The NIF laser at Lawrence Livermore has now achieved ignition ten times. The highest yield, reached in April 2025, produced 8.6 megajoules from 2.08 megajoules of laser input, a gain factor above four. France's WEST tokamak held plasma at 50 million degrees Celsius for 1,337 seconds in February 2025, shattering the previous record. South Korea's KSTAR wants to sustain plasma at 100 million degrees for 300 continuous seconds by the end of this year. Private fusion companies have collectively raised nearly $10 billion. The DOE released a national fusion roadmap in October 2025 targeting power on the grid by the mid-2030s.
Every one of those facts is real. Every one of those facts is also being used, right now, to sell people a timeline that the physics does not yet support.
Stop scrolling and sit with that for a second. The records are genuine. The gap between the records and your electricity bill is also genuine. Both things are true simultaneously, and conflating them is how science communication turns wonder into disappointment.
What 'Net Energy Gain' Actually Means
The December 2022 ignition at NIF was, without question, a scientific milestone. More fusion energy came out of the fuel target than laser energy went in. That is a real threshold. What the headlines reliably omit is the wall-plug picture. The NIF laser facility requires roughly 100 times more energy to run than the laser delivers to the target. The target produced about 1.1 net megajoules, which, after conversion to electricity, would power a 5-watt LED bulb for roughly 20 hours. As of 2025, NIF remains the only laboratory in the world to have demonstrated a fusion energy gain factor above one, and efficiencies orders of magnitude higher are still required to reach engineering breakeven, where a net electricity-producing plant actually pays back its energy debt.
This is not a failure. It is a correct reading of where the science sits. The Wright Brothers' first flight at Kitty Hawk covered 120 feet. Nobody handed them a transcontinental airline route. The flight was the proof of concept. The next forty years were engineering.
Fusion is currently at Kitty Hawk. The physics has been proven. The engineering has not.
The world's flagship project makes the point better than any argument can. ITER, the 35-nation, $25-billion tokamak under construction in southern France, was originally designed to achieve first plasma in 2025. That date has now slipped to 2033 at the earliest, with deuterium-tritium operations, the actual burning-plasma experiments that will tell us what a commercial reactor needs to do, pushed to 2039. Manufacturing defects in the vacuum vessel sections, a safety halt by French nuclear regulators, and the sheer complexity of a first-of-a-kind machine all contributed. The repair bill for malfunctioning components alone is estimated at €5 billion. ITER's delays are not just an administrative inconvenience. They compress the timeline for everything downstream: the demonstration reactors that learn from ITER, the pilot plants that follow those, the commercial fleet that follows those.
The Fuel Problem Nobody Is Talking About Loudly Enough
Assume every prototype works on schedule. Assume SPARC at Commonwealth Fusion Systems achieves first plasma in 2026 as planned and the magnet technology performs. Assume Helion delivers on its agreement to supply Microsoft with at least 50 megawatts of fusion electricity by 2028. Assume all 53 private fusion startups currently operating in the United States hit their marks. You still collide with a wall that has nothing to do with plasma physics.
Tritium. The preferred fusion fuel is a deuterium-tritium mix, and deuterium is easy: pull it from seawater. Tritium is not. It is radioactive, it decays with a half-life of about 12 years, and it barely exists in nature. The entire world's tritium supply sits at approximately 50 kilograms, produced almost entirely as a byproduct of CANDU-style fission reactors in Canada. A single 100-megawatt electric fusion plant running continuously would consume approximately 17 kilograms of tritium per year. A 1-gigawatt plant burns through roughly 400 grams per day. The math is not subtle. You cannot build a fusion-powered civilization on 50 kilograms of anything.
The theoretical solution is tritium breeding: reactors that produce their own fuel by bombarding lithium blankets with fusion neutrons. The engineering for that breeding blanket is, in the candid words of one fusion company, very hard. It has not been demonstrated at scale in any operating device. ITER's mission includes testing breeding blanket concepts. That work begins, at the earliest, in the late 2030s.
The Fusion Industry Association asked the federal government for $10 billion in additional funding this year, on top of the $9.7 billion already in private hands. Estimates circulating in the industry suggest over $77 billion more is required to reach commercialization. Those numbers clarify the distance. They are not reasons for despair. They are honest coordinates.
What the Mid-2030s Target Actually Tells You
The DOE roadmap's target of power on the grid by the mid-2030s is the most optimistic credible timeline from the most well-resourced government fusion effort in history. The UK Atomic Energy Authority targets 2040 for the first commercial fusion plant. One serious analysis of ITER's full operational lifecycle places grid-contributing fusion energy somewhere after 2055. These are not the same number.
The conventional wisdom, as of right now in February 2026, is that commercial fusion arrives in the mid-to-late 2030s. That means the earliest your utility company sources a single kilowatt from a fusion plant is roughly a decade away, under conditions that require every engineering problem currently unsolved, including materials that survive neutron bombardment at scale, a working tritium breeding cycle, regulatory frameworks that do not yet fully exist, and a trained workforce that is currently being built from scratch, to resolve on schedule.
The universe does not care about your timeline. Plasma at 100 million degrees cares only about physics. The physics is cooperating more than anyone dared hope fifteen years ago. The engineering is on the clock.
None of this diminishes what the scientists are doing. In February 2025, France's WEST tokamak maintained hot plasma for 1,337 seconds at 50 million degrees Celsius, breaking the previous world record of 1,066 seconds. China's EAST reactor pushed plasma density well past long-standing empirical limits without triggering the disruptive instabilities that usually end experiments, challenging decades of assumptions about how tokamak plasmas behave at high density. These are not incremental results. They are the kind of data that rewrite textbook chapters.
But they are also not a power plant. The distance between a 1,337-second plasma record and electricity running your refrigerator is roughly the same distance as the distance between a Wright Flyer and a 787 Dreamliner. Both measure progress. Only one measures readiness.
This is bigger than you think, and not in the way the press releases suggest. The scientific achievement unfolding right now in fusion labs across four continents is genuinely one of the most extraordinary things humans have ever attempted. A star, briefly, in a bottle. Held by magnetic fields we designed. Fueled by hydrogen we pulled from water.
The honest answer to "when does this power my home" is: not yet. But the honest answer to "will it ever" has never been more clearly yes.
