Skip to content
Technology How it works Breeder — Hyperion Burner — Aegis Burner — MetroVolt AI-Native Architecture Magnets Fuel cycle Safety Roadmap
Solutions AI & Data Centers Defense & Government Grid & Baseload Neutron Detection Quantum
Learn Technical Library
Proof Publications Whitepapers Technical Library Open Science & Reproducibility The Honest Gates
Company About / Mission Leadership Environment Health & Safety Investors Careers Press Contact
3D Model
AI Architecture › L0 · Foundation
L0 · Foundation

Petabyte Pulse Histories

The complete, replayable archive of every pulse and simulation campaign that grounds retraining, refinement, and audit.

THE STACK · click to jumpL7Ecosystem & StrategyL6Experience & VisualizationL5Applications & CopilotsL4OrchestrationL3Twin Modeling & AIL2Data FabricL1Control PlaneL0Foundation▲tlmctl▼L0 · FOUNDATIONThe offline compute substrate — multi-physics & batch training.1Cloud HPCelastic burst2Bare-Metal ClusterGPU / CPU3Supercomputingmulti-physics runs4Batch Trainingmodel builds5Simulation FarmGrad-Shafranov · MHD6Object StorecheckpointsMACHINE TIETrains the models that ship UP to L3 — no real-time path to the machine.KRONOS FUSION ENERGYAI-NATIVE S.M.A.R.T. GENERATORFOUNDATIONSHEET 02REV. 2026-08L0 · AI-NATIVE STACK
L0 · Foundation — its place in the stack (left, click any layer) and its internal components (right). Telemetry rises; control descends.

The archive as ground truth

At the base of L0 sits the pulse-history archive: a petabyte-scale store of complete records from every pulse and every major simulation campaign for both machines. It is the memory of the whole system. Every retraining cycle, every refinement, and every audit reaches into this archive, which is why it is engineered for completeness and replay, not just capacity.

What a pulse record contains

A stored pulse is more than raw waveforms. It bundles the validated diagnostic streams from the full constellation, the engineered features, the twin's predictions and residuals, the control decisions, and the metadata and lineage needed to interpret them. A record is meant to be sufficient to reconstruct and re-analyze the pulse long after it occurred.

Replay, not just storage

The archive is designed for replay. A refinement study can pull an old pulse, feed it back through a new twin model, and compare, exactly the mechanism the refinement loop relies on. This demands not only capacity but read bandwidth and indexing, so studies can find and stream the pulses they need without scanning petabytes.

Because the machines are still design and simulation until FOAK around 2030, today's archive is dominated by simulation campaigns rather than hardware pulses: neutronics sweeps, equilibrium and stability scans, transport runs. These populate the archive so the retraining and refinement machinery is exercised and trusted before the first real pulse arrives.

Archive integrity is treated as sacred. Records are immutable and hashed; a stored pulse cannot silently change, or the models trained on it lose their provenance. This immutability is what lets a deployed twin model be traced, years later, to the exact data that produced it, closing the loop on reproducibility.

Content reviewed August 2026 · design-and-simulation stage