Athena Caesura, Josh Y. Mutus

Published Tue, 10 Ma
📖 5 min read🧠 Deep dive

Imagine you are trying to build a supercomputer so powerful it can solve problems that would take today's best supercomputers millions of years to finish. This is the goal of a Fault-Tolerant Quantum Computer (FTQC).

However, building one is like trying to build a skyscraper out of Jell-O. The individual pieces (quantum bits, or "qubits") are incredibly fragile and prone to making mistakes. To fix this, we have to wrap every single fragile piece in a "safety blanket" made of thousands of other pieces. This is called error correction.

This paper, written by researchers from Rigetti and various universities, asks a very practical question: "If we want to build a quantum computer with millions of these safety-blanketed pieces, what does the actual building plan look like, and how much will it cost?"

Here is the breakdown of their findings using simple analogies:

1. The Problem: One Giant Room vs. Many Small Offices

Current quantum computers are like small, noisy rooms. You can't just keep adding more qubits to one room because the heat and wiring get too messy, and the "noise" (errors) becomes unmanageable.

The authors propose a Modular Architecture. Instead of one giant room, imagine a massive office building with thousands of small, identical offices (modules).

  • The Office (Module): Each office holds about one million physical qubits. Inside, they are tightly connected and work very fast.
  • The Hallway (Interconnect): To make the whole building work as one computer, these offices need to talk to each other. The authors propose connecting them with "hallways" (coherent links).
  • The Trade-off: Talking down the hallway is slower and a bit noisier than talking within the office. The paper calculates exactly how much this "hallway penalty" slows down the computer.

2. The Blueprint: The "Ladder" Design

The researchers designed a specific layout for this building, which they call a "Ladder" architecture.

  • Imagine a ladder with two long rails.
  • The "rungs" of the ladder are the connections between the offices.
  • The work is split between the two rails. While one rail is preparing the data (like setting up a chessboard), the other rail is consuming the data (making the moves).
  • Once a move is made, the "chessboard" is teleported across the hallway to the other rail for the next move. This allows the computer to keep working without stopping.

3. The "Magic" Ingredient: T-Factories

To do complex math, quantum computers need a special ingredient called a T-state (or "magic state"). Think of this like a rare, high-quality spice needed for a recipe.

  • The Problem: Making this spice is hard and slow. It requires a dedicated factory (a T-factory) inside every office.
  • The Bottleneck: If you have too many "chefs" (logical qubits) working on the recipe, you might run out of spice factories. If you build too many factories, you don't have enough space for the chefs.
  • The Finding: The paper shows that as problems get bigger, you have to constantly balance how much space you give to the "chefs" vs. the "factories." If you don't balance it right, the computer slows down dramatically.

4. The Software Tool: "Rigetti Resource Estimations" (RRE)

The authors didn't just guess; they built a software tool called RRE.

  • What it does: You feed it a quantum algorithm (like a recipe for simulating a new drug or breaking a code), and it tells you:
    • How many physical qubits you need (the size of the building).
    • How much electricity it will use (the power bill).
    • How long it will take to run (the cooking time).
  • The Result: They tested it on real-world problems, like simulating Fermi-Hubbard models (which helps us understand high-temperature superconductivity).
    • Example: To simulate a specific chemical system, they found you would need about 5 million physical qubits and it would take less than two days to run.
    • The Catch: While two days sounds fast, the energy cost is massive (equivalent to powering a small town for a while), and the physical size is enormous.

5. The Big Takeaway

The paper concludes that while building a quantum computer with millions of qubits is physically possible, it is extremely expensive and complex.

  • It's not just about the math: You have to worry about heat, wiring, and the physical space required.
  • The "Hallway" matters: If the connections between modules aren't fast enough, the whole computer slows down.
  • Optimization is key: We need better software to shrink the "recipes" (algorithms) so they don't require as much space and energy.

In summary: The authors have drawn up a realistic, detailed blueprint for a "Quantum Skyscraper." They show us that while we can build it to solve humanity's hardest problems, we need to be very careful about how we design the elevators (connections), the kitchens (factories), and the power supply, or else the building will be too slow or too expensive to use.