Fusion power’s biggest question remains unanswered: how do you ensure the cost to start the fusion reaction isn’t higher than the price at which you can sell the power?
Plenty of people have ideas, but no one has cracked it yet. Commonwealth Fusion Systems, for example, is confident enough that it’s building a massive reactor that costs several hundred million dollars. But the device won’t be turned on until next year, leaving the question unanswered for now.
Other companies that were founded more recently think they have a shot at building a fusion power plant for less, including Pacific Fusion, which today announced the results of a series of experiments it performed at Sandia National Laboratory that it says will eliminate some costly parts of its approach. The company exclusively shared the results with TechCrunch.
Fusion power promises to generate large amounts of electricity 24/7 and deliver it in a way that’s familiar to today’s grid operators. Most fusion startups are targeting the early to mid-2030s to switch on their first commercial fusion power plant.
Pacific Fusion is chasing an approach known as pulser-driven inertial confinement fusion (ICF). At its core, it’s similar to the experiments carried out at the National Ignition Facility (NIF). The company compresses small fuel pellets in rapid succession, and that compression causes atoms inside the fuel to fuse and release energy.
But where NIF uses lasers to kick off the compression, Pacific Fusion wants to use massive pulses of electricity. Those pulses will create a magnetic field that encircles the fuel pellet — about the size of a pencil eraser — causing it to compress in less than 100 billionths of a second.
“The faster you can implode it, the hotter it’ll get,” Keith LeChien, co-founder and CTO of Pacific Fusion, told TechCrunch.
Techcrunch event
Boston, MA
|
June 23, 2026
One of the challenges with pulser-driven ICF is that the process has typically needed a bit of a kickstart to work properly. To create conditions in the fuel pellet hot enough for fusion, researchers have been using both lasers and magnets to warm it up beforehand. “It’s just a little bit of energy just to give it a little bit of a boost before you compress it,” LeChien said, on the order of 5% to 10% of the total energy.
But the added lasers and magnets add upfront complexity, cost, and maintenance requirements to the machine, making it that much harder to sell power at competitive prices.
So in the experiments at Sandia, Pacific Fusion tweaked the design of the cylinder encasing the fuel pellet and adjusted the electrical current delivered to it. Before the big pulse of electricity that ignites the fusion reaction, the company allowed a bit of the magnetic field to leak through to the fuel before compressing it, warming it in the process.
“We can make very subtle changes to how this cylinder is manufactured that allow the magnetic field to leak or to seep into the fuel before it’s compressed,” LeChien said.
Pacific Fusion’s fuel is loaded in a plastic target that’s wrapped in aluminum. By varying the thickness of the aluminum, the company can adjust how much of the magnetic field makes its way to the fuel. The casing needs to be manufactured with some precision, but nothing crazy, LeChien said — something on the order of what’s required for a .22 caliber bullet casing. “That’s a process that’s been honed and manufactured and perfected over 100 plus years,” he added.
The tweaks don’t significantly change how much energy Pacific Fusion needs to deliver to the target. “It doesn’t take much energy to actually allow that magnetic field into the center of the fuel,” he said. “It’s a tiny fraction, much less than 1%. It’s a very, very, very small fraction of the overall energy in the system, so it’s effectively unnoticeable.”
Eliminating the magnetic system would simplify the system and its maintenance requirements, which would have a modest effect on overall cost, he said. But getting rid of the laser would cut costs significantly. “The scale of laser [needed] to preheat these types of systems at high gain is north of $100 million.”
LeChien said experiments like this also help refine the company’s simulations to ensure they’re in agreement with what happens in the real world. “A lot of people have simulated things and said, ‘Oh, this will work or that will work,’” he said. “It’s a very different game to simulate something, build it, test it, and have it work. Closing that loop is hard.”