⚛️ quantum physics

A Perspective on Quantum Computing Applications in Quantum Chemistry using 25--100 Logical Qubits

This perspective identifies scientifically meaningful use cases for early fault-tolerant quantum computers equipped with 25–100 logical qubits to tackle challenging quantum chemistry problems—such as multireference charge-transfer and conical-intersection states—by leveraging qualitatively distinct strategies like polynomial-scaling phase estimation that remain difficult for classical solvers.

Original authors: Yuri Alexeev, Victor S. Batista, Nicholas Bauman, Luke Bertels, Daniel Claudino, Rishab Dutta, Laura Gagliardi, Scott Godwin, Niranjan Govind, Martin Head-Gordon, Matthew Hermes, Karol Kowalski, Ang L

Published 2026-02-20

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Yuri Alexeev, Victor S. Batista, Nicholas Bauman, Luke Bertels, Daniel Claudino, Rishab Dutta, Laura Gagliardi, Scott Godwin, Niranjan Govind, Martin Head-Gordon, Matthew Hermes, Karol Kowalski, Ang Li, Chenxu Liu, Junyu Liu, Ping Liu, Juan M. Garcia-Lustra, Daniel Mejia-Rodriguez, Karl Mueller, Matthew Otten, Bo Peng, Mark Raugus, Markus Reiher, Paul Rigor, Wendy Shaw, Mark van Schilfgaarde, Tejs Vegge, Yu Zhang, Muqing Zheng, Linghua Zhu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to solve a massive, incredibly complex jigsaw puzzle. This puzzle represents the behavior of atoms and molecules—the building blocks of everything from new medicines to better batteries.

For the last 100 years, scientists have been using classical computers (like the laptop you're reading this on) to try to solve this puzzle. They've gotten very good at it, but they've hit a wall. As the puzzle gets bigger (more atoms involved), the number of possible ways the pieces can fit together explodes exponentially. It's like trying to find a specific grain of sand on a beach that doubles in size every time you blink. Eventually, even the world's most powerful supercomputers run out of time and memory.

This paper, written by a huge team of experts from places like NVIDIA, Yale, and national labs, argues that we are finally ready to bring in a new kind of solver: a Quantum Computer.

But here's the twist: We don't need a giant, perfect quantum computer to start. We just need a "small" one with 25 to 100 "logical qubits" (the quantum equivalent of puzzle-solving brains).

Here is the breakdown of their plan, using simple analogies:

1. The "Small" Quantum Computer is Actually a Big Deal

Usually, people think we need a quantum computer the size of a city to do anything useful. This paper says, "Nope, we can start with a modest one."

The Analogy: Think of a classical computer as a very fast, very organized librarian who reads one book at a time. A quantum computer is like a magical librarian who can read every book in the library simultaneously.
The Catch: The magical librarian is currently very noisy and prone to making mistakes. To fix this, we need to group many physical "noisy" librarians together to act as one "logical" (perfect) librarian. The authors say that once we have 25 to 100 of these perfect logical librarians, we can start solving puzzles that the best human librarians (classical supercomputers) simply cannot crack.

2. What Problems Can We Solve?

The paper focuses on three specific types of chemical puzzles where these 25–100 logical qubits will shine:

The "Strongly Correlated" Team (The Tangled Knot):
- The Problem: In some molecules (like the iron-molybdenum cofactor in bacteria that helps make fertilizer), the electrons are so tangled together that they act as a single, chaotic unit. Classical computers try to untie them one by one and fail.
- The Quantum Fix: Quantum computers are naturally good at handling "entanglement" (tangled states). They can look at the whole knot at once.
- The Strategy: Instead of trying to solve the whole molecule, we use a technique called "Active Space Decomposition." Imagine you have a giant knot, but you only need to untie the center to understand it. We use the quantum computer to untie just the center (the active space) while the classical computer handles the rest of the string.
The "Time-Travel" Problem (Quantum Dynamics):
- The Problem: Chemistry isn't static; it moves. Electrons jump, bonds break, and light hits molecules. Simulating this movement over time is incredibly hard for classical computers because the "movie" gets too long and complex.
- The Quantum Fix: Quantum computers can simulate time evolution naturally.
- The Twist: The paper suggests a clever trick: Using Noise as a Feature. Usually, noise (static) is bad. But in open systems (like a molecule interacting with its environment), noise is actually part of the physics. The authors suggest we might be able to use the natural "noise" of the quantum hardware to mimic the environment, saving us from having to simulate the environment separately.
The "Hybrid" Teamwork (AI + Humans + Robots):
- The Problem: The quantum computer is powerful but fragile. It can't do everything alone.
- The Strategy: We need a Hybrid Workflow.
  - Classical Computers (The Manager): They do the heavy lifting of organizing the data and breaking the big problem into small chunks.
  - AI (The Coach): It helps tune the quantum computer, corrects errors in real-time, and predicts the best way to set up the puzzle.
  - Quantum Computer (The Specialist): It only tackles the tiny, hardest chunk of the puzzle that the Manager and Coach can't solve.

3. The Roadmap: How Do We Get There?

The authors propose a very practical, step-by-step plan rather than waiting for a "magic day" in the distant future.

Step 1: Benchmarking (The Practice Games): Before we try to solve real-world drug discovery, we need to test our tools on known problems. They suggest creating a "scorecard" for quantum computers. If a computer can solve a specific molecule (like a small iron-sulfur cluster) with a known error margin, we know it's ready for bigger challenges.
Step 2: Resource Awareness (Don't Waste Energy): We have limited "logical qubits." We can't just throw a massive algorithm at the problem. We need "measurement-efficient" algorithms—ways to get the answer with the fewest possible tries (shots).
Step 3: Co-Design (Building the House and the Furniture Together): We can't just build a quantum computer and then try to fit chemistry into it. We need to design the hardware, the software, and the chemistry problems together. It's like designing a house and its furniture at the same time so they fit perfectly.

The Bottom Line

This paper is a call to action. It says: "Stop waiting for the perfect, million-qubit quantum computer. Start building and testing with 25–100 logical qubits right now."

By focusing on specific, hard chemistry problems (like designing better catalysts for green energy or understanding how photosynthesis works) and using a team approach (Classical + Quantum + AI), we can achieve "Quantum Utility." This means not just showing off that a quantum computer is fast, but actually using it to discover new scientific truths that were previously impossible to find.

It's the difference between building a toy car that drives in a circle and building a real car that can actually take you to the store. We are finally ready to build the real car.

1. Problem Statement

Quantum chemistry faces an "exponential wall" when simulating strongly correlated electronic systems, open quantum dynamics, and complex excited states (e.g., conical intersections, charge-transfer states). While classical methods like Density Functional Theory (DFT) and Coupled Cluster (CC) have achieved remarkable success, they struggle with:

Strong Correlation: Systems where single-reference methods fail (e.g., transition metal catalysts like FeMoco).
Dynamics: Real-time evolution of open quantum systems and non-adiabatic processes.
Scalability: Full Configuration Interaction (FCI) scales factorially with system size, making exact solutions impossible for chemically relevant active spaces (20–50+ orbitals).

The paper addresses the gap between current Noisy Intermediate-Scale Quantum (NISQ) devices and large-scale fault-tolerant quantum computers (FTQC). It argues that waiting for thousands of logical qubits is unnecessary; a transitional regime of 25–100 logical qubits (expected within 5–10 years) is sufficient to achieve "Quantum Utility"—reliable, validated computations on domain-relevant tasks that outperform brute-force classical methods.

2. Methodology and Framework

The authors propose a holistic framework centered on resource-aware co-design, integrating algorithms, hardware, and classical high-performance computing (HPC) with Artificial Intelligence (AI).

A. The 25–100 Logical Qubit Regime

Definition: This regime represents the first window where quantum devices can access qualitatively distinct strategies (e.g., polynomial-scaling phase estimation) that are intractable for classical solvers.
Target Systems: Compact active spaces (20–50 orbitals) derived from larger molecules via embedding or downfolding. Examples include Fe $_2$ S $_2$ clusters, photochemical chromophores, and transition metal oxides.
Resource Estimates: The paper provides concrete estimates for this regime:
- Logical Qubits: 25–100.
- Circuit Depth: $10^3$ – $10^5$ (depending on the algorithm).
- Shot Budgets: $10^5$ – $10^7$ measurements per observable.

B. Algorithmic Innovations

The paper categorizes algorithms suitable for this regime into four main pillars:

Structured Ansatzes & Measurement Efficiency:
- Moving beyond generic Variational Quantum Eigensolvers (VQE) to chemically inspired ansatzes like Unitary Cluster Jastrow (uCJ).
- Utilizing subspace methods (Quantum Subspace Expansion, Krylov methods) to access multiple states with shallow circuits.
- Implementing measurement reduction techniques (classical shadows, grouping commuting terms) to mitigate the shot noise bottleneck.
Downfolding and Renormalization:
- Techniques like Subsystem Embedding Subalgebras (SES) and Double Unitary Coupled Cluster (DUCC) to integrate out external degrees of freedom.
- This creates compact effective Hamiltonians ( $H_{eff}$ ) that retain essential many-body physics within the limited qubit budget.
Quantum Phase Estimation (QPE) Variants:
- Adopting Iterative, Bayesian, and Statistical QPE to reduce ancilla qubit requirements and circuit depth while maintaining "gold-standard" precision.
- These variants trade depth for repeated measurements, making them viable for early FTQC.
Alternative Paradigms:
- Non-VQA approaches such as Quantum-assisted Monte Carlo and Hamiltonian-moment expansions to bypass barren plateaus and trainability issues.

C. Hybrid Integration Pathways

AI Integration: AI is used end-to-end for hardware calibration, error decoding, circuit synthesis, and measurement reduction (e.g., using generative models for circuit design).
Hybrid Workflows: A "quantum kernel" solves the hardest subproblem (active space) while classical HPC handles the environment (embedding).
Validation Framework: A three-tier validation ladder is proposed:
- Class 0: Infrastructure benchmarks (classically tractable) for reproducibility.
- Class 1: Moderately multireference systems (25–100 qubits).
- Class 2: Strongly multiconfigurational systems (>1000 qubits).
- Use of Stabilizer Circuits (Clifford-only) as classically verifiable checkpoints before running intractable circuits.

3. Key Contributions

Redefining "Quantum Utility": The paper shifts the focus from "Quantum Supremacy" (solving contrived problems) to "Quantum Utility" (solving scientifically meaningful chemistry problems with auditable error bars and resource annotations).
Resource Roadmap: It provides the first detailed resource estimates (qubits, depth, shots) for specific chemical problems (e.g., Fe $_2$ S $_2$ , chromophores) mapped to the 25–100 logical qubit regime.
Strategic Roadmap: A performance-based roadmap (Phases I–III) linking algorithmic milestones to hardware capabilities (code distance, logical error rates) and validation metrics, rather than calendar dates.
Benchmarking Standards: Proposes a standardized "cost-vector" for reporting results (logical qubits, T-counts, shot budgets, post-downfolding term counts) to enable fair cross-platform comparisons.
Hardware-Aware Design: Highlights the potential of hybrid qubit-qumode architectures (using microwave cavities) for problems with strong bosonic character (vibronic dynamics), which may offer advantages over qubit-only circuits.

4. Results and Projections

Feasibility: The authors demonstrate that systems like Fe $_2$ S $_2$ clusters (bioinorganic models) and photochemical chromophores can be mapped to 30–50 logical qubits with circuit depths of $10^4$ – $10^5$ .
Algorithmic Trade-offs: The paper quantifies the trade-off between accuracy and resource cost, showing that downfolded Hamiltonians paired with iterative QPE or subspace methods can achieve chemical accuracy within the 25–100 qubit budget, whereas full FCI would require exponentially more resources.
Validation: The proposed framework allows for the verification of quantum results against high-level classical methods (like DMRG or CCSD(T)) on smaller fragments, ensuring that the quantum solver is providing genuine value before scaling up.

5. Significance and Impact

Paradigm Shift: This paper serves as a strategic guide for the community to move away from waiting for massive FTQCs and instead focus on near-term, resource-constrained utility.
Interdisciplinary Collaboration: It emphasizes that achieving utility requires tight co-design between chemists, computer scientists, and hardware engineers.
Scientific Discovery: By enabling the simulation of strongly correlated transition states and open quantum dynamics, this regime promises to unlock mechanistic insights into catalysis, photochemistry, and materials design that are currently inaccessible to classical computing.
Reproducibility: The call for open, provenance-rich artifacts (inputs, circuits, calibration data) sets a new standard for transparency and reproducibility in quantum chemistry research.

In conclusion, the paper argues that the 25–100 logical qubit regime is not merely a stepping stone but a critical, practical window for demonstrating tangible quantum advantage in chemistry, provided the community adopts resource-aware algorithms, hybrid workflows, and rigorous benchmarking protocols.