DEV Community

dosanko_tousan
dosanko_tousan

Posted on

Nuclear Fusion from Scratch — Vol.8: Alternative Confinement — The Roads Not (Yet) Taken

#nuclear #plasma

Series: Nuclear Fusion from Scratch (Vol.8 of 10)
Author: dosanko_tousan — 50-year-old stay-at-home father, no engineering degree
License: MIT — Use freely. Build on it. Fix what I got wrong.
Prerequisite: Volumes 1–7. Physics constraints established there apply here without re-derivation.
Honest disclosure: I am not a plasma physicist. Every claim is sourced. Gaps are marked "Unknown."

Executive Summary

The tokamak is the most mature path to fusion power. It is also, as Volumes 1–6 demonstrated, burdened with problems that may prove fatal: disruptions that can destroy first walls in milliseconds, a tritium breeding ratio that fails in 88% of simulated scenarios, and materials that have never been tested beyond 20 dpa in a fusion neutron spectrum.

This volume asks: what if the tokamak is the wrong shape?

Six alternative confinement concepts are now being pursued with serious capital and serious physics: the stellarator, the field-reversed configuration (FRC), inertial confinement fusion (ICF), the Z-pinch, the spherical tokamak, and magnetized target fusion (MTF). Each trades one set of problems for another. None escapes the constraints of Volumes 3–5 (tritium, materials, neutronics) if they burn D-T fuel. Some attempt to escape those constraints entirely by burning advanced fuels — and run straight into the wall identified in Volume 2 (bremsstrahlung radiation losses make thermal ignition of p-¹¹B impossible) and Volume 6 (D-³He ignition requires 17× higher triple product than D-T).

The landscape in early 2026:

Concept Leading Experiments / Companies TRL Key Advantage Key Risk
Stellarator W7-X, Proxima Fusion, Type One Energy 4–5 No disruptions, steady-state Manufacturing complexity, no D-T experience
FRC TAE (Norm), Helion (Polaris) 3–4 Compact, high-β, aneutronic fuel compatible Stability scaling unproven, confinement time short
ICF (Laser) NIF 5 Ignition achieved (gain 4.13) Repetition rate, target manufacturing, wall efficiency
ICF (Pulsed power) Pacific Fusion, Sandia Z-machine 2–3 Potentially simpler driver than lasers Unproven at fusion-relevant scale
Z-Pinch Zap Energy (FuZE-3) 3 No magnets, no cryogenics, compact Stability scaling, electrode erosion
Spherical Tokamak Tokamak Energy (ST80-HTS), UK STEP 4–5 Higher β than conventional tokamak, compact Center column neutron damage, plasma exhaust
MTF General Fusion 2–3 Intermediate parameter space Near-death funding crisis (2025), unproven physics

This is not a ranking. It is a map. The physics does not care about our preferences.

Table of Contents

§1. Why Alternatives Matter

The tokamak has consumed the majority of global fusion funding for sixty years. ITER, the flagship international tokamak, has cost an estimated $45–65 billion and will not achieve D-T operations until 2039 — sixteen years behind the original schedule (Vol.7, §2). Commonwealth Fusion Systems (CFS), the best-funded private tokamak company at $3 billion in total capital, plans first plasma for SPARC in late 2026 or early 2027.

The tokamak's dominance rests on one fact: it has the highest demonstrated triple product ($n T \tau_E$) of any magnetic confinement device. JET achieved Q = 0.67 in 1997. No other magnetic confinement concept has come close to breakeven.

But the tokamak carries intrinsic problems that are not merely engineering challenges — they are consequences of the topology itself:

  1. Disruptions: The large plasma current required for confinement in a tokamak can terminate abruptly, dumping megajoules of energy into the first wall in milliseconds. Stellarators, FRCs, and Z-pinches either eliminate or drastically reduce this risk.

  2. Pulsed vs. steady-state: A tokamak's plasma current must be driven either inductively (inherently pulsed) or by auxiliary current drive systems that consume significant power. Stellarators are inherently steady-state. This matters for baseload electricity generation.

  3. Aspect ratio constraints: The conventional tokamak's geometry (aspect ratio $A = R_0/a \approx 3$) limits the achievable plasma pressure relative to magnetic pressure (β). Spherical tokamaks ($A \approx 1.5$) and FRCs ($\beta \approx 1$) can operate at much higher β, meaning more fusion power per unit of magnetic field.

  4. Neutron damage to the center stack: In a conventional tokamak, the inner leg of the toroidal field coil is relatively protected. In a spherical tokamak, the center column is exposed to the full neutron flux. In a stellarator, the complex 3D coil geometry creates both shielding challenges and opportunities.

None of these problems is guaranteed to be fatal. But the fact that $15.2 billion in private capital is now distributed across at least six distinct confinement concepts (FIA, September 2025) suggests that the market — if not the physics community — is hedging its bets.

§2. The Stellarator — Twisted Perfection

2.1 The Physics

A stellarator confines plasma using external magnetic coils alone, with no net toroidal plasma current. The confining magnetic field is entirely produced by external coils that are twisted into complex three-dimensional shapes. This eliminates the tokamak's most dangerous failure mode: current-driven disruptions.

The key physics parameter is rotational transform ($\iota$), which describes how magnetic field lines wind helically around the torus. In a tokamak, the rotational transform comes primarily from the plasma current. In a stellarator, it comes from the coil geometry. This distinction has profound consequences:

  • No disruptions: No large plasma current means no sudden current termination. The plasma can be switched off simply by turning off the heating — it does not store magnetic energy in a way that can be released destructively.
  • Steady-state operation: The external coils provide the confining field continuously. There is no need for current drive, no transformer flux consumption, no duty cycle. A stellarator can, in principle, run for months or years without interruption.
  • Neoclassical transport: The 3D magnetic field geometry introduces additional particle drift orbits (banana orbits are more complex). Historically, stellarators suffered from poor neoclassical confinement. The breakthrough of quasi-isodynamic (QI) optimization — pioneered at the Max Planck Institute for Plasma Physics and embodied in Wendelstein 7-X — dramatically reduced these losses.

The price is manufacturing complexity. A tokamak's toroidal field coils are planar and identical. A stellarator's coils are non-planar, each with a unique 3D shape, and must be aligned to sub-millimeter tolerances over structures spanning several meters.

2.2 Wendelstein 7-X: The World's Most Advanced Stellarator

Wendelstein 7-X (W7-X), operated by the Max Planck Institute for Plasma Physics in Greifswald, Germany, is the world's largest and most powerful stellarator. It cost approximately €1.06 billion (1997–2014) and began experimental operation in December 2015.

Key results (through OP2.3 campaign, ending May 22, 2025):

Parameter Value Significance
Triple product record $n T \tau_E > 3 \times 10^{21}$ m⁻³ keV s (43 seconds) World record for sustained duration (>30 s), surpassing JET and EAST for long pulses
Energy turnover 1.8 GJ (360-second plasma) Previous record: 1.3 GJ (February 2023)
Volume-averaged β 3% Target for power plant: 4–5%
Ion temperature ~40 million °C At record β
Plasma duration >8 minutes demonstrated (2023) Stellarator world record

The May 2025 results are significant because they show that a stellarator can match tokamak-level triple product at the pulse durations relevant for power plant operation. At short pulses (<1 second), tokamaks still hold the record. But a power plant does not run for one second.

W7-X is currently in a one-year maintenance phase and will resume operations in September 2026.

2.3 Private Stellarator Companies

Proxima Fusion (Munich, Germany)

Proxima Fusion, the first spin-out from the Max Planck IPP, has raised over €185 million ($200M) including a €130 million ($150M) Series A in June 2025 — the largest private fusion investment in European history.

Proxima's approach: QI-HTS stellarators — combining the quasi-isodynamic optimization demonstrated in W7-X with high-temperature superconducting (HTS) magnets. HTS enables much stronger magnetic fields, which means smaller devices. Their published power plant concept, Stellaris, is the first peer-reviewed stellarator design integrating physics, engineering, and maintenance from the outset.

Timeline:

  • 2027: Stellarator Model Coil (SMC) — de-risk HTS magnet technology for non-planar stellarator coils
  • 2031: Alpha demonstration stellarator — target Q > 1
  • 2030s: Stellaris commercial power plant

Type One Energy (Madison, WI / Knoxville, TN, USA)

Type One has raised $160+ million (including an $87M raise in January 2026, ahead of a $250M Series B). Backed by Bill Gates' Breakthrough Energy Ventures.

Type One's approach differs from Proxima's: they use a different stellarator optimization (based on HSX heritage from the University of Wisconsin), and have partnered directly with the Tennessee Valley Authority (TVA), the largest U.S. public utility, to build a 350 MWe fusion power plant (Infinity Two) at TVA's retired Bull Run coal plant site near Oak Ridge, Tennessee.

Timeline:

  • 2026: Construction of Infinity One prototype stellarator begins
  • 2029: Infinity One operational
  • Mid-2030s: Infinity Two (350 MWe) operational

Type One completed the first formal design review of Infinity Two in May 2025, with external reviewers from Princeton Plasma Physics Laboratory and Westinghouse.

2.4 Stellarator: Honest Assessment

Strengths:

  • No disruptions — the single most dangerous failure mode in tokamaks is eliminated by design
  • Inherently steady-state — no current drive power required
  • W7-X has demonstrated tokamak-level triple product at long pulse durations
  • Two well-funded private companies (Proxima €185M, Type One $160M+) with distinct technical approaches
  • IEEE Spectrum and Journal of Plasma Physics peer-reviewed power plant designs published for both companies

Weaknesses:

  • No stellarator has ever operated with D-T fuel. W7-X has produced zero fusion neutrons. The Large Helical Device (LHD) in Japan is the only stellarator to have produced fusion neutrons (via neutral beam injection)
  • Manufacturing complexity of non-planar coils remains a critical risk. HTS tape is expensive and supply chains are immature
  • The β = 3% achieved in W7-X is still below the 4–5% needed for a power plant
  • All materials and tritium breeding challenges from Volumes 3–5 apply equally to D-T stellarators
  • Stellarators are approximately 20 years behind tokamaks in demonstrated fusion performance

§3. Field-Reversed Configuration — The Compact Rebel

3.1 The Physics

A Field-Reversed Configuration (FRC) is a compact toroid in which the confining magnetic field is created primarily by the plasma itself. The plasma carries a toroidal current that generates a poloidal magnetic field. Inside the separatrix (the boundary between closed and open field lines), the magnetic field reverses direction relative to the external applied field — hence "field-reversed."

Key physics properties:

  • β ≈ 1: The FRC operates at plasma pressures comparable to the magnetic pressure. This is dramatically higher than tokamaks (β ~ 5%) or stellarators (β ~ 3%). High β means more fusion power output for a given magnetic field strength, which translates directly to smaller, cheaper magnets.
  • Compact linear geometry: An FRC is essentially a cigar-shaped plasma confined within a cylindrical coil set. No complex 3D coils, no massive toroidal structures. This simplicity is commercially attractive.
  • No toroidal field: Unlike tokamaks and stellarators, FRCs have no (or negligible) toroidal magnetic field component. This makes them attractive for advanced fuel cycles (D-³He, p-¹¹B) because the low interior magnetic field reduces synchrotron radiation losses at extreme temperatures.

The fundamental challenge: FRC stability. The classic tilt instability should destroy an FRC on the Alfvén timescale (~microseconds). In practice, large-orbit kinetic ion effects stabilize FRCs far beyond MHD predictions, but the physics of this stabilization is not fully understood, and it is unclear how it scales to reactor-relevant parameters.

3.2 TAE Technologies: The NBI-Only Breakthrough

TAE Technologies (Foothill Ranch, CA), founded in 1998, has raised $1.79 billion and is the most experienced private FRC company. TAE pursues the ambitious goal of p-¹¹B (proton-boron) fusion — the cleanest possible fuel, which produces no neutrons in the primary reaction.

The Norman → Norm breakthrough (2025):

In April 2025, TAE published results in Nature Communications demonstrating the first-ever formation of an FRC plasma using only neutral beam injection (NBI). This eliminated the traditional theta-pinch formation hardware (lengthy quartz tubes and supersonic collision sections), reducing machine length and complexity by up to 50%.

The new machine, Norm (a shortened version of its predecessor Norman), routinely achieves TAE's highest steady-state plasma performance. The results were strong enough that TAE skipped its planned sixth-generation machine (Copernicus) entirely, moving directly to development of Da Vinci, its first prototype power plant.

Updated TAE roadmap:

  • Norm (current): NBI-only FRC formation, continued data collection
  • Da Vinci: First prototype power plant (early 2030s)
  • Commercial power: Target mid-2030s (pending net energy demonstration)

In December 2025, TAE announced a merger with Trump Media & Technology Group in a $6 billion deal, making it the first publicly traded fusion company (expected mid-2026).

The p-¹¹B question: As established in Volume 2, thermal ignition of p-¹¹B is physically impossible because bremsstrahlung radiation losses exceed fusion power by a factor of ~23 at relevant temperatures. TAE's response is that they do not pursue thermal ignition — their beam-driven approach maintains non-Maxwellian ion distributions that shift the effective reaction rate. Whether this approach can achieve net energy at reactor scale is an open physics question with no experimental answer yet.

3.3 Helion Energy: The Pulsed FRC

Helion Energy (Everett, WA) has raised $1.03 billion and pursues D-³He fusion using a pulsed, magneto-inertial approach. Two FRC plasmoids are formed at opposite ends of the machine, accelerated to supersonic velocities, and collided in the center. The merged FRC is then magnetically compressed.

Polaris, Helion's seventh prototype, was turned on in January 2025. Key features:

  • Target: demonstrate net electricity generation
  • Microsoft Power Purchase Agreement: 50 MWe by 2028 (extremely aggressive timeline)
  • Nucor Steel: 500 MWe for industrial heat
  • Omega manufacturing facility: Production of thousands of capacitor units for the Orion commercial power plant

Helion's approach is physically distinct from TAE's: where TAE pursues steady-state beam-driven FRCs at high temperature, Helion pursues pulsed compression at high density. The confinement time is short (microseconds), compensated by extremely high density during compression.

The D-³He fuel choice avoids the tritium breeding problem (Vol.3) but faces the 17× higher triple product requirement established in Vol.6. Helion's published scaling from Trenta (9 keV, 8 T) to Polaris requires significant extrapolation.

3.4 FRC: Honest Assessment

Strengths:

  • Highest β of any magnetic confinement device — orders of magnitude more efficient use of magnetic field
  • Simple linear geometry enables modular, compact designs
  • TAE's NBI-only breakthrough dramatically simplifies reactor design
  • Advanced fuel compatibility (low interior B-field favors p-¹¹B and D-³He)

Weaknesses:

  • No FRC has demonstrated Q > 0.01. The gap to breakeven is enormous
  • Stability at reactor-relevant parameters is unproven. Kinetic stabilization may not scale
  • Confinement times are orders of magnitude shorter than tokamaks
  • Helion's 2028 Microsoft PPA timeline requires physics breakthroughs not yet demonstrated
  • TAE's p-¹¹B pathway faces the bremsstrahlung barrier (Vol.2, §4)
  • Neither company has published peer-reviewed data from their latest machines (Polaris, Norm)

§4. Inertial Confinement Fusion — The Other Path

4.1 The Physics

Inertial confinement fusion (ICF) takes the opposite approach to magnetic confinement: instead of holding a low-density plasma for seconds with magnetic fields, ICF compresses a tiny fuel pellet to extreme density for nanoseconds, relying on the fuel's own inertia to hold it together long enough for fusion to occur.

The key parameter is the areal density ($\rho R$), which must exceed approximately 0.3 g/cm² for D-T ignition. At these conditions, the alpha particles produced by fusion reactions are stopped within the fuel itself, depositing their energy and creating a self-sustaining burn wave.

The Lawson criterion for ICF, expressed differently than for magnetic confinement, requires:

$$\rho R \gtrsim 0.3 \text{ g/cm}^2, \quad T_i \gtrsim 5 \text{ keV}$$

Achieving these conditions requires compressing the fuel by a factor of 30–40 in radius, reaching densities of 300–1000 g/cm³ (roughly 1000× the density of lead) and temperatures of 50–100 million °C.

4.2 NIF: Ignition Achieved, Repeated, and Scaled

The National Ignition Facility (NIF) at Lawrence Livermore National Laboratory achieved fusion ignition on December 5, 2022, producing 3.15 MJ of fusion energy from 2.05 MJ of laser energy delivered to the target — the first controlled fusion experiment to exceed unity target gain.

NIF ignition shots (through October 2025):

Date Laser Energy (MJ) Fusion Yield (MJ) Target Gain
Dec 5, 2022 2.05 3.15 1.54
Jul 30, 2023 2.05 3.88 1.89
Oct 8, 2023 1.9 2.4 1.26
Oct 30, 2023 2.2 3.4 1.55
Feb 12, 2024 2.2 5.2 2.36
Feb 23, 2025 2.05 5.0 2.44
Apr 7, 2025 2.08 8.6 4.13
Jun 22, 2025 — 2.4 >1 (LANL-led THOR experiment)
Oct 1, 2025 2.065 3.5 1.74

The April 2025 shot is extraordinary: target gain of 4.13 — quadrupling the input energy. This was achieved with essentially the same laser energy (~2 MJ), through improved target design and implosion symmetry.

However, NIF is not a power plant prototype. The 192-beam laser system consumes approximately 300 MJ of wall-plug electricity per shot. The overall wall-plug gain ($Q_{\text{eng}}$) is therefore approximately $8.6 / 300 \approx 0.03$. A commercial IFE plant would need $Q_{\text{eng}} > 10$ and a repetition rate of 5–15 Hz (NIF fires a few shots per day).

4.3 Pacific Fusion: Pulsed Power ICF

Pacific Fusion (Fremont, CA) raised a record $900+ million Series A in 2024 to pursue pulsed-power ICF — replacing NIF's laser driver with electromagnetic pulses.

The core technology: Impedance-Matched Marx Generators (IMGs), co-invented by Pacific Fusion CTO Keith LeChien. The system uses 156 pulser modules, each containing 320 capacitor "bricks" organized into 32 rings, to deliver ~2 TW for 100 nanoseconds. The pulses converge simultaneously on a D-T fuel capsule inside a deionized water tank.

This approach is inspired by Sandia National Laboratories' Z-machine, which uses pulsed power to drive Z-pinch and MagLIF experiments. Pacific Fusion's innovation is in the IMG architecture, which achieves ~90% electrical-to-pulse efficiency (vs. NIF's ~1% wall-plug to laser efficiency).

Status (early 2026):

  • Phase 1 milestones completed November 2024 (ahead of June 2025 schedule)
  • Brick and stage prototypes built and tested
  • Collaboration with General Atomics on production-scale IMG module testing
  • CRADAs signed with Sandia (December 2024) and LLNL (January 2025)
  • Target: net facility gain by 2030

Demonstration system specifications:

  • 156 IMG pulser modules, spherically arranged
  • Footprint: 73 × 80 meters (approximately one soccer field)
  • Target yield: >100 MJ per shot
  • Target: "100-fold higher gain and 10-fold less costly than NIF"

Pacific Fusion is led by CEO Eric Lander (who led the Human Genome Project) and president Will Regan. The team includes veterans from NIF, Sandia, and the national laboratory ICF programs.

4.4 Other ICF Companies

Focused Energy (Austin, TX / Darmstadt, Germany): Pursuing laser-driven ICF with direct-drive targets, distinct from NIF's indirect-drive approach. Team includes former NIF campaign leaders. Developing high-repetition-rate diode-pumped solid-state lasers.

First Light Fusion (Oxford, UK): Originally pursued projectile-driven ICF (no lasers, no pulses — a physical projectile fired at a target). In March 2025, First Light pivoted away from building its own power plant, instead offering its core technologies to other companies and developing a "pulsed power capability" with defense and science applications. Total funding: $100+ million.

4.5 ICF: Honest Assessment

Strengths:

  • NIF has achieved ignition — the only confinement approach with demonstrated target gain > 1
  • Target gain of 4.13 (April 2025) shows rapid improvement through target engineering
  • Pulsed-power approaches (Pacific Fusion) offer dramatically higher driver efficiency than lasers
  • ICF physics is on solid theoretical footing with decades of weapons-program validation
  • The modularity of pulsed-power systems enables cost reduction through mass production

Weaknesses:

  • Wall-plug gain remains ~0.03 for NIF. The gap to $Q_{\text{eng}} > 10$ is enormous
  • Repetition rate: NIF fires a few shots per day. A power plant needs 5–15 Hz. Target manufacturing at this rate (and cost) is an unsolved problem
  • First wall survival: each fusion shot deposits MJ of energy (neutrons, X-rays, debris) onto chamber walls. Material survival at 5+ Hz is unproven
  • Pacific Fusion's IMG technology has never driven a fusion target. The extrapolation from Sandia Z-machine to commercial scale is large
  • NIF was built for weapons stockpile stewardship, not energy. Converting NIF-class physics to commercial IFE requires an entirely new facility class
  • First Light's pivot away from power plants is a warning signal for the ICF pathway

§5. Z-Pinch — The Oldest Idea, Reinvented

5.1 The Physics

A Z-pinch uses electrical current flowing through a plasma column to generate a magnetic field that compresses (pinches) the plasma. It is, conceptually, the simplest possible magnetic confinement device: no external magnets, no cryogenics, no complex coil geometries. Just current through plasma.

The compressive magnetic pressure scales as $B_\theta^2 / 2\mu_0 \propto I^2$, where $I$ is the current. At sufficient current (~MA), the pinch can heat and compress the plasma to fusion conditions.

The problem — known since the 1950s ZETA experiments — is instability. A static Z-pinch is violently unstable to sausage ($m = 0$) and kink ($m = 1$) modes on the Alfvén timescale. By the early 1960s, Z-pinch fusion research was largely abandoned.

5.2 Zap Energy: Sheared-Flow Stabilization

Zap Energy (Everett, WA / San Diego, CA) has revived the Z-pinch using sheared-flow stabilization (SFS), a concept developed in the 1990s by co-founder Uri Shumlak at the University of Washington with LLNL collaborators.

The idea: if the plasma flows axially at different velocities at different radii (like a river where the center flows faster than the edges), the velocity shear can suppress the growth of instabilities. Theoretical analysis shows that if the flow velocity gradient exceeds the Alfvén speed gradient, all MHD modes are stabilized.

FuZE-3 results (November 2025, APS-DPP):

  • Electron pressure: 830 MPa (1.6 GPa total plasma pressure)
  • This is ~10,000× atmospheric pressure, comparable to pressures deep in the Earth's crust
  • First Zap device with a three-electrode configuration, separating plasma acceleration from compression
  • Density exceeded $10^{24}$ m⁻³
  • Previous device FuZE-Q demonstrated D-D fusion neutron production rates of $5 \times 10^7$ neutrons/μs at 3 keV electron temperature

Century platform (power plant technology):

  • Repetitive pulsed power, high-duty-cycle cathodes, liquid metal wall systems
  • October 2025: >100 plasma shots at one shot every five seconds (12 per minute), delivering 39 kW average power
  • The liquid metal first wall concept addresses the neutron damage problem differently from magnetic confinement — by using a flowing liquid (lithium or lead-lithium) that self-heals and breeds tritium

Zap's reactor concept is a ~2-meter-long device. The company employs 150 people and has raised ~$200 million. The plasma is a ~50 cm long flowing pinch, ~0.15 mm in radius, at densities of $10^{26}$ m⁻³. This is a fundamentally different parameter space from tokamaks (meters of plasma, $10^{20}$ m⁻³ density).

5.3 Z-Pinch: Honest Assessment

Strengths:

  • Radical simplicity: no magnets, no cryogenics, no lasers. The cheapest possible fusion device per unit of confinement
  • Compact: a 2-meter core vs. a 30-meter tokamak
  • Liquid metal first wall concept elegantly addresses neutron damage and tritium breeding simultaneously
  • High repetition rate already demonstrated (12 Hz on Century platform)
  • Published peer-reviewed results in Fusion Science and Technology (2025)

Weaknesses:

  • 1.6 GPa pressure is impressive, but the triple product achieved is still far from breakeven
  • Stability at reactor-relevant parameters is an extrapolation from current experiments
  • Electrode erosion at high repetition rates is an unsolved engineering problem
  • The physics of sheared-flow stabilization at high performance is not fully understood theoretically
  • No company or laboratory has demonstrated Q > 0.001 in a Z-pinch
  • The technology readiness level is lower than stellarators or tokamaks

§6. Spherical Tokamak — The Squeezed Doughnut

6.1 The Physics

A spherical tokamak (ST) is a tokamak with a very low aspect ratio ($A = R_0/a \approx 1.5$, vs. $A \approx 3$ for conventional tokamaks). The result looks less like a doughnut and more like a cored apple.

Physics advantages of the spherical geometry:

  • Higher achievable β: The Troyon β-limit scales as $\beta_N \propto I_p / (a B_T)$. At low aspect ratio, the plasma current $I_p$ can be higher relative to $a B_T$, allowing β values of 20–40% (vs. ~5% in conventional tokamaks). More fusion power per unit of magnetic field.
  • Strong natural elongation: The geometry naturally supports highly elongated plasmas, which have better confinement.
  • Compact size: A 200 MWe spherical tokamak can potentially be much smaller than a conventional tokamak of the same power, reducing capital cost.

The fundamental challenge: the center column. In a conventional tokamak, the inner leg of the toroidal field coil is shielded from the plasma by significant distance. In a spherical tokamak, the center column passes directly through the plasma — it is thin, close to the plasma, and exposed to intense neutron irradiation. This column must carry enormous current density while surviving the full fusion neutron spectrum.

6.2 Tokamak Energy: ST80-HTS

Tokamak Energy (Oxford, UK) is the leading private spherical tokamak company, having raised $250 million. Founded in 2009 as a spin-out from the Culham Centre for Fusion Energy (home of JET).

ST40: Their previous device achieved 100 million °C plasma temperature in 2022 — a world record for a compact spherical tokamak — and the highest triple product of any private fusion company.

ST80-HTS (planned completion: 2026):

  • World's first high-field spherical tokamak with HTS magnets at scale
  • Target: higher sustained triple product than any previous fusion device
  • Long pulse: ~15 minutes (vs. seconds for most tokamaks)
  • Will inform design of ST-E1 fusion pilot plant

ST-E1 (planned: early 2030s):

  • Target: up to 200 MWe net electrical power
  • Grid-ready fusion

6.3 UK STEP Program

The UK government's STEP (Spherical Tokamak for Energy Production) program, managed by the UK Atomic Energy Authority, has received £2.5 billion in funding. The site is the former West Burton A coal plant in Nottinghamshire.

STEP is a government-backed program using the spherical tokamak concept, distinct from Tokamak Energy's private effort. Target: ~100 MWe by approximately 2040.

6.4 Spherical Tokamak: Honest Assessment

Strengths:

  • Tokamak physics is the most mature in fusion — decades of validated confinement scaling
  • Higher β than conventional tokamaks means more efficient use of magnetic field
  • Compact size enables factory manufacturing and standardized deployment
  • HTS magnets (shared with CFS's approach) are rapidly maturing
  • UK government commitment (STEP, £2.5B) provides long-term stability

Weaknesses:

  • Center column neutron damage is the critical unsolved problem. No material has been demonstrated to survive the neutron fluence a center column would experience in a power plant
  • The center column limits shielding for the HTS magnets — they may degrade faster than in a conventional tokamak
  • ST80-HTS has not yet operated. Claims of record-setting performance are projections, not data
  • All D-T constraints from Volumes 3–5 apply: tritium breeding, first wall damage, neutron activation
  • Disruption risk is reduced (higher natural β and stronger shaping) but not eliminated — spherical tokamaks still have a plasma current

§7. Magnetized Target Fusion — The Middle Path

7.1 The Physics

Magnetized target fusion (MTF), also called magneto-inertial fusion, occupies the parameter space between magnetic confinement (low density, long confinement time) and inertial confinement (ultra-high density, nanosecond confinement time). MTF operates at intermediate densities ($10^{24}$–$10^{26}$ m⁻³) and microsecond timescales.

The concept: create a magnetized plasma target, then compress it mechanically or electromagnetically. The embedded magnetic field helps retain heat during compression, reducing the required compression ratio compared to pure ICF.

7.2 General Fusion: A Cautionary Tale

General Fusion (Vancouver, Canada) has been the most prominent MTF company, having raised $492 million in total. Their concept: compress a magnetized plasma target using a spherical array of pneumatic pistons driving a liquid metal liner.

2025 was a crisis year:

  • Spring 2025: ran short of cash while building LM26 (its latest device targeting breakeven demonstration in 2026)
  • 25% staff layoff days after hitting a key milestone
  • CEO Greg Twinney published an open letter pleading for emergency funding
  • August 2025: $22 million pay-to-play round (described by one investor as "the least amount of capital possible" to keep the company alive)
  • November 2025: $51.1 million in SAFE notes from nearly 70 investors

General Fusion's near-death experience is a preview of what happens when private fusion companies burn through capital before demonstrating physics milestones. The company survives, but its credibility and timeline are severely damaged.

7.3 MTF: Honest Assessment

Strengths:

  • Intermediate parameter space may offer advantages over both magnetic and inertial extremes
  • Liquid metal compression simultaneously serves as first wall, neutron shield, and tritium breeder
  • Lower required magnetic field than magnetic confinement

Weaknesses:

  • General Fusion's 2025 crisis demonstrates the funding vulnerability of MTF
  • No MTF experiment has demonstrated significant fusion output
  • The physics basis is less mature than either magnetic or inertial confinement
  • Mechanical compression introduces engineering complexity (piston synchronization, liquid metal management)
  • The concept has not attracted the capital that stellarators, FRCs, or ICF have

§8. Cross-Cutting Physics Comparison

All magnetic confinement approaches face the same fundamental challenge: achieving sufficient triple product ($n T \tau_E$) for net energy production. The Lawson criterion requires:

$$n T \tau_E \gtrsim 3 \times 10^{21} \text{ m}^{-3} \text{ keV s} \quad (\text{for D-T at } T \approx 15 \text{ keV})$$

Each concept occupies a different region of the ($n$, $T$, $\tau_E$) parameter space:

Concept Density (m⁻³) Temperature (keV) Confinement Time (s) Path to Lawson
Tokamak $10^{20}$ 10–25 1–10 Increase $\tau_E$ through better confinement
Stellarator $10^{20}$ 5–10 (current) 1–100 (demonstrated 8+ min) Increase $T$ and $n$ while maintaining steady-state
FRC (TAE) $10^{20}$–$10^{21}$ 5–10 (current) 0.001–0.01 Increase $\tau_E$ through better stabilization
FRC (Helion) $10^{22}$–$10^{23}$ 10+ (during compression) $10^{-6}$–$10^{-3}$ Increase $n$ and $T$ through stronger compression
Z-Pinch $10^{24}$–$10^{26}$ 1–5 (current) $10^{-6}$–$10^{-4}$ Increase $T$ while maintaining stability
Spherical Tokamak $10^{20}$ 10–20 1–10 (target 15 min) Same as tokamak but at higher β

ICF operates in a fundamentally different regime: $n \sim 10^{31}$ m⁻³, $T \sim 5$ keV, $\tau_E \sim 10^{-10}$ s. The Lawson criterion is replaced by the ρR criterion.

Python: Figure 1 — Alternative Confinement Physics Landscape (click to expand)

"""
Vol.8 Figure 1: Alternative Confinement — Physics Comparison
Seed-fixed, fully reproducible.
"""
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches

np.random.seed(42)

fig, axes = plt.subplots(2, 2, figsize=(16, 12))
fig.suptitle('Figure 1: Alternative Confinement — Physics Landscape',
             fontsize=16, fontweight='bold', y=0.98)

# ── Panel A: Parameter Space Map (density vs temperature) ──
ax1 = axes[0, 0]

concepts = {
    'Tokamak\n(ITER target)':  {'n': 1e20, 'T': 15, 'tau': 3.0, 'color': '#2C5F8A', 'marker': 's', 'size': 300},
    'Stellarator\n(W7-X 2025)': {'n': 8e19, 'T': 3.5, 'tau': 0.3, 'color': '#2ECC71', 'marker': '^', 'size': 250},
    'FRC\n(TAE Norm)':        {'n': 3e19, 'T': 5, 'tau': 0.005, 'color': '#F39C12', 'marker': 'D', 'size': 200},
    'FRC\n(Helion compress)': {'n': 1e23, 'T': 10, 'tau': 1e-5, 'color': '#E67E22', 'marker': 'D', 'size': 200},
    'Z-Pinch\n(Zap FuZE-3)':  {'n': 1e24, 'T': 3, 'tau': 1e-5, 'color': '#E74C3C', 'marker': 'v', 'size': 200},
    'ICF\n(NIF ignition)':    {'n': 1e31, 'T': 5, 'tau': 1e-10, 'color': '#9B59B6', 'marker': '*', 'size': 400},
}

for name, d in concepts.items():
    ax1.scatter(d['T'], d['n'], s=d['size'], color=d['color'],
                marker=d['marker'], alpha=0.85, edgecolors='black', linewidth=1, zorder=5)
    offset_x = 0.05 * d['T'] + 0.5
    ax1.annotate(name, xy=(d['T'], d['n']),
                 xytext=(d['T'] + offset_x, d['n']),
                 fontsize=7, fontweight='bold', va='center')

T_range = np.linspace(1, 30, 100)
n_lawson = 3e21 / (T_range * 1.0)
ax1.plot(T_range, n_lawson, 'k--', alpha=0.3, linewidth=1, label='Lawson (τ=1s)')

ax1.set_xlabel('Ion Temperature (keV)', fontsize=11)
ax1.set_ylabel('Plasma Density (m⁻³)', fontsize=11)
ax1.set_title('(A) Parameter Space: Density vs Temperature', fontweight='bold')
ax1.set_yscale('log')
ax1.set_xlim(0, 25)
ax1.set_ylim(1e18, 1e32)
ax1.legend(fontsize=8)
ax1.grid(alpha=0.3)

# ── Panel B: Triple Product Progress ──
ax2 = axes[0, 1]

devices = {
    'JET (1997)':        {'tp': 1.5e21, 'year': 1997, 'color': '#2C5F8A'},
    'JT-60U (1998)':     {'tp': 1.5e21, 'year': 1998, 'color': '#2C5F8A'},
    'TFTR (1994)':       {'tp': 1e21, 'year': 1994, 'color': '#2C5F8A'},
    'W7-X (2025)':       {'tp': 3e21, 'year': 2025, 'color': '#2ECC71'},
    'EAST (2025)':       {'tp': 5e20, 'year': 2025, 'color': '#E74C3C'},
    'ST40 (2022)':       {'tp': 5e19, 'year': 2022, 'color': '#9B59B6'},
    'FuZE-Q (2024)':     {'tp': 1e16, 'year': 2024, 'color': '#E74C3C'},
    'NIF (2025)\n(ICF, ρR equiv)': {'tp': 5e21, 'year': 2025, 'color': '#9B59B6'},
}

for name, d in devices.items():
    ax2.scatter(d['year'], d['tp'], s=150, color=d['color'],
                alpha=0.8, edgecolors='black', linewidth=1, zorder=5)
    ax2.annotate(name, xy=(d['year'], d['tp']),
                 xytext=(d['year'] + 0.5, d['tp']),
                 fontsize=7, fontweight='bold', va='center')

ax2.axhline(y=3e21, color='red', linestyle='--', alpha=0.5, linewidth=2, label='D-T Ignition Threshold')
ax2.set_xlabel('Year', fontsize=11)
ax2.set_ylabel('Triple Product (m⁻³ keV s)', fontsize=11)
ax2.set_title('(B) Triple Product Progress by Concept', fontweight='bold')
ax2.set_yscale('log')
ax2.set_xlim(1990, 2028)
ax2.set_ylim(1e15, 1e23)
ax2.legend(fontsize=9)
ax2.grid(alpha=0.3)

# ── Panel C: β comparison ──
ax3 = axes[1, 0]

concepts_beta = {
    'Conv. Tokamak\n(ITER)': 2.5,
    'Spherical\nTokamak': 25,
    'Stellarator\n(W7-X)': 3,
    'FRC\n(TAE/Helion)': 90,
    'Z-Pinch\n(Zap)': 95,
}

names = list(concepts_beta.keys())
betas = [concepts_beta[n] for n in names]
colors_beta = ['#2C5F8A', '#9B59B6', '#2ECC71', '#F39C12', '#E74C3C']

bars = ax3.barh(names, betas, color=colors_beta, alpha=0.85, height=0.6, edgecolor='black', linewidth=0.5)
for bar, val in zip(bars, betas):
    ax3.text(bar.get_width() + 1, bar.get_y() + bar.get_height()/2,
             f'{val}%', va='center', fontsize=10, fontweight='bold')

ax3.set_xlabel('Plasma β (%)', fontsize=11)
ax3.set_title('(C) Achievable β by Concept', fontweight='bold')
ax3.set_xlim(0, 110)
ax3.axvline(x=5, color='gray', linestyle=':', alpha=0.5)
ax3.text(6, -0.3, 'Tokamak β limit', fontsize=8, color='gray', alpha=0.7)
ax3.grid(axis='x', alpha=0.3)

# ── Panel D: Capital raised vs TRL ──
ax4 = axes[1, 1]

companies_trl = {
    'CFS':          {'capital': 3.0, 'trl': 7, 'color': '#2C5F8A', 'concept': 'Tokamak'},
    'Proxima':      {'capital': 0.2, 'trl': 4, 'color': '#2ECC71', 'concept': 'Stellarator'},
    'Type One':     {'capital': 0.16, 'trl': 4, 'color': '#27AE60', 'concept': 'Stellarator'},
    'TAE':          {'capital': 1.79, 'trl': 4, 'color': '#F39C12', 'concept': 'FRC'},
    'Helion':       {'capital': 1.03, 'trl': 3, 'color': '#E67E22', 'concept': 'FRC'},
    'Pacific':      {'capital': 0.9, 'trl': 2, 'color': '#9B59B6', 'concept': 'ICF'},
    'Zap':          {'capital': 0.2, 'trl': 3, 'color': '#E74C3C', 'concept': 'Z-Pinch'},
    'Tok. Energy':  {'capital': 0.25, 'trl': 5, 'color': '#3498DB', 'concept': 'Sph. Tok.'},
    'Gen Fusion':   {'capital': 0.49, 'trl': 2, 'color': '#95A5A6', 'concept': 'MTF'},
}

for name, d in companies_trl.items():
    ax4.scatter(d['trl'], d['capital'], s=d['capital']*200 + 80,
                color=d['color'], alpha=0.75, edgecolors='black', linewidth=1, zorder=5)
    ax4.annotate(f"{name}\n({d['concept']})", xy=(d['trl'], d['capital']),
                 xytext=(d['trl'] + 0.2, d['capital'] + 0.08),
                 fontsize=7, fontweight='bold')

ax4.set_xlabel('Technology Readiness Level (TRL)', fontsize=11)
ax4.set_ylabel('Total Capital Raised ($B)', fontsize=11)
ax4.set_title('(D) Capital vs. TRL by Concept (size ∝ capital)', fontweight='bold')
ax4.set_xlim(0.5, 8.5)
ax4.set_ylim(0, 3.8)
ax4.grid(alpha=0.3)
ax4.axvspan(6, 9, alpha=0.05, color='green', label='Demonstration+')
ax4.axvspan(1, 3, alpha=0.05, color='red', label='Lab scale')
ax4.legend(fontsize=8, loc='upper left')

plt.tight_layout(rect=[0, 0, 1, 0.95])
plt.savefig('vol8_fig1_physics.png', dpi=200, bbox_inches='tight',
            facecolor='white', edgecolor='none')
plt.close()
print("Figure 1 saved.")
Enter fullscreen mode Exit fullscreen mode

IMAGE_URL_PLACEHOLDER

graph TD
    A[Fusion Confinement Approaches] --> B[Magnetic Confinement]
    A --> C[Inertial Confinement]
    A --> D[Hybrid / Intermediate]

    B --> B1[Tokamak<br/>JET, ITER, CFS SPARC<br/>TRL 7-8]
    B --> B2[Stellarator<br/>W7-X, Proxima, Type One<br/>TRL 4-5]
    B --> B3[Spherical Tokamak<br/>Tokamak Energy, STEP<br/>TRL 4-5]
    B --> B4[FRC<br/>TAE, Helion<br/>TRL 3-4]
    B --> B5[Z-Pinch<br/>Zap Energy<br/>TRL 3]

    C --> C1[Laser ICF<br/>NIF, Focused Energy<br/>TRL 5]
    C --> C2[Pulsed Power ICF<br/>Pacific Fusion, Z-machine<br/>TRL 2-3]

    D --> D1[MTF<br/>General Fusion<br/>TRL 2-3]

    style B1 fill:#2C5F8A,color:#fff
    style C1 fill:#E74C3C,color:#fff
    style B2 fill:#2ECC71,color:#fff
    style B4 fill:#F39C12,color:#fff
Enter fullscreen mode Exit fullscreen mode

§9. The Fuel Question Revisited

Python: Figure 2 — Funding & Commercialization Landscape (click to expand)

"""
Vol.8 Figure 2: Alternative Confinement — Funding &amp; Timeline Landscape
Seed-fixed, fully reproducible.
"""
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches
from matplotlib.patches import FancyArrowPatch

np.random.seed(42)

fig, axes = plt.subplots(2, 2, figsize=(16, 12))
fig.suptitle('Figure 2: Alternative Confinement — Funding &amp; Commercialization Landscape',
             fontsize=16, fontweight='bold', y=0.98)

# ── Panel A: Funding by Concept (stacked bar) ──
ax1 = axes[0, 0]

concepts = ['Tokamak', 'FRC', 'Stellarator', 'ICF\n(pulsed power)', 'Sph. Tokamak', 'Z-Pinch', 'MTF']
public_funding = [20.0, 0.5, 1.2, 15.0, 2.8, 0.3, 0.1]
private_funding = [4.5, 2.9, 0.4, 1.0, 0.3, 0.2, 0.5]

x = np.arange(len(concepts))
width = 0.5

bars1 = ax1.bar(x, public_funding, width, label='Public/Government', color='#2C5F8A', alpha=0.85, edgecolor='black', linewidth=0.5)
bars2 = ax1.bar(x, private_funding, width, bottom=public_funding, label='Private', color='#E74C3C', alpha=0.85, edgecolor='black', linewidth=0.5)

for i, (pub, priv) in enumerate(zip(public_funding, private_funding)):
    total = pub + priv
    ax1.text(i, total + 0.3, f'${total:.1f}B', ha='center', fontsize=8, fontweight='bold')

ax1.set_xticks(x)
ax1.set_xticklabels(concepts, fontsize=9)
ax1.set_ylabel('Cumulative Funding ($B)', fontsize=11)
ax1.set_title('(A) Estimated Cumulative Funding by Concept', fontweight='bold')
ax1.legend(fontsize=10)
ax1.grid(axis='y', alpha=0.3)
ax1.set_ylim(0, 28)

# ── Panel B: Timeline to First Electricity ──
ax2 = axes[0, 1]

projects = [
    ('ITER\n(D-T operations)', 2039, 2042, '#2C5F8A', 'Tokamak'),
    ('CFS SPARC\n(first plasma)', 2027, 2029, '#3498DB', 'Tokamak'),
    ('CFS ARC\n(commercial)', 2032, 2035, '#3498DB', 'Tokamak'),
    ('Proxima Alpha\n(demo)', 2031, 2033, '#2ECC71', 'Stellarator'),
    ('Type One\nInfinity Two', 2033, 2037, '#27AE60', 'Stellarator'),
    ('TAE Da Vinci\n(pilot)', 2032, 2035, '#F39C12', 'FRC'),
    ('Helion Polaris→\nOrion', 2028, 2032, '#E67E22', 'FRC'),
    ('Pacific Fusion\nDS (demo)', 2029, 2031, '#9B59B6', 'ICF'),
    ('Tok. Energy\nST-E1', 2032, 2035, '#E74C3C', 'Sph. Tok'),
    ('UK STEP', 2038, 2042, '#C0392B', 'Sph. Tok'),
    ('Zap Energy\n(pilot)', 2032, 2036, '#E74C3C', 'Z-Pinch'),
]

for i, (name, start, end, color, concept) in enumerate(projects):
    ax2.barh(i, end - start, left=start, height=0.6, color=color, alpha=0.75,
             edgecolor='black', linewidth=0.5)
    ax2.text(start - 0.3, i, name, ha='right', va='center', fontsize=7, fontweight='bold')

ax2.axvline(x=2026, color='red', linestyle='--', alpha=0.5, linewidth=2, label='Now (Feb 2026)')
ax2.set_xlabel('Year', fontsize=11)
ax2.set_title('(B) Projected Timeline to Key Milestones', fontweight='bold')
ax2.set_xlim(2025, 2045)
ax2.set_yticks([])
ax2.legend(fontsize=9)
ax2.grid(axis='x', alpha=0.3)

# ── Panel C: NIF Ignition Gain Progression ──
ax3 = axes[1, 0]

nif_dates = ['Dec\n2022', 'Jul\n2023', 'Oct\n2023', 'Oct\n2023\n(2.2MJ)', 'Feb\n2024', 'Feb\n2025', 'Apr\n2025', 'Jun\n2025\n(THOR)', 'Oct\n2025']
nif_gains = [1.54, 1.89, 1.26, 1.55, 2.36, 2.44, 4.13, 1.0, 1.74]
nif_yields = [3.15, 3.88, 2.4, 3.4, 5.2, 5.0, 8.6, 2.4, 3.5]

colors_nif = ['#9B59B6' if g &gt;= 2.0 else '#3498DB' for g in nif_gains]

bars = ax3.bar(range(len(nif_dates)), nif_gains, color=colors_nif, alpha=0.8,
               edgecolor='black', linewidth=0.5)

for i, (bar, gain, yld) in enumerate(zip(bars, nif_gains, nif_yields)):
    ax3.text(i, bar.get_height() + 0.08, f'Q={gain:.2f}\n{yld} MJ',
             ha='center', fontsize=7, fontweight='bold')

ax3.set_xticks(range(len(nif_dates)))
ax3.set_xticklabels(nif_dates, fontsize=7)
ax3.set_ylabel('Target Gain (Yield / Laser Energy)', fontsize=11)
ax3.set_title('(C) NIF Ignition Shots — Target Gain Progression', fontweight='bold')
ax3.axhline(y=1.0, color='red', linestyle='--', alpha=0.5, linewidth=1, label='Gain = 1')
ax3.set_ylim(0, 5.0)
ax3.legend(fontsize=9)
ax3.grid(axis='y', alpha=0.3)

ax3.annotate('Record: Q = 4.13\n8.6 MJ yield',
             xy=(6, 4.13), xytext=(3.5, 4.5),
             fontsize=9, fontweight='bold', color='#9B59B6',
             arrowprops=dict(arrowstyle='-&gt;', color='#9B59B6', lw=2))

# ── Panel D: Confinement Trade-offs ──
ax4 = axes[1, 1]

categories = ['Disruption\nFreedom', 'Steady\nState', 'Compactness', 'Physics\nMaturity', 'Capital\nEfficiency']
concepts_data = {
    'Tokamak':    [1, 2, 2, 5, 2],
    'Stellarator': [5, 5, 2, 3, 3],
    'FRC':         [5, 3, 5, 2, 3],
    'ICF':         [5, 1, 2, 4, 2],
    'Z-Pinch':     [4, 2, 5, 2, 5],
    'Sph. Tok.':   [3, 3, 4, 4, 3],
}
concept_colors = {
    'Tokamak': '#2C5F8A', 'Stellarator': '#2ECC71', 'FRC': '#F39C12',
    'ICF': '#9B59B6', 'Z-Pinch': '#E74C3C', 'Sph. Tok.': '#3498DB'
}

x_cat = np.arange(len(categories))
n_concepts = len(concepts_data)
bar_width = 0.12
offsets = np.linspace(-bar_width * (n_concepts - 1) / 2, bar_width * (n_concepts - 1) / 2, n_concepts)

for i, (name, scores) in enumerate(concepts_data.items()):
    ax4.bar(x_cat + offsets[i], scores, bar_width,
            color=concept_colors[name], alpha=0.8, label=name,
            edgecolor='black', linewidth=0.3)

ax4.set_xticks(x_cat)
ax4.set_xticklabels(categories, fontsize=9)
ax4.set_ylabel('Score (1–5)', fontsize=11)
ax4.set_title('(D) Confinement Concept Trade-offs', fontweight='bold')
ax4.set_ylim(0, 6)
ax4.legend(fontsize=7, loc='upper right', ncol=2)
ax4.grid(axis='y', alpha=0.3)

plt.tight_layout(rect=[0, 0, 1, 0.95])
plt.savefig('vol8_fig2_landscape.png', dpi=200, bbox_inches='tight',
            facecolor='white', edgecolor='none')
plt.close()
print("Figure 2 saved.")
Enter fullscreen mode Exit fullscreen mode

IMAGE_URL_PLACEHOLDER

Every alternative confinement concept must choose a fuel cycle. That choice determines whether the constraints of Volumes 3–5 apply:

Fuel Approaches Using It Vol.3 (Tritium) Vol.4 (Materials) Vol.5 (Neutronics) Key Physics Barrier
D-T Tokamaks, stellarators, spherical tokamaks, ICF, Z-pinch Full constraint Full constraint (14.1 MeV neutrons) Full constraint TBR < 1.0 in 88% of scenarios
D-³He Helion, PFRC Eliminated (but D-D side reactions produce 3–10% neutrons) Reduced (residual neutrons) Reduced 17× higher triple product than D-T (Vol.6)
p-¹¹B TAE Eliminated Eliminated (aneutronic primary) Eliminated Thermal ignition impossible (Vol.2, §4). Beam-driven approach unproven at reactor scale
D-D Some FRC concepts Eliminated Full constraint (2.45 MeV neutrons) Full constraint ~7× harder than D-T

The "aneutronic paradise" that p-¹¹B and D-³He promise is real in terms of materials and tritium — but the physics price is enormous. Volume 2 showed that the bremsstrahlung power loss for a thermal p-¹¹B plasma exceeds the fusion power by a factor of ~23 at optimal temperature. Volume 6 showed that D-³He requires 17× the triple product of D-T.

The paradox remains: the fuels that solve the engineering problems (tritium, materials, activation) are the ones that make the physics hardest. The fuel that makes the physics easiest (D-T) is the one that creates the worst engineering problems.

No alternative confinement concept escapes this paradox. It is a property of nuclear physics, not of machine geometry.

§10. Uncertainties (Honest Section)

What this article does not know:

  1. Stellarator D-T performance is a complete unknown. W7-X has never operated with D-T fuel, and there is no stellarator data on alpha particle confinement, self-heating, or burning plasma behavior. The extrapolation from helium/hydrogen plasmas to D-T is significant.

  2. FRC stability scaling is unproven. The kinetic effects that stabilize current FRC experiments may or may not persist at reactor-relevant parameters. No theoretical framework reliably predicts FRC stability at arbitrary $s$ (ratio of plasma radius to ion gyroradius).

  3. Private company claims are unverified. Neither Helion (Polaris), TAE (Norm's full performance), nor Pacific Fusion (IMG at fusion-relevant scale) have published peer-reviewed data from their latest machines. We are relying on company press releases and conference presentations.

  4. NIF's ignition gains do not translate to IFE. The gain 4.13 is target gain, not facility gain. The path from NIF-class target physics to a 5 Hz, 200+ MWe IFE plant requires several orders of magnitude improvement in driver efficiency, target manufacturing, and chamber survival. None of these has been demonstrated.

  5. Zap Energy's Z-pinch physics may not scale. The sheared-flow stabilization theory is validated at current parameters, but extrapolation to reactor-relevant current (~10 MA), density ($10^{26}$ m⁻³), and confinement time is uncharted territory.

  6. General Fusion's near-death is a warning, not an anomaly. Any private fusion company that fails to demonstrate physics milestones before running out of capital faces the same fate. The 2025 funding cycle was favorable; the next one may not be.

  7. This article has a selection bias. I cover the concepts with the most capital and visibility. Several promising approaches (dense plasma focus, mirror machines, magnetically confined laser-heated plasmas, polyhedral devices) receive no attention here simply due to space constraints.

  8. I am not a physicist. I have no access to unpublished data, no ability to independently verify simulation results, and no way to resolve disagreements between competing theoretical frameworks. This article synthesizes public information. Errors are mine.

§11. Decision Matrix

Technology Selection Matrix

Criterion (weight) Stellarator FRC (TAE) FRC (Helion) ICF (NIF-class) ICF (Pulsed Power) Z-Pinch Spherical Tokamak MTF
Physics Maturity (25%) ★★★☆☆ ★★☆☆☆ ★★☆☆☆ ★★★★★ ★★☆☆☆ ★★☆☆☆ ★★★★☆ ★★☆☆☆
Disruption Risk (15%) ★★★★★ ★★★★★ ★★★★★ N/A N/A ★★★★☆ ★★★☆☆ ★★★★☆
Steady-State (15%) ★★★★★ ★★★★☆ ★☆☆☆☆ ★☆☆☆☆ ★☆☆☆☆ ★★☆☆☆ ★★★☆☆ ★☆☆☆☆
Tritium Avoidance (10%) ★☆☆☆☆ ★★★★★ ★★★★☆ ★☆☆☆☆ ★☆☆☆☆ ★☆☆☆☆ ★☆☆☆☆ ★☆☆☆☆
Compactness (10%) ★★☆☆☆ ★★★★★ ★★★★☆ ★★☆☆☆ ★★☆☆☆ ★★★★★ ★★★★☆ ★★★☆☆
Capital Efficiency (10%) ★★★☆☆ ★★★☆☆ ★★★☆☆ ★☆☆☆☆ ★★★★☆ ★★★★★ ★★★☆☆ ★★☆☆☆
Funding Strength (10%) ★★★★☆ ★★★★★ ★★★★☆ ★★★★★ ★★★★★ ★★★☆☆ ★★★☆☆ ★★☆☆☆
Time to First Electricity (5%) ★★☆☆☆ ★★☆☆☆ ★★★☆☆ ★☆☆☆☆ ★★☆☆☆ ★★☆☆☆ ★★★☆☆ ★☆☆☆☆

Reading the matrix: No column dominates. This is the central message of the article. Each approach trades strengths for weaknesses. A portfolio strategy — funding multiple approaches simultaneously — is the rational response to this uncertainty.

The Convergence Thesis

Despite their differences in confinement geometry, the alternative approaches are converging on shared technologies:

  1. HTS magnets: Used by CFS (tokamak), Proxima (stellarator), Type One (stellarator), Tokamak Energy (spherical tokamak). This is not a coincidence — HTS enables stronger fields in smaller devices regardless of geometry.

  2. Pulsed power: Used by Pacific Fusion (ICF), Zap Energy (Z-pinch), Helion (FRC). Advances in capacitor technology, solid-state switching, and power supply engineering benefit all three.

  3. Liquid metal technology: Used by Zap Energy (first wall), General Fusion (compression), and proposed for multiple other concepts. Flowing liquid metal simultaneously solves first wall protection, tritium breeding, and heat extraction.

  4. AI/ML for plasma control: Used by every major experiment. Real-time feedback control of plasma instabilities is a universal need.

These shared technology platforms mean that progress in one approach spills over to others. The fusion ecosystem is not zero-sum.

§12. Conclusion

The tokamak is not the only path to fusion power. It is the most mature path, but maturity is not destiny.

What the data show:

  • Stellarators have reached tokamak-level triple product at long pulse durations and eliminated the disruption problem entirely. Two companies (Proxima, Type One) have peer-reviewed power plant designs and serious capital.
  • ICF has achieved the only demonstrated fusion ignition in history, with target gain reaching 4.13 in April 2025. Pacific Fusion's pulsed-power approach could make ICF commercially viable if IMG technology performs at scale.
  • FRCs offer the highest β and most compact geometry of any magnetic confinement device. TAE's NBI-only breakthrough and Helion's Microsoft/Nucor PPAs show commercial intent, but peer-reviewed performance data from latest machines is absent.
  • Z-pinch has demonstrated gigapascal pressures and high repetition rates in a 2-meter device with no magnets. If stability scaling holds, this is the cheapest path to fusion. That "if" is large.
  • Spherical tokamaks combine the proven physics of the tokamak with higher β and more compact geometry. Tokamak Energy's ST80-HTS and the UK STEP program are the most concrete efforts.

What the data do not show:

No alternative confinement concept has demonstrated Q > 1 in a magnetic confinement device. Only NIF (ICF) has achieved ignition, and only in terms of target gain. The gap between laboratory demonstrations and a functioning power plant remains enormous for every approach.

The physics constraints of Volumes 1–6 are geometry-independent. Tritium does not know whether it is inside a tokamak or a stellarator. Materials do not care whether the 14.1 MeV neutrons came from a Z-pinch or an ICF capsule. The only escape is advanced fuels — and advanced fuels require physics breakthroughs that have not occurred.

The portfolio argument:

Given the fundamental uncertainty across approaches, a rational fusion strategy funds multiple concepts simultaneously. The global fusion ecosystem is, in fact, doing this: of the $15.2 billion in private capital invested through September 2025, approximately 45% is in tokamaks, 20% in FRCs, 15% in stellarators, 10% in ICF, and the remaining 10% across Z-pinch, MTF, and other concepts.

The question is not "which concept will win?" It is "how many concepts can survive long enough to demonstrate — or fail at — the physics?" General Fusion's 2025 near-death experience shows what happens when the answer is "not enough."

The shape of the bottle matters. But the fire inside obeys the same laws in every bottle.

References

  1. Max Planck Institute for Plasma Physics, "Wendelstein 7-X sets new performance records in fusion research," IPP Press Release, June 3, 2025.
  2. Princeton Plasma Physics Laboratory, "Wendelstein 7-X sets new performance records," PPPL News, 2025.
  3. Proxima Fusion, "Stellaris Fusion Power Plant Concept," Fusion Engineering and Design, February 2025.
  4. Proxima Fusion Press Release, "€130M Series A," June 11, 2025.
  5. Type One Energy, "Initial Design Review of Infinity Two," Press Release, May 27, 2025.
  6. Type One Energy / TVA, "Letter of Intent," September 19, 2025.
  7. TechCrunch, "Bill Gates-backed Type One Energy raises $87M," January 14, 2026.
  8. TAE Technologies, "Fusion Breakthrough: NBI-Only FRC Formation," Nature Communications, April 2025.
  9. TAE Technologies Press Release, "Shortened Device Roadmap," November 17, 2025.
  10. Helion Energy, Polaris prototype commissioning, January 2025.
  11. Kirtley, D., Milroy, R., "Fundamental Scaling of Adiabatic Compression of FRC Thermonuclear Fusion Plasmas," J. Fusion Energy 42, 30 (2023).
  12. Lawrence Livermore National Laboratory, "Achieving Fusion Ignition," NIF & Photon Science, updated 2025.
  13. LLNL, NIF Shot Record: April 7, 2025 — 8.6 MJ yield, target gain 4.13.
  14. Pacific Fusion, "Affordable, manageable, practical, and scalable (AMPS) high-yield and high-gain inertial fusion," arXiv preprint, April 2025.
  15. Pacific Fusion / General Atomics Press Release, "Collaboration on IMG Module Testing," April 24, 2025.
  16. Zap Energy, "FuZE-3 Exceeds Gigapascal Pressures," Press Release, November 18, 2025.
  17. Zap Energy, "Century: 100-kW-Scale Repetitive Sheared-Flow-Stabilized Z-Pinch System," Fusion Science and Technology, 2025.
  18. Tokamak Energy, "ST80-HTS Advanced Prototype," Press Release, October 25, 2022.
  19. UK STEP Programme, UK Atomic Energy Authority, 2024.
  20. Fusion Industry Association, "The Global Fusion Industry in 2025," July 2025.
  21. IEEE Spectrum, "Stellarator Showdown: Proxima Fusion vs. Type One Energy," May 2025.
  22. ITER Organization, schedule updates through 2025.

Nuclear Fusion from Scratch — Vol.8 of 10
Next: Vol.9 — Fusion Propulsion: From Reactor to Rocket
The shape of the bottle matters. But the fire inside obeys the same laws in every bottle.

vol8_fig1_physics.png
vol8_fig2_landscape.png

Top comments (0)