THE TETRA

01 / Geometric foundation

A tetrahedron, adapted to coastlines.

Station ONE fixes the Australian anchor. TWO, THREE and FOUR are locked reference candidates near the rotated ideal vertices. They are not land acquisitions, surveyed foundations or verified permit boundaries. The earlier ideal coordinates are preserved below exactly as reported to five decimal places.

Coordinate diagram in longitude/latitude; no coastline or territorial boundaries implied. Solid points are coastal candidates; hollow points are the retained ideal vertices. The great-circle table is the authoritative distance view.

02 / Three power classes

The seed, the reference, the horizon.

Why 666.667 MW appears in the GW case: it is a calculated result: 1,000 MW × (1/0.60 − 1) = 666.666… MW of rejected heat at the stated 60% conversion efficiency and zero recirculation. The 60% value is a screening assumption to replace with demonstrated performance. Although 666 is a recurring DCLXVI placeholder motif elsewhere, this output follows the energy balance and changes with the inputs; it was not imposed as a numerical target.

GW: experimental and early-commercial target class for sustained net fusion output, the engine cycle, monolithic fabrication, passive thermal modules, automated inspection, rat net/grid integration and first fabrication cells. Smaller continuous structures and modular thermal circuits precede the TW reference.

TW: 1 TW per station, 4 TW total. The reference envelope has a 2,500 m outer diameter, 76.2 m (250 ft) seawall crest above a future surveyed datum, 2,000 m initial inner tower diameter, 300–500 m tower height and a 250 m radial annulus. These are design inputs, not validated dimensions. If cooling needs more area, grow both diameters.

EW: 1 EW = 1,000,000 TW per station; 4 EW total. The 0.667 EW station heat load alone is about 3.84 times intercepted sunlight. A single 2.5 km surface structure is not a viable extrapolation. Study distributed space industry, direct recovery, off-Earth end use and orbital radiators, with complete energy export accounting. LSM matter/energy conversion is a separately labeled research branch, not an assumed heat sink.

Even perfect direct conversion leaves Earth with heat when the electricity is consumed here. Transporting low-temperature waste heat to a hotter radiator requires work; the radiator equation does not supply that work or its hardware.

03 / Fusion engine

The reaction helps drive the next cycle.

The defining research objective is an integrated engine: introduce fuel → compress → ignite → expand energetic products → extract work → recover compression and confinement energy → prepare the next charge. Like a turbine engine, part of the released fuel energy sustains operation. Self-driving operation after ignition is a target, not fuel-free or loss-free operation.

Conventional fusion research also investigates self-heating and power recirculation. “Fusion engine” names this project’s emphasis on deliberately coupling compression, expansion and recovery; it is not a claim that all conventional reactors ignore those functions.

This repairs the broken equation markup in the supplied brief and distinguishes its accounting boundaries. Internally circulated energy is not counted again as an external output. The calculator assumes Pother = 0, no stored-energy change, and eventual dissipation of recirculated work within the station.

For each candidate record fuel cycle, reaction products, plasma gain Q = Pfusion / Pinjected, driver wall-plug efficiency, gross electrical efficiency, recirculating fraction, startup energy, pulse rate, shutdown inventory and thermal transients. Plasma Q is not plant electrical gain.

Pnet = (η − r) Pfusion; sustained positive net output requires η > r.
Hybrid recovery: η = fcharged ηdirect + (1 − fcharged) ηthermal.

For example, η = 0.60 and r = 0.15 need 2.222 TW of fusion input to deliver 1 TW net, rejecting 1.222 TW under this boundary. The brief’s 1.667 TW input is recovered with r = 0, or with 60% specified as an already-net plant efficiency. Those are different definitions.

D–T provides a useful established reaction reference: about 80% of its reaction energy leaves in neutrons, which are not directly collected by charged-particle converters. Advanced-fuel branches require separate reaction and radiation-loss ledgers. Fuel processing, isotope separation/breeding, cryogenics, magnets, vacuum, control electronics, coolant conditioning and replacement-component manufacturing remain auxiliary loads.

Startup: meter stored bank/flywheel energy, magnet energization, vacuum and heating before ignition. E_start = ∫P_externaldt + ΔE_storage, with the storage term’s sign and boundary declared. No numerical startup energy or gain is asserted before a driver and confinement geometry exist. Shutdown: independently interrupt fuel and drive, safely discharge magnetic/electrical stores, retain local protection, and reject stored/activation heat without the grid.

LSM fusion workstream

Derive a local state/update rule, energy functional, coupling to fields, reaction-product spectra and quantitative rate predictions. Choose an observable where it differs from conventional plasma/nuclear physics; freeze that prediction and uncertainty before testing. A geometric docking path is not a fusion rate. The staff fusion program retains its access restrictions.

04 / Monolithic crystal megastructure

One body. Replaceable organs.

The bulk of the station is one continuous structural body, preferably engineered monocrystalline SiC. Incorporate the seawall, cooling shell, webs, foundation interfaces, fluid channels, exchanger mounts, access galleries, sensors and electrical/electromagnetic service paths during growth. Keep controlled interfaces for replaceable fusion components and equipment.

Cast or grow the station as a single piece

Investigate single-seed growth, controlled directional solidification where the phase diagram permits it, atomically precise deposition, field-directed assembly, self-propagating crystal growth, LSM-derived matter synthesis and in-situ crystalline regeneration. A cast-looking continuous architecture does not mean ordinary molten SiC can simply be poured and cooled into a kilometer-scale monocrystal. Existing SiC growth uses specialized vapor or solution processes; polytype control and growth-front defects are central research problems.

Manufacturing progression: verify composition and crystallographic registry on a coupon; grow embedded channel and sensor features; demonstrate multiple synchronized growth heads retaining registry; measure throughput, impurity and defect budgets; scale to a structural demonstrator with continuous load paths. Conventional feedstock assembly and hypothetical vacuum/substrate synthesis have separate mass and energy ledgers.

Prevent one crack becoming a station-wide failure

Study prestress, graded composition, orientation-aware load paths, redundant webs, crack-blunting channel geometry, residual-stress management, compliant foundation zones and active lattice healing. True single-crystal continuity may conflict with interface-based crack arrest; explicitly compare these tradeoffs rather than silently substituting a block-built structure. Differential settlement, thermal cycling and irradiation are coupled load cases.

Fracture screen: K_I = Yσ√(πa) < measured K_IC
Thermal restraint screen: σ ~ EαΔT / (1 − ν)
Manufacturing rate: ṁ = ρ dV/dt; repair capacity must exceed damage accumulation.

These are first screens, not an anisotropic fracture solution or validated design allowables. Characterize each polytype, orientation and environment. Permanent monitoring needs a defect map, nondestructive inspection, sacrificial test regions, fault-isolation channels and redundant load paths.

Indefinitely renewable, not intrinsically immortal

No known material is immortal. The long-horizon requirement is continuous inspection, lattice repair, replacement machinery and reproduction of the manufacturing system from declared feedstocks. Include energy, consumables, waste separation, tool calibration, sensor replacement and a reserve of independent repair capability. Record repair yield and whether repairs restore load capacity over repeated cycles. Geological-scale operation depends on this renewal system, not on a material name.

05 / THERMOSIPHON TERRA

Let buoyancy do the circulation.

Thousands of independent natural-circulation modules each contain a hot riser, elevated heat rejection, cold downcomer, gravity return, monitoring and isolation. Passive heat exchangers and phase-change transport are candidate branches. Eliminate mechanically driven coolant pumps from the passive heat-rejection circuit while explicitly accounting for auxiliaries elsewhere.

06 / Permanent coastal defense

Two walls, two distinct jobs.

The outer wall protects the land or marine footprint. The inner shell creates natural draft. The 250 m radial annulus houses conversion equipment, thermal plant, distribution, robotic factories, material stores, data centers, seawater processing and inspection galleries. Preserve distinct fire, flood and maintenance zones within the continuous load-bearing body.

The 250 ft crest concept is motivated by the roughly 230 ft (70 m) global sea-level equivalent of melting land ice. It is neither a near-term forecast nor a local flood-height guarantee. Set a surveyed reference datum and combine relative sea level, tides, storm surge, tsunami runup, wave overtopping, uplift/subsidence, erosion, groundwater and long-term circulation change.

For TWO and FOUR, compare seabed-founded and buoyancy-supported configurations with settlement, cyclic waves, uplift, hydrodynamic added mass, tsunami loads and sediment interaction. For ONE and THREE, investigate groundwater ingress, drainage, coastal recession and preemptive defenses. A marine setting is not automatically a deep foundation or unrestricted heat sink.

Site gates: independent property/tenure review, permitting pathway, ecological baseline, geotechnical and bathymetric survey, cooling plume limits, emergency access and grid/logistics study. These investigations have not been completed for the locked coordinates.

07 / The DCLXVI ecosystem

A first unit that can help build its successors.

From four vertices to a planetary network

Compare tetrahedron (4), octahedron (6), cube (8), icosahedron (12), dodecahedron (20) and finer geodesic subdivisions as distinct deployment families. Their vertex sets are not all nested. Preserve the four locked candidates as legacy hubs, then optimize additional stations for coverage, demand, cooling, geology and ecological exclusions. Mathematical symmetry is an organizing tool, not a substitute for siting.

Each station combines fabrication, AI computation, inspection robotics, communications and dispatch with local independent protection. Replication requires a bill of materials, supply-chain closure, measured fabrication yield and manufacturing-time model. Power abundance alone does not establish autonomous reproduction.

At 1 TW, study regional sink capacity, conversion substations, stability, storage, transmission corridors and curtailment. Specify delivered power after losses and the thermal burden of superconducting refrigeration. Communications continuity and grid stability are separate requirements.

08 / Research gates

Prove the enabling systems in parallel.

All gates are proposed, not completed. No guaranteed dates, construction price or achieved performance are assigned without a design and evidence.

Primary reference points / reviewed October 16, 2026

• NASA solar irradiance — 1,361 W/m²; intercepted power here is SπR² ≈ 173.55 PW, distinct from absorbed solar power.
• ITER: making fusion work and fusion-gain measurement — reaction products and the plasma-Q boundary.
• IAEA: natural circulation in water-cooled nuclear power plants — natural-circulation modeling and experimental qualification.
• USGS: all-land-ice sea-level equivalent — approximately 70 m; not a local design flood.
• Wellmann et al.: SiC bulk growth — vapor/solution growth, process control and characterization; no kilometer-scale fabrication demonstration.
• Helion: technology — a developer’s pulsed compression and direct-recovery approach, included as a research comparison rather than independent proof of TETRA feasibility.

Coordinates: author-locked candidates and October 15 ideal-vertex record. Numerical results: the published deterministic model, not measured station performance. Terminology: the project brief’s “Lattice Structure Model” refers to the existing LSM / Lattice Structure Manifold research program.