When it comes to fusion power, there are two basic approaches: One, create a small star here on Earth that’s held in place by powerful magnetic fields. Two, use intense lasers to make a succession of even smaller stars, but repeat the process several times per second.

Moritz von der Linden, the co-founder and CEO of Marvel Fusion, likes his odds with the latter. In a world racing to wean itself off fossil fuels, fusion power promises to be an effectively limitless supply of energy, using widely available materials to re-create conditions that are hotter than the surface of the sun.

But most analysts believe commercial fusion power is still at least a decade away, and it’s up against renewable energy and battery storage, which continue to grow cheaper by the year. “Fusion has to come fast and it has to come cheap,” von der Linden told TechCrunch. “Otherwise nobody needs it, and nobody will be willing to pay for it.”

Marvel Fusion is one of several companies pursuing what’s known as inertial confinement fusion. It’s the same basic approach used at the National Ignition Facility (NIF), a Department of Energy lab that proved in 2022 that controlled fusion reactions could generate more power than it took to ignite them. That’s a helpful milestone for any startup chasing the thus-far-elusive technology.

But where NIF’s lasers are based on decades-old designs, Marvel is using cutting-edge technology to improve its lasers’ power and efficiency. The startup will soon build a demonstration facility in collaboration with Colorado State University, where it hopes two 100-Joule lasers will prove its core technology. Shovels hit the dirt on October 16, and von der Linden expects it’ll be operational by early 2027.

Those lasers will fire faster than the blink of an eye — in the femtosecond range, or one-billionth of second — bombarding a nanostructured target with photons that blast away its electrons and scatter the remaining positively charged ions. Those ions will then hit Marvel’s fuel, igniting a fusion reaction. Currently, Marvel Fusion is using a mix of mainly hydrogen and boron, though von der Linden says the company is taking a “mixed fuel” approach to keep its options open should a more advantageous combination come along.

Compared with NIF’s fuel pellet, which is ensconced in a one-centimeter gold hohlraum that takes two weeks to manufacture and load, Marvel’s fuel and target were designed for mass manufacturing. The fuel itself is solid at room temperature, making it simpler to handle than NIF’s fuel, which relies on either gaseous or cryogenically frozen hydrogen isotopes. Marvel’s target is simpler, too, with nanostructures made of silicon, not gold.

“That was kind of an awakening,” von der Linden said. “When the physics guys found out silicon works better, the target guys were like, ‘Hallelujah! We can use standard lithography from chip manufacturing.’” At the dimensions Marvel intends to manufacture, about 50 to 80 nanometers per feature, the company can use semiconductor manufacturing equipment that’s up to a decade old. It can produce around 5,000 targets on a standard 300-millimeter wafer.

If the Colorado experiments go as planned, the company will increase the lasers’ energy and multiply their number in another facility starting in 2028 or 2029. “With 20 lasers, we have the ability to really engineer the acceleration of the ions,” von der Linden said.

To hit those milestones, Marvel recently raised €62.8 million in a Series B round, the company exclusively told TechCrunch. HV Capital led the round with participation from b2venture, Bayern Kapital, Deutsche Telekom, SPRIND, and Tengelmann Ventures. The company was also selected by the European Innovation Council for a €2.5 million grant and up to €15 million in equity investment, which if made will be an extension of this round.

Marvel’s first prototype should be finalized around 2032 or 2033, von der Linden said, and it will contain hundreds of kilojoule-class lasers. Each will fire around 10 times per second.

That’ll be the moment of truth. While the company’s Colorado facility will be a useful milestone, “it’s like driving a Ferrari with a two-cylinder engine,” von der Linden said. “It will move, but it won’t do what it’s supposed to do,” which in Marvel’s case is generate useful amounts of power. If the full-scale prototype fires on all, er, lasers, then the startup has a chance of crossing the fusion finish line. The race is on.