Cogitan

Simulation economics

Cheap simulations,
without cheap answers.

Most simulation budgets are spent re-solving the same physics at thousands of design points. A learned surrogate solves that physics once, at training time — then answers your whole sweep at millisecond latency, reserving your solver for the candidates that deserve it.

Why simulation gets expensive

A single high-fidelity solve is affordable. The cost explodes because design questions are sweeps, not solves: every grasp angle on a part, every gain combination that keeps a controller stable, every conductivity a material batch might ship with.

At minutes per solve and thousands of points per question, teams end up rationing their own engineering curiosity — exploring less of the design space than they should because compute is metered in cluster-hours.

A worked example · battery aging

One 500-cycle aging simulation (PyBaMM, measured)~2.5 min
Screening 10,000 candidate designs with the solver~17 CPU-days
Same screen through the validated surrogate~3 seconds
Solver runs still needed — for the shortlist that survivesdozens, not 10,000

Live models, published validation

Cheap only counts if the answers hold up.

Every model ships with a validity report: held-out validation against its reference solver, the classical baseline it had to beat, a risk percentile on every prediction — and abstention in the regions the reference physics itself cannot label.

ModelHeld-out accuracySpeed vs. solver

RSFQ screening — Fluxus

Margin, timing, yield & design-rule screening for superconducting logic (MIT-LL SFQ5ee+)

100% DRV detection · 99.1% chain functional match, held-out~10³× vs. JoSIM margin runs

Battery-life screening

500-cycle degradation from 11 cell-design + duty knobs (PyBaMM aging emulator)

0.49% retention-curve MAE · 91.0% conformal coverage @ 90% target~5×10⁵× vs. the aging simulation

When a surrogate is the wrong tool

If you need one certification-grade answer, run the full simulator — that is what it is for. Surrogates also stay inside the input ranges they were validated on; they will not extrapolate to physics they were never trained against.

The economic case is exploration: screen the whole design space for the cost of a lunch, then spend your solver budget on the handful of candidates that survive.

Where the savings come from

Training happens once, on physics generated by the reference solver. Serving a trained model costs milliseconds of CPU time per design — so the screening loop that used to be metered in cluster-hours becomes effectively free, and your solver budget concentrates on the candidates that matter.

FAQ

Simulation cost, answered plainly.

Why are physics simulations so expensive?

A single high-fidelity solve (FEM, CFD, dynamics) costs minutes to hours of CPU time, and real design questions are never a single solve — they are sweeps: every gain setting, every grasp angle, every material variation. Multiply solver time by thousands of design points and you get cluster bills and week-long queues.

What is a simulation surrogate?

A surrogate is a neural network trained on many runs of a reference simulator until it reproduces the simulator's outputs for a validated range of inputs. Inference takes microseconds to milliseconds instead of minutes, which is what makes large sweeps cheap. Cogitan trains surrogates purely on synthetic physics — never on customer data.

How accurate are surrogate models?

Accuracy is per-model and must be measured on held-out cases, not claimed. Cogitan publishes a validity report with every model — including the classical baseline it had to beat: RSFQ screening at 100% design-rule detection and 99.1% chain functional match on held-out cells; battery-life at 0.49% retention-curve error with 91.0% conformal coverage at a 90% target. Inputs outside a model's validated envelope are refused rather than answered, and every prediction carries a risk percentile.

How do engagements work?

A fixed-scope feasibility study first: we take your simulator, build the labeled dataset, benchmark a surrogate against a strong classical baseline, and deliver the validity report — you see the evidence before committing. If the numbers justify it, the surrogate ships as a managed endpoint your team calls at millisecond latency, with retraining as your process or design library evolves.

When is a surrogate the wrong tool?

When you need one high-stakes, certification-grade solve, run the full simulator. Surrogates are also restricted to the input ranges they were validated on — they will not extrapolate to novel physics. The economic case is exploration: screening a large design space cheaply, then spending your solver budget on the handful of candidates that matter. Our feasibility study will tell you honestly if a classical model or your existing solver is the better answer.

Do I need my own GPUs or solver licenses?

No. The delivered surrogate runs as a managed endpoint on Cogitan infrastructure (or on-prem where the data can't leave), and screening results come back at interactive latency. Your solver seats get reserved for the shortlisted candidates that deserve a full-fidelity run.

See the economics live, on real models.

demo.cogitan.ai