How a fusion startup plans to reach the grid by 2028

HostWe have heard for decades that fusion power is always thirty years away. It's the holy grail of energy, the idea that we can copy the way the sun works to get clean, endless power, but it always feels like a dream for the next generation. Now, a small company in Washington state is saying they can do it in less than four years, and they already have a giant tech company waiting to buy the electricity. How does a startup jump ahead of every government lab in the world to meet a deadline that most experts think is impossible?

GuestIt's a massive claim. The company is called Helion Energy, and they have signed a deal to start providing power to Microsoft by 2028. This isn't just a pinky promise or a research grant. It's a real contract with big fines if they don't deliver. They have promised to provide at least fifty megawatts, which is enough to power tens of thousands of homes. To understand how they plan to hit that date, you have to look at how they're building their machine. Most of the world is focused on these giant, billion dollar donuts called tokamaks that use huge magnets to hold a floating ring of fire. Helion is doing something much smaller and much faster. They're basically building a big metal tube where they fire two rings of gas at each other from opposite ends. When those rings smash together in the middle, they get squeezed by magnets until the atoms fuse.

HostSo instead of trying to keep a sun burning forever in a bottle, they're doing it in quick bursts?

GuestExactly. Think of it like a car engine. A regular fusion project is like trying to keep a fire burning steadily to heat water for steam. Helion is building an internal combustion engine. They do a burst, get the energy, and then do it again, many times a second. Because the machine is smaller and simpler, they can build new versions quickly. They're on their seventh big machine right now, called Polaris. They don't have to wait ten years to see if a new idea works. They can build it, break it, and fix it in months. That speed is the only reason they can even talk about a 2028 goal.

HostI have to push back a bit on the 2028 date. We're talking about the most difficult physics problem on earth. Even if the machine works in the lab, building a power plant that hooks into the grid is a huge task involving tons of red tape. Is this just a case of a startup over promising to get more money?

GuestThere's plenty of doubt from the old guard in the physics world. Most experts look at 2028 and think it's pure fantasy. But the deal with Microsoft changed the conversation. Microsoft is desperate for carbon free power to run their huge data centers, especially with the boom in artificial intelligence. By signing this deal, Helion gets the cash they need to build, but they also get a lot of pressure. If they don't hit the mark, they lose money and their name is shot. The big hurdle isn't just the physics, but the wires. They have to prove to the people who run the power grid that this machine is safe and steady. Most nuclear plants take a decade just to get through the paperwork. Helion is betting that because their machine doesn't use high level radioactive fuel or create long term waste, the rules will be a lot easier on them.

HostYou mentioned the fuel. I thought fusion usually needed stuff like tritium, which is super rare and hard to handle. Is that what they're using?

GuestNo, and that's another way they're trying to cheat the timeline. Most fusion projects use deuterium and tritium. Tritium is very hard to find and it makes the walls of the machine brittle and radioactive over time. Helion uses deuterium and a rare form of helium called helium-3. The catch is that helium-3 is also very hard to find on earth. Most people talk about mining it on the moon. Helion says they can make it themselves inside their own machines. If they can do that, they skip the biggest supply chain headache in the industry. But there's an even bigger trick to their setup. Most power plants, even nuclear ones, are just fancy ways to boil water. You use the heat to make steam, which spins a big fan, which makes electricity. You lose a lot of energy in that process. Helion plans to skip the steam entirely.

HostHow do you get electricity out of a machine without spinning a turbine?

GuestIt's called direct energy recovery. When those gas rings smash together and fuse, the whole plasma ball expands. As it pushes outward, it pushes back against the magnetic field that was squeezing it. That movement of the magnetic field creates a current in the wires nearby. It's like a piston pushing back. You get electricity directly out of the magnetic interaction. It's much more efficient because you don't need all the heavy gear for boiling water and cooling steam. That makes the whole plant smaller and cheaper to build.

HostIt sounds like they're trying to reinvent every single part of the process at once, from the fuel to the way the power is caught. If even one of those pieces fails, the whole 2028 plan falls apart. Does the government group that watches over nuclear power even have a category for this kind of machine yet?

GuestThat was a huge win for them recently. The group that watches over nuclear power in the United States decided they won't treat fusion machines like the big fission plants we have today. They're going to treat them more like particle accelerators or medical labs. This is a big deal because it means the rules are much simpler. They don't have to build massive concrete domes or prepare for a meltdown, because a fusion machine can't melt down. If something goes wrong, the gas just cools off and the reaction stops instantly. That decision alone probably shaved years off the timeline for any fusion company trying to sell power.

HostSo the path is clear, at least on paper. They have the customer, they have a lighter set of rules, and they have a machine that skips the steam. But we still haven't seen this machine actually produce more energy than it takes to run, right?

GuestThat's the final boss. Their current machine, Polaris, is supposed to be the one that finally crosses that line. They need to show they can get more juice out than they put in, and they need to do it by next year or 2026 to stay on track. If they hit that mark, then the 2028 goal becomes a question of how fast they can bolt the parts together. If they fail to hit that net energy gain in the next eighteen months, then 2028 becomes just another year on the long list of missed fusion deadlines.

HostThe dream of a sun in a bottle used to be something for governments to chase over decades, but now it's a race against a contract and a calendar.

GuestThe stakes are high because if they pull this off, the way we build everything from data centers to cities changes overnight.

HostThat metal tube in Washington might finally be the thing that turns a fifty year old promise into a light switch.