In early 2014, Sam Altman made an influential visit to Helion Energy’s modest office in Redmond, Washington, arriving equipped with nuclear-fusion textbooks under his arm. At that time, Helion was primarily focused on research and development, navigating the complex challenges of harnessing fusion energy. According to Helion's CEO, David Kirtley, Altman persuaded the startup to adopt a more assertive strategy aimed at accelerating fusion deployment. By the following year, Altman had taken on the role of chairman and invested $9.5 million, marking one of the earliest significant private capital commitments to Helion. He later injected an additional $375 million in 2021, turning it into one of the largest private investments within his substantial personal portfolio.
Once predominantly directed by government agencies, nuclear fusion has attracted substantial private sector interest and financial backing, notably from those involved in the burgeoning field of artificial intelligence (AI). These investors are responding to the energy-intensive nature of AI development and the broader ambition to achieve artificial general intelligence (AGI) — AI that can perform wide-ranging intellectual tasks akin to human cognition. According to a Fusion for Energy report, published by a European Union body tracking the industry, total capital flowing into fusion companies surged dramatically, rising from $1.7 billion in 2020 to an estimated $15 billion by September 2025.
Altman, who has explicitly linked the future of AI advancement to breakthroughs in energy production, is among a cohort of prominent investors including SoftBank and Dustin Moskovitz — co-founder of Facebook and early backer of Anthropic — who have directed funds into Helion. Other fusion ventures have attracted significant tech company support, including Nvidia's investment in Commonwealth Fusion Systems (CFS), and Google's involvement with both CFS and TAE Technologies. Troy Carter, who oversees the fusion-energy division at Oak Ridge National Laboratory, underscores the motivation behind this surge: "AI is a big driver [due to] the energy needs ... to power their data centers."
Technological advancements coupled with abundant venture capital willing to embrace high-risk, high-reward projects have brought fusion's promise closer to reality. Several companies now project that commercial fusion power plants could come online within years instead of decades. Nevertheless, they face the fundamental challenge of validating their technology's functionality at scale. Successful fusion would revolutionize energy production by delivering clean, carbon-free electricity without the intermittency issues that hamper solar and wind power or the long-lasting radioactive waste concerns posed by nuclear fission.
Fusion energy emanates from the same process fueling the sun, producing power through the fusion of lightweight atomic nuclei rather than the fissioning of heavy atoms utilized in existing nuclear power plants. This process occurs within plasma — a highly energetic, electrically charged gas. Replicating the extreme conditions of a star’s core on Earth has long been recognized as a formidable engineering obstacle. Historically, even when fusion reactions were initiated in laboratory settings, they consumed more energy than they generated, failing to achieve what is known as scientific break-even energy.
However, a milestone was reached in 2022 when scientists at Lawrence Livermore National Laboratory successfully initiated a fusion reaction producing more energy than was used to heat the plasma, marking the first recorded instance of scientific break-even. Despite this breakthrough, no private enterprise has yet demonstrated this achievement.
Moving beyond scientific break-even, the next critical milestone is engineering break-even—where a fusion reactor produces enough energy not only to sustain the plasma but also to power the entire generation system and feed electricity into the grid. Helion Energy positions itself at the forefront of this race, aiming to deliver a commercial-grade fusion generator by 2028 at a site in Malaga, Washington, where construction commenced in July 2024. The company has substantiated its confidence by securing a contract to supply Microsoft with 50 megawatts of fusion power, with stipulated financial penalties should the project fail to meet its timeline.
Helion’s method diverges from conventional fusion designs that generate steam to turn turbines. Instead, it involves accelerating two plasma toroids to collision speeds approaching one million miles an hour. The resultant fusion event disrupts a magnetic field, inducing electrical power harvest without relying on thermal cycles to turn turbines. According to Kirtley, this technology currently recaptures approximately 96% of its input energy—comparable to regenerative braking systems in electric vehicles—placing it tantalizingly close to energy break-even. The company’s seventh-generation device, Polaris, was scheduled for its first test firing in late 2024, with plans to demonstrate engineering break-even in 2024. Kirtley, however, has withheld specific test results.
Encouraged by Altman’s push to accelerate scale and speed, Kirtley envisions Helion not only as a pioneer in fusion power plant construction but as a prolific manufacturer aiming to produce a fusion generator daily and deploy systems globally with rapid pace. Such ambitions underscore a growing belief that fusion energy may enter the market much sooner than previously expected.
This shift is echoed by Carter of Oak Ridge, who led a 2020 Department of Energy report forecasting a demonstration fusion pilot plant by the early 2040s. He now anticipates this milestone could arrive in the mid-2030s, crediting increased investment and AI-driven innovation. AI tools are facilitating progress by tackling complex problems in plasma physics and reactor design, says Nuno Loureiro, director of MIT’s Plasma Science and Fusion Center.
Even if Helion experiences delays, it would still likely secure several industry firsts. Moreover, alternative companies are racing to reach or surpass similar targets. Among them is California-based startup Pacific Fusion, which announced a $900 million funding round in 2024 with investors including former Google CEO Eric Schmidt and Microsoft AI CEO Mustafa Suleyman. Pacific Fusion aims to achieve engineering break-even through a methodology akin to the Lawrence Livermore device.
Commonwealth Fusion Systems, an MIT spinout founded in 2018 and recognized as the best-funded startup in the fusion space, pursues a different technical approach: a magnetic confinement system that maintains ultra-hot plasma within a magnetic bottle. Its pilot plant targets achieving scientific break-even by 2027, with parallel development of a commercial plant planned for early 2030s operation. Google has committed to purchasing 200 megawatts from CFS. Brandon Sorbom, CFS's co-founder and chief science officer, highlights the importance of backing from major tech companies, stating that such support signals suppliers that fusion energy development transcends isolated research efforts and aims at commercial-scale infrastructure. Notably, TIME’s co-chair and owner, Marc Benioff, also invests in CFS.
In contrast, OpenStar, a fusion startup based in New Zealand, demonstrated plasma generation in late 2024 on a modest $10 million budget, subsequently raising an additional $14 million. OpenStar's design inverts the magnetic bottle concept, positioning an exceptionally strong magnet at the reactor's core surrounded by plasma. Despite growing enthusiasm from multiple startups, Carter cautions about the field's vulnerabilities: a single, high-profile failure among notable companies could deter investors and impede broader progress.
The urgency for fusion power adoption is especially acute for companies like Google and Microsoft, which are building new data centers to sustain AI operations. Both companies have set ambitious environmental goals, with Microsoft aiming for carbon-negative status by 2030 and Google targeting net-zero emissions. AI data centers run continuously, making energy storage solutions critical since existing renewables like wind and solar are variable. Compounding these challenges, U.S. power generation capacity has remained largely stagnant since 2010, when China surpassed the United States as the world's top electricity producer. Fossil fuel resources may also face scaling difficulties amid increasing computational demands. Consequently, Altman and Nvidia's Jensen Huang identify energy availability as a principal bottleneck for AI advancement.
Beyond data center applications, successful fusion energy deployment could have transformative geopolitical and economic consequences. Fusion plants can be established close to areas of high energy demand, eliminating the dependency on locations favorable to solar or wind resources. The resulting abundant, carbon-free power could reshape global dynamics, reducing conflicts historically linked to energy scarcity. As Carter observes, "Most of our wars are fought over energy. If that’s no longer the driver, that changes things dramatically." Fusion energy thus stands poised to not only revolutionize how the world powers AI but potentially redefine international relations and economic development in the decades ahead.