---
title: 'This $15 Billion Nuclear Fusion Machine Can’t Work, Physicists Say'
source: 'https://youtube.com/watch?v=wwB4hB9QJX8'
video_id: 'wwB4hB9QJX8'
date: 2026-11-11
duration_sec: 424
channel: 'Sabine Hossenfelder'
---

# This $15 Billion Nuclear Fusion Machine Can’t Work, Physicists Say

> Source: [This $15 Billion Nuclear Fusion Machine Can’t Work, Physicists Say](https://youtube.com/watch?v=wwB4hB9QJX8)

## Summary

The video analyzes recent setbacks in the nuclear fusion industry, focusing on Helion Energy's $15 billion valuation despite mounting scientific criticism. It compares different fusion approaches and assesses their likelihood of success, concluding that net power before 2030 is unlikely.

### Key Points

- **Fusion startups face delays** [00:00] — Nuclear fusion startups are postponing completion dates, and it's becoming clear which approaches are succeeding and which are not.
- **Record private investment** [00:36] — Private investment in fusion has exploded, with $4.48 billion raised in the past 12 months and over $14 billion since 2021.
- **Circular investment patterns** [01:06] — Investors like Google, Microsoft, and Nucor are both funding fusion startups and signing contracts to buy future electricity, creating potential conflicts of interest.
- **Helion's valuation triples** [01:44] — Despite criticism, Helion Energy is now valued at $15 billion, three times more than in January last year.
- **Helion's approach explained** [01:56] — Helion uses field-reversed configuration, accelerating two plasma rings to collide and compress, aiming for direct electricity extraction.
- **First criticism from John Slough** [02:50] — John Slough, whose work led to Helion's machine, said it will never work due to a fundamental physics aspect.
- **Max Planck Institute critique** [03:15] — A group from Max Planck found that energy redistributes too quickly after collision, preventing net energy gain.
- **French group's energy loss finding** [04:14] — A French group found only 1% of magnetic field energy goes into the plasma, concluding net energy is never achieved even at 50 Tesla.
- **Helion's response and delay** [04:41] — Helion dismisses critics as looking at old designs and silently moved net power plans from 2028 to 2029/2030.
- **Traditional approaches progress** [05:06] — Tokamaks and stellarators (like Open Star's levitating dipole) are making steady incremental progress, while inertial confinement (First Light) has stalled.
- **Net power unlikely before 2030** [05:35] — It seems very unlikely anyone will produce net power from fusion before the end of the decade.

### Conclusion

The video concludes that understood physics is slow but reliable, while poorly understood physics is fast until it fails. It advises caution in investing in high-risk fusion startups like Helion.

## Transcript

Nuclear fusion startups are postponing and&nbsp; postponing their completion dates. This&nbsp;&nbsp; unfortunately isn't surprising. More interesting&nbsp; is that it's becoming increasingly clear who's&nbsp;&nbsp; doing well and who isn't. In particular, the&nbsp; outlier approach helium fusion has encountered&nbsp;&nbsp;
some problems recently. Today, I want to look&nbsp; at what happened with Helon, but also give you&nbsp;&nbsp; an assessment of how I think it's going with&nbsp; the different approaches. Make sure you stick&nbsp;&nbsp; around till the end. I've got a great offer from&nbsp; today's sponsor, Outskill. In the past decade,&nbsp;&nbsp;
private investment in nuclear fusion technology&nbsp; has exploded. Fusion companies raised an record&nbsp;&nbsp; $4.48 billion in the past 12 months. The total&nbsp; investment since 2021 now sum up to more than&nbsp;&nbsp;
$14 billion. These investments show some of the&nbsp; same circularity that we're seeing for AI. The&nbsp;&nbsp; companies giving money to nuclear fusion startups&nbsp; are also signing contracts to buy the electricity&nbsp;&nbsp;
that may or may not one day come in. Google, for&nbsp; example, has invested into Commonwealth Fusion and&nbsp;&nbsp; also said they'd buy back from them. Microsoft&nbsp; has invested into openai and openmeai CEO Sam&nbsp;&nbsp;
Alman has invested into Helium's fusion machine&nbsp; which Microsoft said they want to buy from. The&nbsp;&nbsp; American steel manufacturer new car invested&nbsp; into Helion and also said they'll buy back&nbsp;&nbsp;
electricity if that electricity ever comes into&nbsp; being. There's nothing legally wrong with this,&nbsp;&nbsp; but it's certainly part of the reason why Helium&nbsp; was recently valued at a staggering 15 billion,&nbsp;&nbsp;
three times as much as January last year.&nbsp; That's despite the harsh criticism they've&nbsp;&nbsp; received earlier this year. As a reminder, Helon&nbsp; Energy works at a particularly innovative approach&nbsp;&nbsp;
to nuclear fusion called a field reversed&nbsp; configuration. They want to accelerate two&nbsp;&nbsp; blobs of fuel particles with magnets. The fuel&nbsp; forms rings which then collide and get compressed&nbsp;&nbsp;
in the middle. The compression increases pressure&nbsp; and temperature and fuses the nuclei. The magnetic&nbsp;&nbsp; field inside the few rings is reversed, hence&nbsp; the name. One of the appeals of Helen's approach&nbsp;&nbsp;
is that they want to extract electric energy&nbsp; directly rather than using the fusion to create&nbsp;&nbsp; heat and then extract energy from that. Now, when&nbsp; I first heard about this, I thought it's basically&nbsp;&nbsp;
a particle collider. There's no way this thing&nbsp; will break even faster than a celerator. But okay,&nbsp;&nbsp; not my research area. What's happened now though&nbsp; is that three groups independently have come&nbsp;&nbsp; forward and said this isn't going to work.&nbsp; First, there is one of the researchers whose&nbsp;&nbsp;
work led directly to the helion machine, John&nbsp; Slow. He said in an interview with Scientific&nbsp;&nbsp; American that helon's machine would never work.&nbsp; They have quote run up against a fundamental&nbsp;&nbsp;
aspect and that he can't see anything in the&nbsp; physics. Slow has proposed a different approach&nbsp;&nbsp; in which there's only one plasma ring that gets&nbsp; gradually squeezed until fusion sets in. Second,&nbsp;&nbsp;
a group from the Max Plunk Institute for Plasma&nbsp; Physics have looked at Helon's proposed mechanism&nbsp;&nbsp; and they say that it just doesn't work. You see,&nbsp; Helon says that they have a good chance to produce&nbsp;&nbsp;
net energy because when the two pulses collide,&nbsp; most of the kinetic energy goes into the heavy&nbsp;&nbsp; particles that are the nuclei, not the electrons.&nbsp; So, you have the energy where you want it. But&nbsp;&nbsp;
the Maxplank group says that while this might&nbsp; be the case immediately after the collision,&nbsp;&nbsp; the energy gets redistributed extremely quickly,&nbsp; meaning helium will never get anywhere close to&nbsp;&nbsp; nuclear ignition. They write, quote, only&nbsp; for very small values of gain which are of&nbsp;&nbsp;
no practical interest for power production with&nbsp; the initial temperature differences result in the&nbsp;&nbsp; improvement predicted by helium. [clears throat]&nbsp; Then third, we have a paper from a group in France&nbsp;&nbsp; which has found that only a small fraction of the&nbsp; magnetic field energy in the range of 1% actually&nbsp;&nbsp;
goes into the plasma which is a huge energy&nbsp; loss. They conclude for the realistic case&nbsp;&nbsp; net energy is never achieved even at a magnetic&nbsp; field strength of 50 Tesla. Helon has basically&nbsp;&nbsp;
addressed this criticism by saying yes, yes,&nbsp; but the critics are looking at an old design&nbsp;&nbsp; plan. We now know how to fix the problem, but we&nbsp; won't tell you how. Meanwhile, they have silently&nbsp;&nbsp;
moved their plan for net power from 2028 to maybe&nbsp; 2029 or 2030. As you probably guessed, personally,&nbsp;&nbsp; I wouldn't invest into helium. We've also seen&nbsp; in the past year that inertial confinement fusion&nbsp;&nbsp;
like it was pursued by first light fusion hasn't&nbsp; been going very well. Basically their progress&nbsp;&nbsp; stored. The approaches that have made steady if&nbsp; incremental progress are the more traditional&nbsp;&nbsp;
ones. Tokma accelerators and I'll count open&nbsp; stars levitating dipole among those. There are&nbsp;&nbsp; also a few hybrid approaches still out there such&nbsp; as the Zet pinch from Z Energy or Pacific Fusion,&nbsp;&nbsp;
but the pulsed version of magnetic confinement.&nbsp; The physics of the hybrid approaches is generally&nbsp;&nbsp; not as well understood. So, the risk that they&nbsp; fail is much higher. On the other hand, if it&nbsp;&nbsp; works, they might get the job done faster. Either&nbsp; way, though, at this point, it seems very unlikely&nbsp;&nbsp;
we'll see anyone producing net power from nuclear&nbsp; fusion before the end of the decade. In summary,&nbsp;&nbsp; understood physics is slow, poorly understood&nbsp; physics is fast, right up until you understand&nbsp;&nbsp;
it. Artificial intelligence is developing rapidly&nbsp; and open eyes GPT6 Astra's closer to AGI than ever.&nbsp;&nbsp; Are you keeping up? Are you sure? Today's sponsor,&nbsp; Outskill will bring you up to speed. Outskill is&nbsp;&nbsp;
offering a GPT6 Astra crash course. It's 3 hours&nbsp; long and it'll teach you how to build your own&nbsp;&nbsp; Astra workflow hands on. Outskill is an official&nbsp; OpenAI partner, so their team has exclusive access&nbsp;&nbsp;
before anyone else. They've tested it upside&nbsp; down and had time to figure out how to get the&nbsp;&nbsp; most out of it. In the workshop, you'll learn, for&nbsp; example, how you can use Astra to find jobs that&nbsp;&nbsp; are tailored to your skills or how to alert you&nbsp; of information concerning your stocks. When you&nbsp;&nbsp;
sign up, you'll also get the Astra Power Playbook&nbsp; with prompt formulas and workflows ready for you&nbsp;&nbsp; to plug and play. The link to sign ups in the&nbsp; description, or you can scan the QR code on&nbsp;&nbsp;
screen. I've signed up for it because it's an easy&nbsp; way to stay up to date. So, go ahead, sign up for&nbsp;&nbsp; this and learn what less than 1% on this planet&nbsp; know. Thanks for watching. See you tomorrow.
