Find the observable.
Identify the quantity that changes scientific understanding. Connect a physical question to a Hamiltonian, an estimator, or a measurable correlation.
Quanta / Scientific frontiers
What if we could
ask nature
directly?
A molecule does not approximate its own quantum mechanics. It lives it. Our ambition is to bring computation closer to that same physical language.
01 / The molecular frontier
A better catalyst begins with a better description of electrons.
Industrial chemistry turns microscopic interactions into consequences at enormous scale. A catalyst changes the pathway of a reaction. Understanding why one candidate works and another does not can depend on subtle differences between electronic states.
In strongly correlated systems, a single electronic configuration may be an inadequate description. Quantum algorithms offer a route to representing the interacting state directly and extracting its energy through carefully designed measurements. The scientific target is a sufficiently accurate energy difference to distinguish competing mechanisms.
H = ∑pq hpqap†aq + ½ ∑pqrs hpqrsap†aq†asar
Here, the one- and two-electron integrals encode a chosen molecular basis; creation and annihilation operators enforce fermionic structure. Mapping this Hamiltonian to qubits is one stage of a larger calculation that includes selecting an active space, preparing a useful state, and controlling approximation error.
Bellonzi and colleagues examine catalyst-discovery problems related to nitrogen fixation through three connected lenses: scientific value, classical difficulty, and fault-tolerant resources. Their analysis makes the opportunity concrete while showing how strongly feasibility depends on the particular chemical problem.
Read the catalyst-discovery resource study ↗The question for Quanta
Which molecular models preserve the decisive chemistry while placing the required calculation within a credible quantum resource budget? That is where chemical insight and algorithm design must meet.
02 / Energy & resilient materials
Look beneath performance.
Understand the mechanism.
A battery electrolyte, a protective coating, and a reactive metal surface seem like different engineering problems. At the molecular scale, each depends on how electrons redistribute, bonds change, and competing configurations exchange stability.
Battery chemistry
Kim and colleagues study fault-tolerant simulations of lithium-ion battery electrolyte molecules. Their resource analysis connects chemistry to the architecture of a photonic quantum computer, including the preparation and consumption of magic states. It illustrates a crucial design opportunity: the way logical operations are scheduled can change the time needed for a calculation.
Read the battery-electrolyte resource analysis ↗Corrosion and interfaces
Corrosion is a multiscale phenomenon, but its initiation involves atomistic interactions at a surface. A 2026 study in npj Quantum Information develops a quantum-computing workflow and resource analysis for corrosion simulation, addressing the difficulty of correlated electronic behavior. The potential service to engineering is a more faithful account of the mechanisms that ultimately govern material degradation.
Read the corrosion-simulation study ↗For Quanta, the objective is to connect a microscopic calculation to a meaningful engineering decision: which observable changes the interpretation, what accuracy is sufficient, and what new experiment would it motivate?
03 / Many-body dynamics
Let a quantum system reveal how information moves.
Equilibrium energies are only part of the story. Interacting systems evolve: correlations spread, local disturbances propagate, and initially accessible information becomes distributed across many degrees of freedom.
W(t) = eiHtWe−iHt
F(t) = ⟨W†(t)V†W(t)V⟩
F(t) = ⟨W†(t)V†W(t)V⟩
This out-of-time-order correlator probes how a time-evolved operator relates to another perturbation. It provides a window into information scrambling, linking the mathematics of operator growth to an experimentally accessible signal.
Google Quantum AI’s 2025 Quantum Echoes work uses forward evolution, a perturbation, and reversed evolution to extract such signals. The team reports a verifiable advantage for a defined family of benchmark circuits and investigates a route toward Hamiltonian learning from nuclear magnetic resonance data. Its significance is the move toward reproducible observables that can encode information about a physical system.
Explore the Quantum Echoes research ↗The question for Quanta
Can carefully selected dynamical observables make an otherwise opaque system identifiable? We connect this question to model reconstruction: finding the interactions that best explain an observed response.
04 / Computing with uncertainty
Estimate the answer.
Change the cost of precision.
Many consequential calculations ask for an expectation rather than an exact trajectory: the average behavior of a model, the likelihood of an event, or a quantity accumulated across a distribution of possible outcomes.
Montanaro’s work on quantum Monte Carlo describes near-quadratic improvements in mean-estimation query complexity under appropriate access and variance assumptions. The key is coherent access to a sampling procedure, enabling interference to extract information more efficiently than independent samples alone.
μ = 𝔼[f(X)] · precision ε · confidence 1 − δ
The opportunity is to change how the cost grows as the requested precision becomes tighter. Realizing it requires an efficient reversible implementation of the model, sufficient coherent depth, and a full accounting of preparation and execution. Our focus is on finding where that complete calculation becomes compelling.
Read Quantum speedup of Monte Carlo methods ↗05 / The algorithmic engine
Engineer the transformation itself.
Qubitization and quantum signal processing provide a systematic way to turn an encoded operator into useful spectral transformations. A Hamiltonian can be embedded in a larger unitary operation; carefully chosen control sequences then transform its spectral information.
Low and Chuang’s qubitization framework establishes strong query-complexity guarantees for Hamiltonian simulation under its oracle-access assumptions. This matters because the cost of simulating evolution is a central ingredient in many energy-estimation algorithms. Efficient operator access, normalization, and the cost of constructing the encoding remain integral to a complete implementation.
Read Hamiltonian Simulation by Qubitization ↗The Quanta approach
From a scientific question
to a computational instrument.
Build the calculation.
Choose a representation, a state-preparation method, and a circuit strategy. Assign accuracy budgets to approximation, evolution, and measurement.
Push the comparison.
Use exact methods where possible, capable classical approximations where necessary, and controlled experiments to understand where each approach succeeds.
Return to the science.
Turn the computational result into a sharper hypothesis, a better model, or a new experiment. The measure of progress is what the answer makes possible.