A Sun in a Box

Nuclear Fusion in ~1k Words

👋Hey!

Regular readers know that I’ve been on a bit of a nuclear kick recently here at Uncredentialed. So far, I’ve been focused exclusively on fission (splitting atoms to generate heat), trying to understand what makes SMRs different from traditional nuclear and who the big players are in the space.

Today, we’re going to talk through the other side of the nuclear coin: fusion. Instead of splitting heavy atoms apart, fusion smashes light ones together. They both fall under the same “nuclear” umbrella but are fundamentally opposite processes with entirely different risk profiles.

If you’re tired of constantly hearing that fusion is a couple decades away, this one’s for you!

---

But first… welcome to 13 new Uncredentialed members! If this is your first time here and interested in tech, startups, and strategy, you could be just like them and subscribe below!

---

How Fusion Works

My SMR post described fission as “a fancy way to boil water.” And, well, so is fusion! The ending is (typically)the same story of heat → steam → turbine → electricity, but with a very different path to create that initial heat.

Most fusion reactors use 2 isotopes of hydrogen, deuterium and tritium, as their fuel. To get these atoms to fuse, you heat them to over 100 million degrees Celsius, about 6x hotter than the core of the sun. At that temperature, the fuel becomes plasma, a state of matter where electrons are stripped from their atoms.

Obviously, no physical material can survive contact with something that hot, so most designs use something called a tokamak, a donut shaped vacuum chamber surrounded by superconducting magnets. The magnets create a field that suspends the plasma in midair so it never touches the walls.

When the atoms fuse, they release helium and high energy neutrons. That energy gets captured, often by a molten salt blanket surrounding the reactor, to generate the heat that eventually boils the water. From there it’s the same turbine story as every other power plant,

Self-Generating Fuel

Ok, so we need deuterium and tritium as our fusion inputs, but where do we get them?

Deuterium is easy, it’s just hydrogen with an extra neutron and occurs naturally in the ocean. There’s enough in a few gallons of seawater to power a home for a year and extraction comes from a known, repeatable process.

Tritium, on the other hand, is tricky. It’s radioactive, with a half-life of 12 years, and is rare enough that researchers choose to use the reactor to breed it.

Here’s how: the fusion reaction produces high energy neutrons as a byproduct. Those neutrons fly out of the plasma and slam into a lithium blanket surrounding the reactor. When a neutron hits a lithium atom, the lithium splits into helium and a fresh atom of tritium, which gets extracted and piped right back into the reactor as fuel.

A key metric to know is the tritium breeding ratio (TBR). If the reactor breeds more tritium than it consumes (a TBR above 1), it’s self-sustaining. Achieving a TBR of, say, 1.05 would mean the reactor produces 5% more fuel than it burns, enough surplus to then eventually start up another reactor, leading to a nice virtuous cycle of waste product (neutrons) creating the feedstock (tritium), keeping the reaction going indefinitely.

Compare that to the HALEU supply chain headaches I discussed in the SMR Field Guide and you start to see why the fuel story alone makes fusion worth watching.

Why Not Stick With Fission?

Fair question, especially since I just spent 2 posts being bullish on SMRs. The short answer is that fusion solves the problems fission can’t fully escape, even in its modern SMR form.

On safety, fission relies on a chain reaction that, however well-managed, is still there. Fusion has no chain reactions at all. If anything goes wrong, the plasma cools and the reaction simply stops. You couldn’t make it meltdown if you tried. On waste, fission produces material that needs storage for thousands of years while fusion’s primary byproduct is helium. Lastly, as discussed in the previous section, the fuel supply chain presents fewer problems with fusion than it does with fission.

That being said, fission works today and fusion does not. While labs like NIF have achieved brief net energy gain at the experiment level, no fusion system has yet produced sustained net electricity to the grid. The looming energy crunch won’t wait for fusion to become feasible at scale to begin. Fission and SMRs will still be pivotal in carrying the economy forward until fusion becomes viable.

Who’s Who in Fusion

For decades, fusion was synonymous with ITER, the massive international tokamak in France that broke ground in 2010, has cost over $20B, and still isn’t operational. Recent years have shown a shift away from government-backed megaprojects to venture-backed startups driving change in the fusion race. There’s two companies in particular that I think are worth highlighting here, Commonwealth Fusion Systems (CFS) and Helion Energy.

CFS is the leading player, an MIT spinoff backed by over $2B from investors including Google and Tiger Global. They’re a compact tokamak called SPARC that’s designed to fit on a tennis court and is currently under construction in Massachusetts.

CFS’s core breakthrough was developing high temperature superconducting (HTS) magnets that are dramatically smaller and more powerful than anything previously available. These magnets are what make a compact, commercially viable reactor possible.

SPARC’s near term goal is to achieve Q > 1, meaning more energy comes out of the plasma than goes in, considered the scientific breakeven point that shows the physics work. Importantly, this is just a first hurdle that comes before the engineering breakeven point where the reactor can power the entire facility including the magnets while still producing net electricity. CFS is aiming to prove the physics by 2027 so they can solve the engineering in the 2030s.

Helion Energy, backed by Sam Altman, is taking a fundamentally different approach. Rather than a continuous tokamak, Helion uses pulsed fusion via Field-Reversed Configuration (FRC). They fire 2 magnetized plasma rings into a central chamber at 1M MPH, where they collide and are compressed by magnets to trigger fusion. Most exciting though, is that they’ve opted to skip the steam-powered turbine entirely, instead using the interaction between the fusion plasma and magnetic field to induce a current directly into the reactor’s coils.

Their 7th prototype, Polaris, is already operational in Washington, and they’re aiming to demonstrate net electricity production by 2028 to fulfill a power agreement they already made with Microsoft.

Why Now?

Healthy skepticism is warranted. This technology has been “almost here” for longer than I’ve been alive, but I really do think this time is different.

First, the magnet breakthrough is real. CFS demonstrated their HTS magnet in 2021 and it worked, making compact fusion physically possible for the first time.

Next, regulators are onboard. In a major shift, the US has started regulating fusion differently than fission. Because there’s no meltdown risk and no long-lived waste, fusion is being treated more like advanced industrial equipment (think MRI machines) than traditional nuclear plants. That drastically cuts the permitting timeline which can be a major drag on energy projects.

Last but not least, compute is accelerating the timeline. Earlier this year, CFS announced a partnership with NVIDIA and Siemens to build a digital twin of the SPARC reactor, using AI to simulate plasma behavior and compress years of manual experimentation into weeks of virtual optimization. 30 years ago, fusion scientists were limited to running a handful of physical experiments and hoping for insights. Today, ML-assisted plasma modeling and control systems mean they can iterate at software speed on a hardware problem, acting as yet another case of compute unlocking progress in the physical world.

The Fusion-Based Future

If fusion works, and I’m increasingly optimistic that it will, it changes the math on just about everything. With seawater and lithium as the only inputs, helium as the only waste product, 24/7 operation regardless of weather, and enough power density to run at a fraction of the footprint of wind or solar, fusion is the physical embodiment of energy abundance.

When energy becomes so cheap and abundant it feels free, the constraints on what we can build disappear. Manufacturing, desalination, transportation, heating/cooling, and just about anything else that requires energy will be effectively costless with fusion at scale.

The engineering challenges between today and a working commercial reactor are immense, but between SMRs bridging the gap today and fusion startups racing toward the 2030s, the path is becoming real. In the next decade, we could build a miniature sun, domesticated and humming along in a tennis court-sized box, powering everything we can dream up.

---

Another exciting week of growth! We grew another 8.5% week over week and are quickly approaching the next 100-sub hurdle. If you enjoyed the read and haven’t subscribed yet, the button below will get you all set up to make sure you see my posts in the future. And if you know anyone else who might like my writing, consider sharing!