The Wellington-based startup OpenStar Technologies has made a breakthrough in sustainable nuclear fusion, as reported by Bloomberg.

The team successfully levitated a 500 kg magnet in a five-meter vacuum chamber filled with glowing gas heated to over one million degrees Celsius.

A select group of observers, including New Zealand Prime Minister Christopher Luxon, witnessed the event.

OpenStar's CEO Ratu Mataira and Prime Minister Luxon. Source: Bloomberg.

Currently, the reactor consumes more energy than it produces. However, the successful levitation is one of the first steps confirming the technology's viability.

CEO and founder of OpenStar, Ratu Mataira, believes that the simplicity of the design will give the company an edge in creating an efficient source of nuclear fusion energy. He stated that the levitation of the superconducting magnet validates the approach and its scalability.

“No one has a working fusion system capable of producing economically viable electricity. Starting with a simpler setup that can be scaled up and made cheaper is an attractive approach,” the physicist said.

The Race for Nuclear Fusion

About 50 companies worldwide are vying to be the first to learn how to collide atomic nuclei to produce cheap energy. OpenStar has attracted nearly $10 billion from various investors, including Bill Gates and Jeff Bezos.

“Nuclear fusion has the potential to revolutionize the energy sector by providing an unlimited source of safe and clean energy. What we witnessed makes it clear that this prospect is closer to reality than ever before,” Luxon emphasized.

The timeline for achieving this goal remains uncertain, with estimates ranging from 10 to 30 years.

Other countries have also reported similar breakthroughs. For instance, in 2022, California scientists were the first to generate more energy from a fusion reaction than was required to initiate it.

OpenStar believes it will take several generations of prototypes to create a system capable of powering an entire urban area.

How It Works

Nuclear fusion requires plasma—the fourth state of matter (the other three being solid, liquid, and gas). It is so hot that electrons are stripped from atoms, creating an ionized gas. Stars, lightning, and auroras are forms of plasma.

Within the Sun, heat and gravity compress plasma, holding it at the center. Under immense pressure, atoms fuse, releasing energy that powers the entire solar system.

One way to replicate this process on Earth is to use magnetic fields to contain plasma and initiate the fusion reaction.

In the 1950s, Soviet physicists developed the Tokamak—a promising concept in controlled nuclear fusion. This donut-shaped reactor uses powerful magnets surrounding a plasma chamber.

This design is employed in the multi-billion dollar international project, the International Thermonuclear Experimental Reactor, located in southern France. Its drawbacks include high costs and potential instability of the fourth state of matter.

“The Tokamak is more like a jet engine in terms of how it needs to be designed and how its performance is extracted. It heavily relies on complex modeling and high-precision manufacturing. The dipole, on the other hand, is more like a campfire. You roughly arrange the elements, add heat—and once the fire ignites, it sustains itself,” Mataira explained.

In 1987, Japanese theoretical physicist and engineer Akira Hasegawa proposed an alternative approach to plasma containment—placing a high-temperature superconducting magnet inside the plasma instead of outside. This design is known as a levitated dipole reactor.

In 2004, MIT and Columbia University successfully implemented this idea, but later research was halted due to funding shortages and technological limitations of the time.

OpenStar's Prospects

The company is currently preparing to launch a new prototype called Tahi, which is expected to be unveiled in two years. A third-generation model (Maui) is anticipated within five years, which will generate neutrons and become commercially viable.

The final stage will be the fourth-generation installation—Tama Nui. Its projected capacity will range from 50 to 200 MW, sufficient to power a small town or a large industrial facility.

Why This Matters

The rapid development of artificial intelligence is leading to exponential growth in electricity consumption. Morgan Stanley predicts a power deficit of 36 GW in the U.S. over the next three years.

This situation is already impacting consumers: rates are rising, and network overloads are causing power outages. Since the launch of ChatGPT, electricity prices in the U.S. have increased by 23%. Since 2020, the figure has risen by 40%, significantly outpacing the overall inflation rate in the country.

Analysts at The Kobeissi Letter consider the development of nuclear energy as one possible solution. Unlike solar and wind installations, nuclear power plants operate around the clock, which is essential for the continuous operation of AI. Additionally, it is one of the most cost-effective energy sources.

However, building nuclear power plants takes a long time. Currently, no major reactor is under construction in the U.S.

Nuclear fusion could address the electricity shortage, believes serial entrepreneur and Dataprana.io founder Arseniy Grusha.

“Yes, in theory, nuclear fusion is one of the most ‘ideal’ sources for training artificial intelligence and mining: stable base generation 24/7, high energy density, minimal emissions, and predictable power output (like nuclear energy),” he commented to ForkLog.

The expert noted that the problem of rising energy consumption will only worsen over time.

“There will only be growth: we estimate an increase in AI capacity of about 75 GW by 2030—five times. For comparison, Germany currently consumes 55 GW. Plus, we will need to service electric cars and robots. Global consumption will double over the next ten years,” Grusha pointed out.

He added that despite its appeal, nuclear fusion will not be a solution in the near future—it remains an early-stage technology.

“Even traditional nuclear generation takes decades to build. My realistic forecast for the mass application of fusion for data centers is around 15-25 years. Therefore, in the next 5-10 years, the growth of AI/HPC will be covered by gas, existing nuclear energy, renewable energy sources, and storage systems,” Grusha stated.

He emphasized that the network infrastructure—substations, transformers, and connection capacities—has become the main limiting factor today.

Is Space the Solution?

Some entrepreneurs believe that the future of data centers lies beyond Earth. They argue that the planet's energy networks are nearing their limits.

In January, Elon Musk announced that Tesla would revive work on Dojo3—a previously shelved project to create a third-generation chip for electric vehicles. Its new purpose will be for space computing.

Among the advantages are virtually unlimited access to solar energy and space for equipment placement. The downside is the high cost of launching rockets with the necessary infrastructure. However, analysts at 33FG calculated that AI computations in orbit will become economically viable by 2030.

Google was one of the first to take the initiative. The company announced a plan to create a network of near-Earth satellites to power data centers. This idea is also supported by OpenAI CEO Sam Altman, but Musk has a strategic advantage—control over delivery means.

Through the upcoming IPO of SpaceX, Musk aims to finance the launch of a constellation of computing satellites using Starship rockets. Once in orbit, these devices will be able to continuously collect solar energy due to constant illumination.

It is worth noting that in January, Alibaba Cloud's Qwen-3 became the world's first AI model uploaded and operating in orbit.