Fusion power has moved from theoretical promise to repeated laboratory demonstration, and the honest answer is: yes, it almost certainly will work as a physics problem. Whether it becomes a practical, affordable source of electricity is a harder question, but the trajectory over the last three years has been more encouraging than at any point in fusion’s 70-year history.
What’s Already Been Proven
The biggest milestone came in December 2022, when the National Ignition Facility in California achieved fusion ignition for the first time, producing 3.15 megajoules of fusion energy from 2.05 megajoules of laser energy delivered to its target. That wasn’t a fluke. As of late 2025, NIF has achieved ignition ten times, with yields climbing sharply. In April 2025, a single shot produced 8.6 megajoules of energy from just 2.08 megajoules of laser input, a target gain of 4.13. That means the fuel capsule released more than four times the energy that hit it.
This is a genuine scientific achievement, but it comes with an important caveat. The lasers themselves consume far more energy than they deliver to the target. NIF’s entire laser system draws roughly 300 megajoules of electricity to fire a 2-megajoule shot. So while the fusion reaction produces net energy relative to the laser light on target, the overall system is still deeply in the red. NIF was built as a weapons research facility, not a power plant prototype. It fires once every few hours and was never designed for efficiency. What it proved is that controlled fusion ignition is physically achievable, reproducible, and improving.
The Machines Designed to Make Power
The real action for practical fusion energy is happening elsewhere. Commonwealth Fusion Systems, a company spun out of MIT, is building SPARC, a compact fusion reactor in Devens, Massachusetts. SPARC is expected to produce its first plasma in 2026 and demonstrate net fusion energy shortly after. Its key innovation is a new generation of high-temperature superconducting magnets that achieved a world-record field strength of 20 tesla at large scale during testing in 2021. Stronger magnets let you build a smaller, cheaper reactor that can still confine plasma hot enough and dense enough to sustain fusion. CFS has already announced plans to build a commercial fusion power plant in Virginia.
Helion Energy is taking a completely different approach. Rather than using magnets to hold plasma in a steady-state ring, Helion slams two plasma rings together at high speed and captures energy directly from the resulting reaction. Its seventh-generation prototype, Polaris, entered operation at the end of 2024 and became the first private fusion machine to use deuterium-tritium fuel. Helion is tentatively committed to bringing a 50-megawatt power plant online by 2028, though that timeline is aggressive by any measure.
Then there’s ITER, the massive international tokamak being built in southern France. Originally projected to be complete by 2016 at a cost of $10 billion, the project has ballooned repeatedly. As of mid-2024, ITER won’t be fully operational with burning plasma until 2039, at an estimated total cost exceeding $20 billion. ITER’s purpose is to demonstrate that a tokamak can produce ten times more energy than it consumes (a Q-factor of 10), but its glacial pace has become a cautionary tale about megaproject management rather than a reflection of whether the physics works.
The Engineering Problems That Remain
Proving that fusion reactions produce net energy is only one piece of the puzzle. Building a power plant means solving a cascade of engineering problems that have never been tackled at scale.
The most punishing challenge is what happens to the materials inside the reactor. The components facing the plasma must endure heat fluxes of 15 to 30 megawatts per square meter during normal operation. For context, the surface of a rocket nozzle during launch experiences roughly 5 to 10 megawatts per square meter. During plasma disruptions, brief instabilities where the plasma loses confinement, heat loads can spike to 100,000 megawatts per square meter for a fraction of a millisecond. Testing with electron beams has shown that even the best current materials develop cracks at brazed interfaces after only about 1,000 thermal cycles at 10 to 15 megawatts per square meter. A commercial reactor would need to survive millions of such cycles over its lifetime. No material in existence has demonstrated that durability under fusion-relevant conditions.
There’s also the tritium problem. Most fusion designs burn a mixture of deuterium and tritium. Deuterium is easily extracted from seawater, but tritium is radioactive, scarce, and expensive. A commercial fusion plant would need to breed its own tritium by surrounding the reactor with lithium blankets that capture neutrons from the fusion reaction. This breeding process has never been demonstrated at reactor scale. Helion’s long-term plan sidesteps this issue by aiming to use a deuterium and helium-3 fuel cycle instead, though that reaction requires significantly higher plasma temperatures.
Can It Compete on Cost?
Even if the engineering works, fusion has to be cheap enough to matter. An analysis published in Energy Policy modeled the economics of magnetically confined fusion reactors and found that early designs will likely produce electricity at costs above $150 per megawatt-hour. For fusion to be competitive with other clean energy sources beyond 2040, costs would need to drop to roughly $80 to $100 per megawatt-hour. For reference, new solar and wind installations in favorable locations already produce power for $30 to $50 per megawatt-hour, and advanced fission reactors aim for the $60 to $90 range.
That $150-plus figure for early fusion plants is not necessarily a dealbreaker. Solar panels were similarly expensive in their early years before manufacturing scale drove costs down. But the path to cheap fusion requires building many reactors to climb the learning curve, and building many reactors requires someone willing to pay premium prices for the first ones. The economic case will depend heavily on whether compact reactor designs like CFS’s can be factory-built in standardized units rather than constructed as one-off megaprojects like ITER.
The Regulatory Landscape Is Favorable
One barrier that has been substantially lowered is regulation. In 2023, the U.S. Nuclear Regulatory Commission voted unanimously to regulate fusion under the same framework used for particle accelerators, formally separating it from nuclear fission. This was a significant decision. Fission reactors require years of licensing review due to risks like meltdown and long-lived radioactive waste. Fusion reactors don’t carry those risks: they can’t melt down because the reaction stops the moment confinement is lost, and they don’t produce the same high-level waste. Regulating fusion like an accelerator rather than a fission plant means developers won’t face the decade-long permitting timelines that have strangled nuclear fission construction in the U.S.
A Realistic Timeline
The physics of fusion clearly works. The question is when, not whether, it transitions from laboratory demonstration to commercial electricity. The most optimistic private companies are targeting the late 2020s to early 2030s for demonstration plants. Mainstream estimates from the fusion research community put the first commercial power on the grid somewhere around 2035 to 2045. That’s a wide range, and it depends on solving the materials, tritium breeding, and cost challenges that remain open.
If you’ve heard the joke that fusion is always 30 years away, there’s a reason it resonated for decades: progress was genuinely slow when funding was limited and the only path was building ever-larger government tokamaks. What’s different now is the convergence of high-temperature superconducting magnets (which make reactors dramatically smaller and cheaper), over $7 billion in private investment flowing into more than 40 fusion companies worldwide, and a regulatory environment designed to let projects move quickly. Fusion power will almost certainly work. The remaining uncertainty is whether it will work soon enough, and cheaply enough, to matter in the energy transition already underway.