Gasoline Is Just Too Damn Strong. After 167 Years, Let's All Just Bet on Fusion.

⛽ Honda just torched roughly $10 billion trying to build the EV future and quietly retreated to hybrids. That same spring, ITER's "beating heart" finished arriving in France, the UK's Pulsar Fusion ignited the first plasma inside a fusion rocket engine, and a Kyoto startup is bringing the world's first integrated fusion fuel cycle online in Canada. So here's a question worth asking, with the math actually done: after 167 years of gasoline, is anything genuinely better — and if so, why isn't humanity dumping its chips on it? Spoiler: there's exactly one answer, and it's not the battery in your driveway.

A Brief, Slightly Petty History of the Thing in Your Tank

Quick recap. Edwin Drake drills the first commercial oil well in Titusville, Pennsylvania, 1859. Twenty-six years later, Karl Benz bolts an internal combustion engine onto three wheels and patents the Motorwagen. Henry Ford's Model T rolls off the assembly line in 1908. Then come the oil shocks of the 1970s, the rise of Toyota's hybrids in the 1990s, Tesla's Roadster in 2008, and an entire decade of every analyst on Earth declaring the internal combustion engine dead.

And then 2026 happens. Donald Trump tears up emissions rules. Ford writes down billions on its EV business. Honda — Japan's pride, the company that built the Civic and the Super Cub — books a roughly ¥1.58 trillion (about $10 billion) impairment on its scrapped North American EV lineup, posts its first annual loss since going public in 1957, and walks back its much-publicized "no more engines by 2040" pledge. The total damage from the EV pivot reversal could hit ¥2.5 trillion (roughly $16 billion) over two fiscal years.

167 years. Gasoline has been pronounced dead about once a decade since 1973. It keeps refusing.

So: why?

Because Gasoline Is, Unfortunately, Just Really Good

Here's the chart nobody at climate conferences likes to put on the screen. Energy density, watt-hours per kilogram, the only honest yardstick when you're trying to move something:

Look at that gap. Gasoline isn't winning by 10 percent. Today's lithium-ion sits at about a fiftieth of gasoline's energy density. Even the solid-state battery Toyota, Nissan, and Honda are sweating about — the one promised "next decade" — tops out at about one-twenty-fourth. The pinball still tilts toward the petroleum side of the table.

This isn't a marketing problem. It's a physics problem.

Going Through the Candidates, Politely

Let's interview each contender. Three questions: does it beat gasoline, can you put it in a car, when does it arrive?

Synthetic e-fuel. Beats gasoline? It is gasoline — same molecules, just synthesized from captured CO₂ and renewable hydrogen. In the car? Trivially, drop-in. Arrival? Porsche's Haru Oni plant in Chile has been making it since 2022, and Formula 1's 2026 regulations mandate 100% advanced sustainable fuel. The catch: production cost is currently several times higher than fossil gasoline, and scaling means a lot of clean electricity. So: a sideways move, not an upgrade.

Solid-state batteries. Beat gasoline? Nope. Mass-production targets put them around 500 Wh/kg, still 1/24 of gasoline. In the car? Yes. Arrival? Toyota says 2027–2028, Nissan also chasing late 2020s. Real progress — just not a vault over the petroleum line.

Liquid hydrogen. Beats gasoline? By weight, yes — 3.3×. By "weight including the −253 °C tank," no, it loses. In the car? Toyota Mirai and Hyundai Nexo prove it works. Arrival? Already here. The blocker is infrastructure: Japan had about 164 hydrogen stations as of 2022.

Ammonia. Beats gasoline? Roughly 0.4×. The point isn't density, it's that it's carbon-free and easier to liquefy than hydrogen. In the car? Mostly a no. Cargo ships and gas turbines, yes — Mitsubishi Heavy Industries and IHI are moving to commercial scale, and Mitsui O.S.K. Lines is putting the world's first ammonia-fueled oceangoing vessel into service. Useful, not the answer to passenger cars.

Lithium-air. Theoretical density around 11,000 Wh/kg — finally in gasoline's neighborhood. Arrival? Honestly: nobody knows. Lab curiosity, real engineering remote.

Nuclear fission. Beats gasoline? By 1.8 million times. In the car? Ford actually tried this. The 1957 Ford Nucleon concept car was meant to be powered by a tiny reactor. The lead shielding alone would have put the curb weight in the multi-tonne range. Reactor-powered icebreakers and aircraft carriers work because they're the size of icebreakers and aircraft carriers. A sedan doesn't have room for the shielding, the cooling loop, or the regulatory inspector who'd live in the back seat.

Nuclear fusion. Beats gasoline? By 7.8 million times — even more than fission. In the car? Currently the smallest experimental reactor still fills a building. Arrival? See below, it's getting interesting.

Antimatter. Beats gasoline by 2 billion times. Producing one gram costs several years of global GDP. We're not going to talk about antimatter.

The pattern is unmistakable. Inside the "chemistry" tier, gasoline keeps winning or tying. To beat gasoline meaningfully you have to leave chemistry entirely — and once you do, "fits in a car" becomes nearly impossible.

So if chemistry is a tie at best, and fission can't shrink, that leaves one card.

The Case for Fusion, Said Plainly

Fusion fuses light atomic nuclei — deuterium and tritium — into helium, releasing energy at 7.8 million times gasoline's density. The fuel: deuterium is everywhere in seawater (one litre yields the energy equivalent of about 300 litres of gasoline), tritium is bred from lithium inside the reactor wall. No CO₂. No long-lived high-level waste of the sort that defined the fission debate. Cannot melt down — the plasma simply extinguishes if you lose containment, the same way a candle goes out when you blow on it. Fuel reserves measured in tens of millions of years.

The problem has always been doing it. Sustaining a plasma at 100 million degrees Celsius — hotter than the core of the Sun — long enough to extract net energy. For seventy years that's been "thirty years away."

Here's where 2026 starts looking different.

What's actually happening, right now

ITER, in Cadarache, France. On April 27, 2026, U.S. ITER delivered the final modules of the central solenoid — a 59-foot, 3,000-tonne superconducting magnet 15 years in the making, often called the reactor's "beating heart." The price tag is roughly €22 billion (~$26 billion). The first plasma target is now 2034, with full deuterium-tritium burning plasma operations targeted for around 2039. Slow, expensive, behind schedule — and the largest scientific collaboration humanity has ever attempted. Assembly inside the tokamak hall is now in its integration phase, with the superconducting magnet systems and reactor components being stitched together ahead of that 2034 first-plasma window.

SPARC, Devens, Massachusetts. Commonwealth Fusion Systems, a 2018 MIT spinout, has now installed the first of 18 high-temperature superconducting magnets in its SPARC tokamak. The company has raised close to $3 billion — more than any other private fusion firm — and signed a 200-MW power purchase agreement with Google for the follow-on ARC plant in Virginia. "Fusion is no longer a science project; it's actually the next big thing in tech," Mumgaard said. First plasma is targeted for late 2026, net fusion energy in 2027.

Helion Energy, Everett, Washington. Sam Altman-backed Helion has signed the world's first fusion power purchase agreement, committing to deliver 50 MW to Microsoft by 2028 from its Orion plant — with real financial penalties paid to Microsoft and to power-marketing partner Constellation if Helion misses the deadline. It is the most aggressive timeline commitment anyone in the fusion sector has put their balance sheet behind. Helion is now building a smaller test machine, Tiny Merge, to iterate faster — a sign the deadline is genuinely tight.

EAST, Hefei, China. In January 2025, the Chinese Academy of Sciences' superconducting tokamak EAST held a 100-million-degree plasma in high-confinement mode for 1,066 seconds, blowing past its own 2023 record of 403 seconds. China has been spending an estimated $6.5 billion on fusion deployment between 2023 and 2025 — more than triple comparable U.S. federal spending — and a new state-backed fusion company was launched in mid-2025 with registered capital of around 15 billion yuan (roughly $2 billion).

JT-60SA, Naka, Ibaraki, Japan. The world's largest currently operating tokamak, jointly built by Japan and the EU, achieved first plasma in October 2023. Power-up modifications are underway in 2025 ahead of full plasma heating experiments.

Kyoto Fusioneering. The Japanese startup — founded only in 2019 — is building two of the most important pieces of supporting infrastructure no fusion reactor can work without. UNITY-1, in Kumiyama, Kyoto, demonstrates the heat-extraction and blanket cycle. UNITY-2, the world's first integrated fusion fuel cycle pilot, is coming online at Canadian Nuclear Laboratories in mid-2026. A new UNITY-3 facility at Oak Ridge National Laboratory was announced in January 2026 under a strategic partnership with the U.S. Department of Energy.

This isn't "thirty years away" anymore. It's not "five years away" either — that's still magical thinking — but it's "the next ten to fifteen years are going to decide whether this works."

Now Watch This Number

Honda's EV write-down for fiscal year 2026: ¥1.58 trillion, about $10 billion. The full anticipated cost of Honda's EV strategy reversal, including FY2027: up to ¥2.5 trillion, about $16 billion. ITER's total budget to date: about €22 billion, or $26 billion.

Honda's mistake, just from one car company in one fiscal year, comes to roughly 40 percent of the entire budget of the largest international fusion experiment ever built. The full two-year EV strategy cost — at $16 billion — would cover about 62 percent of ITER.

Or, put another way: Total funding for the 53 fusion companies surveyed has reached a cumulative USD9.766 billion. Honda alone, in a year, has now spent more correcting a strategic mistake than the entire private fusion industry has raised in its lifetime.

That is the sentence to read twice.

If We Win This Bet, Here's What's on the Other Side

The pitch for fusion is rarely "your electricity bill drops a bit." It's a regime change. Let's stack what actually shifts.

Electricity falls toward effectively free. Deuterium from one litre of seawater carries roughly the energy of 300 litres of gasoline. Fuel cost is functionally zero. What you pay for is amortizing the plant. Forecasts vary, but the floor is: "an order of magnitude cheaper than today, for everyone, indefinitely." Industrial electricity stops being a competitive moat. A country's geological luck with coal and gas stops mattering.

Water and food stop being political weapons. Desalination is economically marginal because it's energy-hungry. Make energy almost free and saltwater becomes drinkable at any scale. Deserts turn green. Vertical farming stops being a hobbyist's pet project and becomes infrastructure. The arguments about who controls a river upstream stop being wars-in-waiting.

Climate change becomes a fixable problem, not just a managed one. Direct Air Capture of CO₂ is currently a thermodynamic loser — pulling 400 ppm out of the air takes more energy than the original carbon released. With fusion electricity, DAC moves from "doesn't work" to "trivially scales." Humanity stops being net-positive on emissions and becomes net-negative. The accumulated CO₂ we've put up there becomes something you can roll back, not just stop adding to.

The dirty industries get clean for free. Steel, cement, ammonia synthesis, glass, aluminum smelting — together about a quarter of all human CO₂. All electric processes if electricity is cheap. You don't have to negotiate with anyone about coal; the economic case kills coal on its own.

AI stops being limited by power. Right now every hyperscale AI buildout in the world is bottlenecked by electricity supply. Fusion at grid scale changes the question from "where do we get the watts" to "what do we want to build with them."

And now the part that's actually fun.

The Solar System Shrinks Ten Times

On March 22, 2026, British startup Pulsar Fusion ignited the first plasma inside a nuclear fusion rocket engine at its Bletchley facility, live-streamed at Jeff Bezos's MARS Conference. Their Sunbird drive is designed for an exhaust speed of around 500,000 mph (≈800,000 km/h) — faster than the fastest object humanity has ever built. The architecture aims at up to roughly a thousand times the thrust of current orbital propulsion, with cruise speeds of about 800,000 km/h, which would compress Mars travel from months to weeks.

The current Earth-Mars trip on chemical rockets is six to nine months one way. On Sunbird-class fusion drives: roughly half that, with an in-orbit demonstration planned for 2027. With more aggressive drives further out — Helicity Space in Pasadena (backed by Lockheed Martin), Princeton Satellite Systems' Direct Fusion Drive — the long-term envelope opens dramatically: Mars in two months, Jupiter in one year, Pluto in roughly four.

The inner solar system goes from "expensive multi-year national project" to "shipping route." NASA is now openly developing what it calls "the first nuclear-powered interplanetary spacecraft," and the agency has stated it wants a nuclear reactor on the moon by 2030.

And the Moon Is Sitting on the Fuel for Round Two

Earth produces about 8,000–10,000 litres of helium-3 a year. Helium-3 is the "second-generation" fusion fuel physicists have been discussing for decades — its reactions produce almost no neutrons, meaning reactor walls don't get destroyed, almost no long-lived waste, much simpler designs, and energy that converts directly to electricity instead of going through a steam turbine.

The reason nobody uses it is supply.

The lunar regolith, blasted by solar wind for four billion years on an airless body, has accumulated an estimated one million tons of helium-3. A single ton of it is currently valued at around US$3.7 billion in energy terms, and roughly 44 tons could meet U.S. annual electricity demand. Once fusion-powered transport is already proven on the deuterium-tritium loop, going to the moon to harvest helium-3 isn't science fiction — it's logistics with a longer haul.

China's Chang'e missions have already returned regolith samples with explicit helium-3 measurements. NASA, ESA, and multiple private firms (Black Moon Energy, LH3M, Pulsar Helium) have begun building business cases. The moon is being re-evaluated — not as a flag-planting prize, but as the first real off-Earth resource basin.

The Chain Reaction

Follow what actually gets unlocked:

1. First-generation fusion (D-T) works on Earth. Power becomes plentiful and cheap.
2. Carbon emissions stop and reverse. Water and food stop being political. Industrial decarbonization happens because it's cheaper, not because of treaties.
3. The same engineering moves to space propulsion. The solar system shrinks tenfold; multi-year missions become months.
4. Lunar helium-3 mining becomes economic. Second-generation fusion (D-He3) gives even cleaner reactors with simpler engineering.
5. Permanent moon bases. Then permanent Mars surface settlements with local fusion power running the greenhouses and the habitat heat. Then orbital colonies — O'Neill cylinders, Stanford torus, whatever shape humanity eventually picks — that don't need to be near the sun for power. Pulsar Fusion's CEO Richard Dinan put it bluntly: as you go into deep space, solar fades, so you need a new power source. Fusion provides it.
6. Humanity stops being a single-planet biological accident.

That last bullet sounds grandiose because it is. It's also literally what the engineering chain at the top of this section ends in.

This is the actual prize. Not "your next sedan is a slightly greener Honda." Not "Toyota or BYD wins the next decade." The prize is that humanity stops being capped by how much sunlight happens to land on Earth — first by tapping the entire deuterium content of the oceans, then by climbing out of the gravity well to harvest the rest of the solar system.

For roughly the cost of about 1.6 Honda EV strategy reversals.

So Maybe the Allocation Is Wrong

The pitch, plainly:

Stop pouring 100× the R&D dollars into solid-state batteries chasing a 10× improvement, when the prize at the end is still 1/24 of gasoline. Stop arguing whether your next car should be a hybrid or a battery EV — the answer is "whichever, who cares." Stop fighting the gasoline–versus–EV cultural war that produces $10 billion Honda write-downs and pretends those are the limits of human ambition.

The thing humanity is actually allowed to be ambitious about lives one category up. Fusion doesn't fit in your car. That's fine. The power plant becomes fusion, and you charge whatever drivetrain you happen to like off the resulting grid. Cars stay cars. The civilization upstream of them changes.

The numbers required to make this happen are, by national-budget standards, embarrassingly small. ITER's full lifetime cost is about 0.1 percent of one year of global military spending, or about 0.3 percent of one year of global fossil-fuel subsidies. Private fusion has raised under $10 billion total, ever — less than a single AI hyperscaler's quarterly capex.

The bottleneck isn't really money anymore. The harder pieces are engineering execution and political willingness to fund and stand behind it.

And Japan Is Quietly Already Doing It

Here's the part nobody outside Japan seems to notice. While Honda is on the front page apologizing for a $10 billion EV miss, the country's quiet fusion sector is building real hardware:

• JT-60SA in Ibaraki Prefecture: the world's largest operating tokamak, a joint Japan-EU machine, with first plasma already achieved.
• Kyoto Fusioneering: ranked by the IAEA among the most promising private fusion firms globally, supplying gyrotron heating systems to Tokamak Energy in the UK and others. Multiple Series C rounds (including extensions) have brought the company well above ¥20 billion (over $125 million) in cumulative funding, with strategic partnerships now spanning Canada and the U.S.
• National Institutes for Quantum Science and Technology (QST): supplying core components — superconducting coils, blankets, ITER's neutral beam injector — for the international project.
• National Institute for Fusion Science (NIFS) in Gifu: research on the stellarator alternative to tokamak designs.

Tucked between Tokyo and Mito, in a city called Naka, the country with a $10 billion EV problem is also building the most advanced fusion experimental device in Asia. The same engineering tradition. Just pointed at a different physics tier.

The Wager

Gasoline has won for 167 years inside the chemistry category. That's not stopping. e-fuels and solid-state batteries are going to share the road for decades, and that's fine. But the next leap — the actual one, the one that's 7.8 million times the energy density, where one Olympic swimming pool's worth of deuterium powers a city for a year — lives in fusion.

And on the other side of that bet sits a payoff list we used to file under "science fiction." Free electricity. Climate change running in reverse. Drinkable water for any country that has a coastline. Mars in two months. The moon as the world's next mining basin. Stanford-torus orbital colonies that don't need to live in line-of-sight of the sun. A species that finally has a backup planet.

Honda's mistake this quarter cost roughly 40 percent of ITER. Globally we keep finding ten-billion-dollar bills under sofa cushions to spend on strategic reversals, advertising budgets, ride-share subsidies, share buybacks, and AI training clusters whose products will mostly still need power from somewhere. The numbers required to actually take the moonshot are smaller than the numbers we're currently lighting on fire correcting strategy reversals.

So the unsubtle proposal: take some of that money, and bet it on the sun.

Honda's $10 billion mistake would cover almost 40 percent of ITER. How much is your country actually betting on fusion — and is it anywhere close to what it bets on the next-quarter version of cars? Whose children will grow up thinking of Mars as a place you can go?

References

• https://global.honda/jp/news/2026/c260312.html
• https://car.watch.impress.co.jp/docs/news/2108657.html
• https://www.nikkei.com/article/DGXZQOUB115RA0R10C26A5000000/
• https://www.enr.com/articles/62975-iter-magnet-milestone-tests-fusions-construction-supply-chain
• https://oilprice.com/Energy/Energy-General/The-Worlds-Biggest-Fusion-Reactor-Just-Hit-a-Milestone.html
• https://www.geekwire.com/2026/helion-makes-big-bet-on-tiny-merge-fusion-testbed-to-meet-aggressive-microsoft-timeline/
• https://thefusionreport.com/commonwealth-fusion-systems-a-hot-start-for-2026/
• https://kyotofusioneering.com/news/2026/01/29/3678
• https://www.fusionindustryassociation.org/over-2-5-billion-invested-in-fusion-industry-in-past-year/
• https://www.world-nuclear-news.org/articles/access-to-funding-remains-a-major-issue-for-fusion-says-fia
• https://gigazine.net/news/20250123-chinese-nuclear-fusion-reactor-record/
• https://www.qst.go.jp/site/jt60/5150.html
• https://www.euronews.com/next/2026/03/26/world-first-uk-startup-ignites-plasma-inside-nuclear-fusion-rocket-in-major-step-for-space
• https://pulsarfusion.com/sunbird-fusion-propulsion/
• https://www.cnn.com/science/nuclear-powered-rocket-pulsar-space-spc
• https://interestingengineering.com/science/mining-the-moon-for-helium3
• https://www.esa.int/Enabling_Support/Preparing_for_the_Future/Space_for_Earth/Energy/Helium-3_mining_on_the_lunar_surface