Helion
blog.samaltman.com
blog.samaltman.com
While it's possible that helion has made improvements to ion density and confinement allowing them to achieve a significantly higher triple product and close the gap to power production, I see no reason why a company looking for investment would hide such a result, especially while putting out press releases celebrating other milestones. I doubt they're anywhere near the point where an economical plant could even be considered, though I'd love to be proven wrong.
E.g.: In the Tech Crunch article linked elsewhere in this discussion there is this quote:
"Helion’s CEO speculates that its first customers may turn out to be data centers"
Do you know what else a 50MW generator could be used for? Anything. Anything that electricity is used for now. Why talk about things we all already know? Why talk about specific applications?
It's like a car company advertising that their new engine could be used to drive to Starbucks to get a coffee.
Once you notice this pattern, you'll see it everywhere in Free Energy / LENR circles...
As you get better at mass producing the machines, the cost will come down and you can make money selling power to the grid at much lower prices.
Untested nuclear reactors (fusion or fission) are about as far away as possible from what they want as it is possible to get.
Not to mention that many data centres are bang in the middle of high density urban areas, or at least reasonably close to them, typically in some industrial area on the outskirts of town.
No council in their right mind would approve any kind of nuclear power without years and years of environmental impact studies, justifications, reams of paperwork, etc...
You might argue that it's "safe", but the bureaucrats won't care about your opinion. They'll be worried about perception in an era where people set communication towers on fire because they want to stop 5G "radiation".
Fusion is precisely the kind of technology that is best implemented as base load in a standard power plant type configuration. Far away from big cities and scaled for efficiency.
It makes zero sense to plop something this esoteric down somewhere downtown to occasionally power a data centre during a rare power outage.
This "suggested" use case is 100% intended to appeal to people like you, the "YC News" crowd type. It is absurd on its face. Hilariously improbable. But it sounds cool and it got you talking, so a successful marketing trick, I suppose.
They also have grid connections, of course, so it's no problem if the energy source has frequent downtime.
Some data centers are in cities but there are many remote ones. https://en.wikipedia.org/wiki/Prineville%2C_Oregon is a small town with large Apple and Facebook centers.
Fusion reactors wouldn't be used just during outages, but for base load, reducing consumption from the grid. They already have this sort of power supply/demand matching to support their solar arrays.
Appealing to cost-insensitive early adopters is an excellent way to start. The Tesla roadster is a well-known example. I don't see the absurdity.
But it is not a risky investment – quite the opposite actually. You can buy panels and inverters with 25yr manufacturer warranties. Maintenance costs are low and predictable (cleaning panels, replacing equipment). Energy output is as predictable as the sun, and in certain regions it is pretty constant.
It is also not "costly" in the sense that the bulk of the cost of installing solar is the labor.
I agree with all the other stuff you said, and I would accept this as a fair characterization of what is in store for nuclear plants historically. I wonder though how much of that experience would transfer in the event that we have real fusion reactions, given climate considerations and a novel technology that won't necessarily become entangled in the same narratives, and (maybe?) comes with a different set of environmental implications than traditional nuclear power plants.
Compared to fission, it's much better overall, but you're still talking about remote manipulation, robots, lead shielding, etc...
There is just no way that a shipping container-sized fusion reactor will be allowed to be "plopped down" anywhere without a metric ton of paperwork and justification.
It wouldn't be safe to go anywhere near it while it is operating! It couldn't possibly have sufficient shielding in that form factor. Even if it had solid lead walls it would still be dangerous.
Fusion would be perfectly fine as base load in dedicated power plants similar to nuclear power plants. There would be thick concrete shielding, containment buildings, etc... Less than you'd need for fission, and also less waste, but there's still nuclear waste that needs to be handled in much the same way.
It's not like you could plop a neutronic fuel as a substitute in a design that expects aneutronic fuel.
The reactor walls will still become "hot", you still need shielding, remote manipulation, etc...
It just that it takes longer for the reactor walls to reach the same level of radioactivity.
If after 1 year of operation the nuclear waste is 50% as radioactive as with a different design, that's nice and all, but it's still... nuclear waste.
"They say the combined reaction will produce only 6% of its energy as neutron radiation, compared to 80% for D-T."
Neutrons are a design requirement for typical tritium breeding fusion reactors, but highly undesirable for Helion's D-He3 D-D reactor. I'd expect a 100x neutron flux reduction at activating energies over typical D-T.
The relative abundance of He3 is, numerically, really quite small (WP says 0.000137%, or 1 He3 per 730k He4 atoms), but that doesn't matter as much as you might think: they don't need much. Process a ton of helium, get 1.03g of He3. But it also says it is 70 to 242 parts per billion, which is a lot smaller than the other, 1370 ppb figure.
It may be cheaper to get it from used-up tritium, from people who are finished with it because it has decayed too much. In fact the US DOE does sell He3 they have extracted from tired-out stocks held ready to inject into bombs before they are sent out to use. The DOE makes its (fresh) tritium by irradiating lithium, but it starts decaying immediately, with a half-life of ~12 years, and the bombs want it fairly fresh.
These FRC reactors generate their own tritium, which is a problem, because when those fuse you get hot neutrons you don't want, and gamma rays. When you use FRC for propulsion, you can expel the tritium as reaction mass, but that doesn't work so well on the ground. On the other hand, lots of shielding is cheap on (under) the ground. But they don't make enough of it to use, and anyway who wants to bank it for years while it decays?
They are rightly taking every opportunity to clarify to investors who their potential market would be.
It has to be someone without vested interests in coal supply contracts and therefore the delayed success of your product, with a huge amount of money to throw at energy security, at a large enough scale for it to be worth a big start up cost. You also need someone to go first, because fusion is scary. This is non obvious. It is an essential part of their pitch, and no amount of cringe from people who know what electricity is is worth omitting it.
It goes down wires and is distributed nationally!
I'm also not in the personal market for: Nuclear power, offshore wind, or gas turbines.
Yet, I get electricity from all of those sources.
If they can make one 50 MW power plant, then they can make ten 50 MW power plants. Put a nice little array of them on some cheap industrial land, hook them up to the grid, and start selling 500 MW like any other power plant. Easy. You can also get funding like any other power plant. Just turn up at a bank. Or issue shares. Whatever. If it works there's no need for specialised applications. It just needs to work!
There is no need to "sell" their investors on the concept of electricity generation and usage. We get it. We all get it, in the most literal sense, right now. No need to talk us into it.
They should be selling me on their capability of producing the thing in the first place, not its utility.
That's much harder if they're faking it, which is why they talk about its utility instead.
You know... if it works.
If.
> Helion’s CEO speculates that its first customers may turn out to be data centers, which have a couple of advantages over other potential customers. Data centers are power-hungry, and often already have power infrastructure in place in order to be able to accept backup generators. In addition, they tend to be a little away from population centers.
They are definitely talking about selling them physical reactors.
For example, solar panels got enthusiastic use in very remote areas even when they theoretically were more expensive than a grid connection. Because there was no grid in remote areas, and no population to support one.
Ironically, their efficient method of energy capture actually makes their job harder than for a thermal system where heat losses due to bremsstrahlung and neutron heating are partially recovered; indeed this is the output for a conventional fusion reactor. For helion, only the energy of the plasma is harvested and thus they must exceed breakeven by enough to not just maintain but to heat the plasma by some economically useful amount.
For short duration fusion (ICF, Magnetized target, Helion, etc) you can exceed the lawson criterion and thus produce gain, but the plasma may still not be long enough lived for the fusion to actually induce more fusion (which is the actual ignition point).
What exactly would they gain from keeping this particular card close to their chest?
Even if Helion is behind the tokamaks, perhaps this play is more about reaching an economically viable reactor design? Not first to fusion, but first to scalable fusion?
The triple product is 16x higher for D-He3.[0] They also need to get D-D reactions going to produce the He3. The triple product for that is 30x the D-T value. (Yeah, you could run the D-D reaction at a loss or at barely-breaking-even I guess).
They're talking about letting the T from the D-D reaction decay to produce more He3. Tritium has a half-life of 12 years, so in steady-state, there is about 20x the annual tritium production sitting in storage. That's a massive amount -- a 1GW D-T reactor would use something like 50kg of T in a year, so that's about 1 tonne of T. (Just getting some rough approximation.) Even if you scale it down to 50 MW, that's still 50kg. It's a major radioactive hazard.
>I see no reason why a company looking for investment would hide such a result
Exactly. Fusion companies tend to trumpet their successes from the rooftop.
[0] https://en.wikipedia.org/wiki/Nuclear_fusion#Neutronicity,_c...
8 crater
1 breaks even
1 does a moonshot
Given the amount of public dollars already put into this without success and the amount of money they are going to have too continually pour into this to make it successful it seems like a serious hail mary. Even if they do have the brightest minds working on it. I wish them the greatest success - we need this.
To your point it's an incredibly privileged investing position to be in and to be honest - he can take a lot of the gains he has already had in relatively uninteresting companies that have been successful and hope to something truly remarkable for humanity.
But an incredibly high return if successful. Nuclear fission (edit... accidentally wrote fusion here), if we can figure it out, is potentially the golden ticket to reducing our carbon footprint. Unlike geothermal energy, it can be done anywhere. Unlike wind or solar, it can be done at any time. It doesn't have the safety issues associated with fusion, nor does it generate waste products nearly as hard to deal with.
Right now, carbon emissions breakdown in the US are broken down by:
Transportation - 29% Electricity production - 25% Industry - 23% Commercial and Residential - 13% Agriculture - 10% Land use and forestry - 12%
By moving to fusion, you can all but eliminate fossil fuel usage in the first two (and largest) categories. You can knock a large chunk out of the next two categories, where much of the emissions is due to burning fossil fuels for energy (heating, etc.). You'll still have emissions from agriculture and land use, but you can clamp down on most emissions in a big way.
If you can figure out fusion and get it working on an industrial scale level on par with other forms of electricity production (which is a big if), then you'll have achieved a monumental technological leap and you'll make a lot of money while at it.
They need to prove that the research works to actually produce net electricity - which requires a scientific breakthrough. Next after a research breakthrough - they need to make this a product -- then a commercial product. During that process they need to make this a commercially viable economically viable product that can compete against other forms of energy in the marketplace. They will need to get through serious regulatory requirements. And remember that they need to make this commercially viable to produce electricity at a very low cost - its super competitive at baseload power cost range.
By the time this comes to market the energy landscape will be completely different. It is already moving incredibly quickly.
Like I've said on other post - we need these kinds of moonshots but let's not have them distract against the other important work of deploying already commercially ready technology into the market.
The UK recently announced a regulatory regime for fusion, with significantly lighter requirements than fission since safety and proliferation issues are much less troublesome. That would be even more the case for aneutronic fusion. Possibly the US would be silly enough to get in the way but many other countries certainly wouldn't, including China.
As an American this is accurate and depressing.
They’re in a really good position here because they actually don’t. Being a no-carbon power source puts them in an almost new market. The government can (should) regulate carbon fuel away, and pour money into this in a non-market way to tip scales. Energy is heavily regulated but also heavily government funded.
To be fair to your comment though air pollution relating to climate warming has been treated as a tragedy of the commons problem for ever.
Fusion is what they are doing here. Fission we have already figured out and have been generating power from for decades.
Fusion does not have the runaway reaction safety issues that fission has.
Also, what you're describing is what we've known since we fairly easily harnessed fusion to make a hydrogen bomb.
Controlling the reaction rate so it doesn't explode is metastable with fission, and nearly impossible with fusion. This solution is basically just using tiny explosions.
[1]: Nuclear is one of the safest kinds of energy we have even including every absurd disaster. We already know how to deal with the waste, and the unit cost of managing nuclear waste is very low (the up-front costs are high, but we're already committed to those costs).
What do you mean "with perfect reliability"? Do you mean it can buy the panels and turbines and batteries? How much reliable energy capacity does $30B buy?
> Displacing a megawatt of carbon-generated power now is worth a lot more than displacing that same megawatt in 10 years, because 10 megawatt-years worth of carbon did not, thereby, go into the atmosphere.
Yes, there are different return-on-investment curves and the short term obviously favors things which can be deployed quickly. The question is which pays off the best for the relevant timescales. Also, we should invest in both--now isn't the time to pinch pennies nor to put all of our eggs in one basket.
Why d'you think? Helion has been super successful with their demos so far. The timelines are optimistic but I don't see why you'd expect the technology itself to fail with, say, >80% probability.
1. Impossible to get anywhere close to good investing rounds. And deal flow requires serious capital on any meaningful technology (sorry carbon accounting software doesn't move the needle, needed but it isn't a game changer).
2. Investing in companies in the market as an equity holder. It feels like it doens't actually help the company - there's an argument that it helps the industry as there is more money/attention/talent attraction. Seems like a poor investment for myself given the P/E ratios on most of the companies.
3. Investing in actual projects - small returns but meaningful results. You don't get the outsized returns on companies growing quickly.
4. I do believe the success of humanity in the climate tech space is actually not through moon shots but a constant deployment of ready tech (read solar, ESS, wind, etc) and getting our politicians to probably signal the value proposition that climate tech brings. I do think moon shots have a place and we should bet on them.
I am open to ideas on how to help and new models if anyone has any!
edit for formatting
I think the place where software could be really useful is with demand response and grid management. There's tons of work in that space already.
Seems to me that long-duration storage and electrochemical production of fuels/chemicals/materials are the places where new technology is really needed, thinking ahead to when we get to 100% decarbonization of the grid and beyond that to when we try to decarbonize everything else. If the exponential growth and learning rates of solar keep up, in 15-20 years it will meet our entire projected primary energy demand and cost 5x less than the going rate for electricity right now. (Those are big 'if's, though!)
Alan Kay[0]: "The key to the Parc approach was to be able to do many experiments in the future without having to optimize." What technology are we going to need 15-20 years from now, when solar energy is 5x cheaper, that doesn't exist yet? My thinking is that electrochemistry is the big missing puzzle piece. If we totally dropped fossil fuels, we'd need to pump something like the entire present-day electric grid's worth of power into making chemicals/fuels.
Some companies with new technologies in this space that I like:
-Form Energy: low-cost iron-air batteries
-Prometheus: solar fuels from CO2
-Boston Metal: zero-carbon steelmaking
-Twelve (formerly Opus 12): other chemicals from CO2
-Carbon Engineering: air capture of CO2
[0]https://news.ycombinator.com/item?id=11955020
Edit: formatting
The bottling industry similarly ran campaigns in the 70s to convince people that litter pollution was a personal responsibility problem, so that it wouldn't have to pay to clean up its mess.
Similarly, rather than making safer cigarettes, the cigarette industry ran commercials and hired "experts" to testify that the cause of household fires was flammable furniture (not cigarettes). As a consequence, several generations grew up around toxic flame retardants.
Ultimately personal responsibility cannot carry the day. Not only is it politically impossible to convince everyone to give up their luxuries and frivolities, but even if we could, these things account for a small share of our pollution. We need to transition our economy to clean energy. Carbon tax (or "pricing" if you chafe at the word "tax") is necessary (but probably not sufficient).
Yes, this will probably "harm the economy" in the same way that limiting one's credit card debt "harms their personal finances".
If electricity gets just a little cheaper (and it's fairly obvious that it will), then Power-to-Gas technology becomes viable, and could displace fossil fuels quite rapidly.
I'm looking into it. Email in profile if you're interested.
When we get to the point of putting Graphene batteries in planes than can fully recharge in the time it takes to unload and reload passengers/luggage it’s going to be pretty incredible.
Once the hydrogen-powered aircraft start flying, kerosene-powered craft will find it impossible to compete.
Main thing I wanted to say is that we all love to shit on how Silicon Valley has basically gotten rich off investing in websites and SaaS products over the past ~15 years, areas which the Internet has provided a natural monopoly to the winner but haven't really been the type of "societal innovation" we've been craving. This, however, is obviously different, and if it works (a huge if), would be on par with the transistor in terms of societal effects. Kudos to Sam for swinging big.
The quotes about total system efficiency is also odd. "95% efficient"? They're not planning on capturing heat energy, so neutron heating is totally wasted, and they're talking about using entirely resistive 12 tesla magnets, which will also throw off a lot of heat.
- Fusion reactors are an order of magnitude physically larger than fission reactors,
- The particle energies are an order of magnitude higher than fission, resulting in much nastier activation
and that results in orders of magnitude more highly radioactive waste.
>orders of magnitude more
Really? Wikipedia says "a typical large 1000 MWe nuclear reactor produces 25–30 tons of spent fuel per year". I don't see a commercial fusion reactor being quite that bad, unless reactor lifetime ends up being very short.
Helion's promotional images keep showing the reactor vessel in a shipping container, which is narrowly true, but doesn't show the support equipment on the other side of a five meter thick concrete radiation shield. No vacuum pump near a running fusion reactor will survive long.
There are all sorts of fascinating engineering problems for a commercial reactor. Photos of the Wendelstein 7-X show it covered with ports for sensors, which you couldn't do with a power reactor, because anything you attach directly to the vacuum vessel will be destroyed. It'll probably end up terribly ugly, the usual intestinal tangle of any industrial plant, a big spiky sausage with pressure sensors at the ends of long tubes to reduce their neutron flux. If you have to mount any sensor directly to the vessel then you use ten or twenty fold redundancy, since replacement is impossible.
After five years of operation, if you go to replace a sensor on the vessel and the threaded fitting crumbles away when you apply the wrench, what do you do? Weld on another one? What's the weld heat-affected zone look like inside the steel after it's spent five years soaked in hot hydrogen, helium, and is richly marbled with transmutation products, dozens of odd elements you don't ordinarily choose to alloy steel with?
Reactor size depends on the design. For tokamaks, output scales well with reactor volume. But it also scales with the fourth power of magnetic field strength. Two well-funded startups are building tokamaks with newer superconductors that support stronger magnetic fields, allowing them to get the same output as ITER from much smaller reactors.
Other designs have different scaling laws so aren't necessarily any particular size. Helion for example is pretty compact.
500 years ago, these people would've chased Jewish financiers out of strongly catholic areas, or shit on Dutch merchants getting rich in the spice trade.
When it comes to human nature, things don't really change.
Also I think you mean "Catholic", not "catholic".
Wow, I sometimes like to think that I speak English at a native level and then I learn about something like this..
Catholic: relating to the Catholic church catholic: including many different things
She is a novelist who is catholic in her interests.
Criticism - "Retired Princeton Plasma Physics Laboratory researcher Dr. Daniel Jassby mentioned Helion Energy in a letter included in the American Physical Society newsletter Physics & Society (April, 2019) as being among fusion start-ups allegedly practicing "voodoo fusion" rather than legitimate science. He noted that the company is one of several that has continually claimed "power in 5 to 10 years, but almost all have apparently never produced a single D-D fusion reaction".[24] However, the Helion team published peer-reviewed research into its colliding FRC system demonstrating D-D neutron production as early as 2011,[11] and further detailed D-D fusion experiments producing neutrons in an October 2018 report at the United States Department of Energy's ARPA-E's annual ALPHA program meeting.[25] According to the independent JASON review team,[26] VENTI, a sub-scale prototype Helion had developed partially for the ALPHA program, achieved initial results of 8·1022 ions/m3, 4·10-5 seconds confinement time and a temperature of 2 keV for a triple product of 6.4·1018keV·s/m3 in 2018."
https://old.reddit.com/r/fusion/comments/qkvzjs/fusion_energ...
I wish there was a good summary article of all the mainstream and alternative fusion approaches out there, like ITER, (SP-)ARC, General Fusion, Wendelstein 7X, etc.
I'm not the best one to explain what it is Helion is doing, but it's not p+B11 -> 4He.
In a sense, them patenting it so early is sort of a public good =)
And within those there are the variety of designs.
I hate to say it, but a tiny part of me doesn't discount the possibility of energy majors subtley killing small-scale fusion approaches and pushing internation-scale huge projects (i.e. ones unlikely to be delivered quickly).
Pretty much, but there are some that are kinda-sorta both (General Fusion's approach is "magneto-inertial"), and also a lot of variety within each bucket. Like, Zap Energy's approach is magnetic, but involves no external magnets, whereas developing the fancy magnets is a key operational challenge for CFS.
[1] https://astronomy.stackexchange.com/questions/328/is-there-a...
After all, I'd be insanity to consume massive amounts of energy without accounting for its future costs.
The general sense that I get (as a purely amateur observer but one who reads a fair bit about the various efforts) is that pretty much everyone thinks ITER will reach a decently high Q, but that it will take forever and be incredibly expensive, and that beyond that, of the people that think any of the startups have a chance (which seems to be a minority among fusion people but a decently big one), there's the most consensus around the potential of the high-field tokamak startups (so that's CFS/(SP)ARC and Tokamak Energy), because the physics is basically the same as with ITER, except facilitated by the much higher-strength fields allowed for by high-temperature-superconductor magnets; there are engineering challenges there still, but it seems like if the physics underlying ITER are sound and the magnets work (which has at this point been demonstrated as a stand-alone thing), the math all adds up. Stellarators (Wendelstein etc.) seem like the next-highest consensus: people seem to think the physics is sound, but it's less far along. Beyond that, other things seem to fall into one of: "the physics basis seems fine but the economics seem questionable" (probably most inertial approaches), "this has been considered at length and many believe this is physically impossible, but it'd be amazing if they're wrong" (TAE, arguably General Fusion, possibly also Helion), or "it'd be incredible if it works and physics doesn't forbid it but it's wildly novel and nobody really knows yet if it has legs" (things like Zap Energy's shear-stabilized Z-pinch plan).
The overarching tldr is that it seems like the approaches that the most people think will work also have some of the most significant operational challenges if they do (breeding tritium, dealing with high neutron flux, etc.), which is why anybody is bothering with the quirkier approaches: if they work, they will probably ultimately work better than the tokamaks do, but they might not work.
I'm curious though, what happens to the plasma before the next pulse - in the animations it neatly dissipates, but I doubt it's that simple.
The big issue with other fusion reactor designs is how to make sure this incredibly hot plasma doesn't touch and melt any part of the reactor. Current designs try to "levitate" and contain the plasma with a magnetic field, but it's really, really difficult to successfully contain something as energetic and chaotic as plasma. I'm guessing part of the point of the pulses here is to answer the problem of "we can't contain it for long" with "we don't have to".
TAE is exploring static FRC which has instabilities over longer timescales. Helion uses a pulsed approach which means they don't need to worry about these long-term stabilities and can simply optimize for peak power in non-equilibrium systems.
Over the years, many fusion designs have been shut down due to losses at equilibrium. It seems that Helion avoids these factors altogether by having a non-equilibrium system.
Happy to see investment in this space, but also tempering my expectations that having a commercial fusion reactor in three years is anything approaching realistic.
[1] https://techcrunch.com/2014/08/14/y-combinator-and-mithril-i...
[0] https://www.reddit.com/r/fusion/comments/qna3dr/-/hjh36wx
How close is nuclear fusion power? https://www.youtube.com/watch?v=LJ4W1g-6JiY
In this environment it's hard not to see all of this as snake oil.
Helion in particular was founded by respectable scientists who have a background in plasma physics, and they've developed their idea pretty quietly for the last 10 years, only de-stealthing this summer (presumably after reaching a key milestone). What they're trying to do is difficult, but I wouldn't say it's snake oil.
I think half of that is that most other engineering professions now come with some non negligible coding skill and the other half is simply that "software can solve anything" is a plain and simple lie.
I can't help but feel a little left out of innovation with "just" software skills. Am I too sorry for myself or is that a shared feeling?
There's a nuclear technology company that open-sourced a tool for measuring fuel efficiency. https://github.com/terrapower/armi
But it's by no means the most important tool. There's a lot of hard, physical engineering problems that can be solved.
I think of software engineering like motor oil. You can apply it across a wide variety of contexts, but it's by no means the engine.
I think this is among the best things someone like Altman can do with his money and I think it's awesome - even if Helion turns out to be Theranos-level nonsense in hindsight. He seems perfectly willing to lose his entire investment, likely an enormous sum of money. I think it's impressive and I hope it becomes some sort of hype among the ultra-wealthy.
In general, there's so many ideas that "don't seem like they wouldn't work" without evidence one way or the other. It wouldn't surprise me that the path with FRCs is more fruitful than tokamaks/stellarators/lasers. FRCs are rooted in the inertial-electrostatic-confinement realm of fusion research which actually does produce neutrons even on the smallest of scales.
So as I said I really hope this works and makes the founders rich while also producing cheap & clean energy, but I remain at least a bit skeptical.
I take this with a big grain of salt too but 2.2 billion lined up is an awful lot of funding. To your point, if they are legit, I'm sure they are building off of prior fusion research and progress.
As some one else mentions, part of Helion's novelty is that they don't use heat to produce the electricity (through steam and traditional generators). They use the Faraday effect on the (pulsed) magnets.
This is (I think) unique to their approach. Therefore, if they actually have "built a generator that produces electricity", it may prove that part of their concept.
Then all they need to do is get the fusion part working for longer than 1ms.
Skipping the heat cycle is possible if they achieve their goal of using aneutronic fuel. With D-T fusion (the easiest) the output energy is 80% neutrons, so you're stuck with heat. With aneutronic, you mostly get fast-moving charged particles.
IIRC YC's investment is 7%, correct? I'm not sure how much they participate in subsequent rounds, but Stripe alone is now valued at about $95 billion.
I'd be shocked if Sam weren't that loaded.
I recently saw an 8-figure round in which the lead put in $25K. But they did the legwork to arrange the round and get everyone else in, so they were the lead. Sam's the board chair, an early investor, and super connected, so he's certainly able to lead a round whether or not he puts in the majority of the cash.
Sabine Hossenfelders explains here:
> Helion has a clear path to net electricity by 2024
...which strongly implies producing more electricity than put in, not just heat.
https://www.cnbc.com/2021/11/05/sam-altman-puts-375-million-...
So with a regular gas generator you go to the gas station and buy some 93 unleaded gas, then pour it into your generator. You might then pull something or use an electric start to turn on the generator and then you have electricity.
With this power plant the site says it requires helium. Where does this helium come from? Presumably energy is needed to get the helium just like energy is needed to pump natural gas and frack for oil.
Will the total output in electricity be greater than the inputs for this?
Extra: what happens if there’s too much expansion or things are too hot? Could this explode or implode? If so would it be a huge deal?
Cool tech!
> Will the total output in electricity be greater than the inputs for this?
Yup, that's the goal!
> Extra: what happens if there’s too much expansion or things are too hot?
Fusion reactions are difficult enough that if it were physically possible to make them run 'hotter' or release energy faster, frankly we would already be doing so.
> Could this explode or implode? If so would it be a huge deal?
The radiation risks from fusion are orders of magnitude smaller than from nuclear fission power. We still need to think about them and make sure that plants are safe, but it should be significantly easier to manage.
Strictly speaking, it depends on the fuel cycle. They're targeting a fuel cycle (D+He3) with minimal neutron flux.
A commercial D+T fusion reactor would generate much larger neutron flux than a fission reactor. But it still wouldn't generate all the long-lived fission byproducts that are so problematic with (non-breeder) fission reactors (assuming proper sheilding).
Wikipedia provides a bit more detail [2]: "The helium-3 is produced by D-D side reactions and is captured and reused, eliminating supply concerns. Helion has a patent on this process."
Conveniently, BCHydro just emailed me my bill this morning. They charge $0.2077/day. Then $0.0939/kWh for the first 688kWh and then $0.1408/kWh after that.
It looks like Helion is targeting to be somewhere between 10x to 15x cheaper than hydro-electricity.
What you've missed is that most scientists have been expecting electricity will be generated from fusion soon for the past 10 years or so, almost definitely before ITER comes online. The development of High-temperature superconducting tape basically guarantees it.
The scientists didn't miss this - on Helion's website they state that idea here was thought about in the 1950s, although they lacked the computational power to test their theories.
Helion gets to draw on that work for magnets and plasma containment, but doesn't have to solve that system's problems. So, it is legitimately possible they could have something.
You might wonder why the international research community has ignored this approach. It is a good question. All I can think of is that the people paying the bills for that want lots of physicists to have experience with hot-neutron processes, so they have a population to draw on for weapons work.
Of course they won't say so. What they say is that they are chasing the system that needs the lowest temperature.
Like, are we actually getting net power output?
edit : The idea to capture the EM energy directly, without using the heat from the system is ingenious. Gives me a huge boost of hope, however limited my exposure / knowledge about fusion reactor design is.
Yes, that's the idea.
If you don't need power generation, several vendors offer commercial deuterium-tritium fusion reactors for sale today, as neutron sources. Here's one: https://www.thermofisher.com/order/catalog/product/151762A?S...
Compared to CFS's approach, Helion's approach is different in two or three key ways.
First, it's a different fusion reaction, which has important engineering consequences. The reaction that Helion wants to use generates fewer damaging neutrons, which makes the rest of the reactor easier to engineer and eliminates several tricky subsystems.
Second, it's a different geometry, which does not require huge powerful steady-state superconducting magnets. However, it does require very fast magnets. It's a pulsed machine (they would fire several shots per second; each shot lasts on the order a milliseconds if I recall). So it requires more pulsed power systems and the components may need a different kind of high-repetition lifecycle testing.
Third, they want to use a 'direct energy conversion' scheme, which means eliminating the need for gas turbines (i.e. steam turbines) coupled to generators. This is important since the heat exchangers and turbines make up very roughly half the cost of a traditional power plant, so this would allow the electricity price to be lower by about a factor of two!
Helion uses Field-reversed configuration [0].
[0] https://en.wikipedia.org/wiki/Field-reversed_configuration
Where are your scientists? researchers? engineers?
Oh i see ycombinator..
Are scientists second class citizens? investors and "people who pitch projects to 'vulture' capitalists" are gods?
Kirtley: An NSF, NASA, and DOD fellow, Dr. David Kirtley has 13 years of experience in nuclear engineering, fusion, and aerospace
Pihl: TitleEngineer / Scientist II Dates EmployedApr 2012 – Jul 2014 Employment Duration2 yrs 4 mos The Plasma Dynamics Laboratory at the University of Washington
Votroubek: University of Washington Degree NamePhDField Of StudyPlasma Physics (Aeronautics & Astronautics)
So not just VC pitching specialists (sorry about the bad cutting pasting)
This kind of information should be on that page
Including the rest of the team
One thing I didn’t really think about with fusion, until reading the comments here: when the energy is basically free, you don’t have to try too hard to capture the output to make something of it. Is that actually the case with harvesting fusion energy?
Is the waste heat an issue? Where does the pink stuff go after it’s been squished into the middle of the ribbed magnetic bonbon thing? Apologies if I’m blinding you with science technobabble.
We still haven't got more usable energy (electricity and/or heat) out of a fusion reactor than it takes to run the thing, so capturing as much of the energy as possible is still absolutely important.
However, in the long-term (e.g. 100 years from now) you're right that working fusion power plants would make energy "basically free". The main reasons would be (a) the abundance of fuel (assuming the source would be heavy water) and (b) the fact that nuclear reactions are so much more energetic than the chemical reactions we're used to (so even a few percent improvement may be a large amount of extra energy).
1) Fusion plants still require site infrastructure, power conversion technology, waste heat removal, and (though not for this particular concept) steam generators (or other fluid cycle generators). These have significant capital costs but finite lifetimes. You're right that the variable cost of energy is pretty low, probably comparable to current fission plants, but that's still more expensive than the variable cost of electricity from solar and wind.
2) The price of electrical transmission and distribution starts to become important (I forget what the typical cost of that today is, but it's a few cents/kWh.) This doesn't matter if you can put a small power plant at your local industrial park though.
3) There's a big difference between $0.05/kWh, $0.02/kWh, $0.01/kWh, and $0.005/kWh, and then a huge difference to 'true zero'. This is because there's probably lots of industrial processes that we might like to do if electricity and heat were cheaper than it is today, each becoming reasonable at a certain price. There could be a large market at each price floor 'step'.
4) Yeah, the fuel is abundant, which is good, but fuel costs are not significant drivers of the cost of fission.
While not free, I'd like to think that fusion will help make a world with energy much cheaper than the world without fusion.
https://www.investopedia.com/articles/investing/092815/how-b...
Explanation required. Does the first sentence have any bearing on fusion, and how does it relate the the second?
I mean, I have a generator that produces electricity (it doesn't use fusion). What is the point of having one that uses fusion, but doesn't have net output?
But we are making measurements on aerodynamic performance - on another level we have learned that if you can make an aircraft that shape and give it an engine that moves it at the same speed as the airflow, then the aeroplane will generate enough lift to keep it in the air. That's a really important result that gets us significantly closer to a useful aeroplane. If separately someone has demonstrated an engine with suitable weight and power characteristics to match, then we can say "now we just need to build the aircraft and it should fly".
Helion involves a unique method of extracting energy from the fusion process (direct extraction from the magnetic field). That's new and therefore uncertain. Producing a reactor that performs something fusion-like and generates electricity in that way is a great result even if the fusion process isn't generating as much power as it should and is fundamentally driven by electricity. Like the plane above, if they have separately demonstrated a fusion process that is powerful enough, "all" they need to do now is put the two things together.
I think if I were a billionaire investing in fusion technologies, I think I'd be sure to mention my investment's special sauce when I drop a few sticks on it.
Getting a fusion reactor to work is trivial; at least one 12 year old has done it [1]. Getting net energy out of the reactor is much more difficult. The point of the funding is to figure out how to increase the gain factor. Once that's figured out, the rest (manufacturing and deploying reactors) is comparatively trivial.
[1] https://www.popularmechanics.com/science/energy/a34312754/12...
Well, this is what the investment is for.
Edit: fix quote
It's like any other business: you invest to establish it, and then you make it profitable.
Machinery to initiate fusion must be at least this big, consume at least this much power, and cannot be scaled down arbitrarily.
However, if one were to scale it up in size, the same doesn't hold. Output power scales faster than increased input requirements.
Consequently, most current fusion work is (a) find a design that theoretically has those scaling characteristics, (b) build a prototype to investigate / prove any unknowns (net negative power, but not ITER/NIF expensive), & (c) if able to prove (b) then scale up into a net positive example.
I see a bunch of patents and hand-waving that seems intentionally complicated. I spent much of my career helping companies raise money based on demos (I'm talking billions of dollars) and this just seems like more bullshit to me. And all of those demos I worked on were smoke and mirrors, despite building quasi-functional prototypes you could interact with.
I'd love to be wrong but they aren't make an effort to convince me otherwise, they just want to raise a shitload of money.
> I spent much of my career helping companies raise money based on demos
And I've made a bunch of money sitting on my ass watching a few stocks go to the moon. Knowing how to invest in webshit or getting lucky picking stocks can make you rich much easier than R&D can make you rich. Sam isn't investing because this is the best risk-adjusted return for his portfolio. He's investing because, if it does pay out, it also very literally changes the course of humanity in the process.
> And all of those demos I worked on were smoke and mirrors, despite building quasi-functional prototypes you could interact with.
The investment isn't being made on the basis of an existing reactor. There are tons of existing fusion reactors. The investment is being made on the basis of the team's plan to get to net positive energy.
* I'd love to be wrong but they aren't make an effort to convince me otherwise*
Luckily there are other folks in this world who are willing to make risky investments in important ideas.
(BTW, no one's getting rich on fusion research until fusion works... every year the fusion community leaks a bunch of folks to finance and tech because even entry level positions pay 3x and offer more stability.)
It seems like you missed my point, because I make all my money investing nowadays too after failing forever trying to turn R&D into viable products. It's easy to raise money on bullshit and almost impossible to actually make it work. In fact, it likely is actually impossible, we just don't know yet.
> Sam isn't investing because this is the best risk-adjusted return for his portfolio. He's investing because, if it does pay out, it also very literally changes the course of humanity in the process.
I've met Sam and I don't think he cares about making the world a better place. I think he just likes money and attention.
You seem awfully idealistic. I'm terribly cynical. We're not going to agree and that's fine.
First you point out that this seems like bullshit and a waste of money.
Then you point out that Sam doesn't care about making the world a better place and just likes money.
You're contradicting yourself, and on top of that, publicly insulting a core figure in this community. I've never met Sam, I don't care about him or if he wants to make the world a better place or not, but your comments are completely uncalled for.
You're a cynic, cool, that's fine. Comment on why you think the technology is bogus and don't publicly insult others.
I'm not celebrating anything. I'm a skeptic at heart and don't believe in any hype until I see meaningful progress. I just don't see the need to shit all over something because my gut tells me it's hype.
> Sam isn't investing because this is the best risk-adjusted return for his portfolio. He's investing because, if it does pay out, it also very literally changes the course of humanity in the process.
I thought I'd reply with my personal experience why I disagree. And again, I stand by what I said about Sam. I really think he deserves much more criticism than he gets. If criticizing someone's character is an insult then I think we need more insults. I don't want a nicer world, I want a world with less bullshit.
Now if you posted about a specific negative experience you had with the individual and actually put something on the line, that would be different. But as it is, there is no reason to believe you. For all I know, you just dislike him because he didn't invest in your company.
I'm willing to concede you're right about Sam. Sure. Investing in fusion is a terrible way to make money. If rich folks' egos get more money thrown at the right problems, so be it.
And if this works out, between (a) Sam funded the project that cracked cheap fusion and saved the planet, and (b) HN's OnlineGladiator met him and disliked him, I think the balance will tilt toward feeling some gratitude towards Sam.
What does it matter what I think, anyway? Think for yourself and come to your own conclusions. I don't care if you disagree.
Edit: I'm not trying to be (too) pedantic. But the brief investment announcement went to the trouble of saying "they built a generator that produces electricity" and (effectively) "but no net electricity".
Isn't thathat generator a nothing-generator then? Why even mention it?
• Better techology (that still needs to be R&Ded) makes it more efficient and net positive in the future.
• Prototype v1 is a required step towards v2(-v3...vn?) that actually accomplishes the goal. An example of this is SpaceX's Starhopper -> Starship.
• Some economy of scale makes it work at some point. Example, put a single box in a big ship from Shanghai to LA, cost of shipping = millions of $; vs. put a million boxes in the same ship, cost of shipping = a few $.
Have they achieved fusion? Or the concept of generating electricity from heat, proven long ago?
It's just that Sam's post didn't mention any (semi-)novel method for generating electricity from fusion, so the actual words - "they and their team have built a generator that produces electricity" - appear as either hype or non-information outside that context.
"The D-3He fusion reaction produces no neutrons as well (D+3He→4He (3.6 MeV)+H (14.7 MeV). However the D-D side reaction, while not as frequent, can generate 14.1 MeV neutrons through one of its fusion product reactions (D+T→4He+n+14.1 MeV). There is also the D-D reaction itself that produces a lower energy neutron (2.45 MeV) which is below the threshold for activation of most nuclear materials and is thus far less detrimental...Example systems and methods described herein may employ a 3He fuel cycle which may reduce or suppress a dangerous D-T side reaction by extracting the tritium ions as they are created. The extracted tritium is unstable and may beta decay in a relatively short period of 11 years to 3He, a primary fuel for the D-3He reaction. Accordingly, example systems, reactors and methods described herein may enjoy a self-sustaining fuel cycle where the required 3He to operate the reactor may be generated by the decay of tritium ions extracted from the reactor itself..."
The FAQ on the website [2] acknowledges their process "does create some 'activated materials' over the operating life of a power plant. Helion’s plants have been specifically designed to only use materials that would result in low activation, similar to what might be created by medical devices or other particle accelerators.
Our expectation is that a Helion plant could be fully decommissioned within a week without any lasting environmental impact."
http://www.projectrho.com/public_html/rocket/fusionfuel.php#...
No source given, but this is usually well-researched.
Most of the energy is released as fast-moving charged particles, which will drive direct electricity production, so efficiency might be pretty good, but I haven't seen numbers.
> Helion is by far the most promising approach to fusion I’ve seen.
> Helion has a clear path to net electricity by 2024
That seems impossibly ambitious.
weren't the doubters right all along?
And yes I understand that it is more exciting to support the new technology, but it would be foolish to dress that up as a proof that the other people were wrong - if it indeed in the past 30 years the clean fusion was right around the corner.
https://upload.wikimedia.org/wikipedia/commons/a/ab/U.S._his...
It was very much a justified position. Humans did underestimate the difficulty. Specifically, codes for predicting behaviors of plasmas routinely gave overly optimistic estimates for what would be required to produce net power, leading to numerous experiments that produced orders of magnitude worse results than predicted. Basically every time a more powerful fusion experiment was built, new instabilities that we were previously unaware of (and thus the simulation codes couldn't possibly account for) were discovered. Fusion funding was initially abundant but dried up after repeated failure.
Again, the progress didn't stop because funding dried up, funding dried up when it became clear the progress didn't matter. Fusion was appealing at first because it seemed easy, and easy things can typically be done economically. As soon as it became clear that fusion was much harder than initially believed, the hope of building power plants that are economically competitive with more mature alternatives was dashed. New startups can make all the claims they want about how their technology will succeed where others failed, but the fact is until they've actually succeeded, it is impossible for them to say they won't run into unforeseeable problems between now and then. Anyone claiming otherwise is lying.
It feels really far out from practicality. They have demoed a system that ran once every ten minutes, and they hope to build a system once per second.
I frankly don't think it is possible to heat things up to 100 million celsius and have that process work reliably. It is just too extreme.