Nuclear Fusion: The Key to Powering AI and Saving the Planet—If It Works

Sam Altman is certain that we'll harness the energy of the stars in the near future.

The wealthy CEO of OpenAI stated an interview with Bloomberg In January, they said "fusion will be successful" within the coming years—owing to Helion, a firm where Altman serves as both the chairperson of the board and a key investor.

If Altman’s view proves accurate, it signals the dawn of a new epoch. Nuclear fusion offers the potential for clean, affordable, plentiful, and consistent energy sources. As these facilities produce zero carbon emissions and can be situated virtually anywhere without significant constraints, robust fusion reactors have the capacity to revolutionize our planet indefinitely. In theory, such power could fuel high-energy demands from burgeoning sectors like generative artificial intelligence, crypto-mining operations, as well as support endeavors like space exploration, all while significantly aiding efforts against global warming.

However, if Altman’s forecast seems reminiscent, it’s due to his previous predictions along these lines not coming true. Back in 2022, he stated that Helion would "answer every question required for designing a commercially viable fusion reactor" by 2024. Additionally, Helion declared at the end of 2021 that it aimed to achieve “net energy production through nuclear fusion” following this schedule. Yet as we know, come 2024, there was still no word about significant progress from the company.

These alternating periods of ambitious promises followed by disillusioning setbacks have been recurring themes throughout history. Fusion power has remained an aspiration for many years, with researchers, governmental bodies, and enterprises globally striving towards this objective—but there is also a substantial record of such predictions not coming true as expected. In fact, there's a well-worn jest suggesting that fusion was always just three decades out, despite being forecasted consistently over the last six decades.

However, things might be different now. Recent scientific advancements indicate that novel methods for achieving fusion might prove successful – and an increasing number of start-ups are asserting that they can bring this technology to market more swiftly than previously thought. Nonetheless, Helion’s promised timeline Even compared to other early-stage fusion ventures, several firms assert they will achieve commercial fusion energy output within approximately ten years, available at prices that rival conventional electricity costs.

Investors are paying attention now. Over the past decade, the number of private fusion projects has grown threefold, with over $2 billion invested in this sector just within the last two years. This influx includes significant funds from well-known figures such as Peter Thiel via Mithril Capital, Bill Gates through Breakthrough Energy Ventures, Masayoshi Son at SoftBank, John Doerr who chairs Kleiner Perkins, and Khosla Ventures.

The participation of technology venture capitalists might signal that fusion is finally ready for prime time—or it could indicate that the backing from influential figures such as Altman is merely fueling a frenzy of exaggerated expectations. Following numerous previous disappointments, are these firms poised for a groundbreaking leap forward? Or does fusion remain just three decades out of reach?

Splitting atoms is simple; merging them remains tricky.

Nuclear fusion operates distinctly from nuclear fission, which is the type of reaction employed by every operational nuclear power plant today. During fission, an oversized and inherently unstable atomic core (such as uranium) breaks down, yielding substantial amounts of energy. This process occurs spontaneously on our planet without any input from humans. Interestingly, nature produced its own uncontrolled fission "reactor" approximately 1.7 billion years ago in what is now Africa; geological processes had accumulated sufficient quantities of uranium ore, leading to a spontaneous chain reaction.

Fusion does not follow this pattern. It occurs when two lightweight, stable atomic nuclei, such as those found in hydrogen, are compelled to combine into a single, even more stable nucleus, similar to that of helium. The process yields approximately four times more energy per unit mass compared to fission and roughly four million times more energy than what you would get from burning coal.

The difficulty lies in getting those lightweight atomic nuclei to fuse initially since they inherently resist combining. These atomic cores carry a positive charge, causing them to push away from one another. To surmount this repulsive force demands an immense amount of energy. In our solar system, this process occurs spontaneously solely within the Sun’s core, where the enormous gravitational compression resulting from the Sun’s mass—which exceeds Earth’s by more than 300,000 times—forces hydrogen nuclei close enough to merge.

Replicating those circumstances on Earth has posed a significant scientific hurdle for decades. Back in the 1930s, physicists found that achieving fusion on a minor scale wasn't overly complex, yet it was quite ineffective, yielding substantially less energy than what was expended. By the 1950s, the creation of hydrogen bombs demonstrated that extracting net energy from controlled fusion could be done; however, this came with uncontrollable and catastrophic consequences.

Ever since, the primary technique for generating viable fusion power has involved using an apparatus called a tokamak – essentially a toroidal vacuum vessel surrounded by strong magnetic fields designed to prevent the scorching plasma (an ionized gas) from coming into contact with the chamber’s sides. Despite significant efforts over almost seven decades, no tokamak has yet managed to produce more energy than what it takes to run it. Nonetheless, researchers have progressively built larger and more potent tokamaks throughout this period, each inching nearer to achieving breakeven, which brings us to ITER—the International Thermonuclear Experimental Reactor—a colossal machine spanning about 100 feet wide presently being assembled in Southern France.

The primary objective of ITER is to produce tenfold the amount of energy absorbed by the plasma. However, issues with the machinery have caused construction hold-ups, pushing back completion to roughly 2034. Although ITER aims to showcase net power generation through fusion, this will not involve producing electricity; instead, it serves as a test site for developing technologies required to construct an actual power facility. If current schedules remain unchanged, a practical fusion plant stemming from ITER studies would start operating no sooner than the early 2050s.

As the world anticipated ITER, an alternative approach to fusion emerged. The National Ignition Facility (NIF) became operational in 2009 at the Lawrence Livermore National Laboratory in Northern California. Unlike tokamak methods, NIF employs a distinct process for achieving fusion: It uses the world's strongest laser system to compress a small hydrogen-filled sphere for just a few nanoseconds, resulting in a surge of energy.

In contrast to ITER, the National Ignition Facility (NIF), using brief laser pulses, wasn’t designed as a prototype for a fusion power station; instead, it was created for "stockpile stewardship," conducting research aimed at understanding bomb physics without requiring actual nuclear tests. However, in 2022, NIF accomplished something that no tokamak had managed before: Its lasers triggered ignition —A self-sustaining fusion reaction spread throughout the small hydrogen sphere, momentarily generating more energy via fusion than what the lasers supplied to the target. This accomplishment at NIF sparked increased enthusiasm for fusion technology and attracted fresh streams of private funding towards fusion start-ups.

New approaches, aggressive timelines

We occasionally face criticism at plasma physics conferences," says Brandon Sorbom, Chief Science Officer at Commonwealth Fusion Systems. He’s referring to SPARC, the tokamak they're developing. "People argue that 'the design was overly cautious.'" However, he adds, "[Of course, SPARC will function based on its current scientific foundation.] Yet, ideally, a power plant should operate in a very steady and reliable manner.

Commonwealth is one of the biggest enterprises In the realm of private fusion initiatives, SPARC stands out as one of the most prominent. The team behind it anticipates completion by 2026 and believes firmly that this apparatus will generate substantially more power than it consumes. According to Sorbom, SPARC employs "precisely the identical physical principles" as ITER but leverages advancements from the past quarter-century to create an immensely compact version. Standing at approximately 24 feet wide, SPARC promises to be considerably less expensive and quicker to construct compared to larger projects like ITER.

In the meantime, several firms are pursuing methods for fusion akin to those used at NIF: they aim to squeeze fusion fuel tightly using rapid shocks to initiate ignition. Among these ventures, one of the latest and best-financed enterprises, Pacific Fusion, declared recently that it secured an inaugural financing round totaling $900 million from previous investors. Google CEO Eric Schmidt, Patrick Collisson, Reid Hoffman, along with other prominent figures from Silicon Valley.

Instead of employing lasers as done at NIF, Pacific aims to develop a method first introduced by another government-funded institution, the Sandia National Laboratory’s Z Machine. At this facility, massive amounts of electrical power are discharged in a brief pulse lasting under one microsecond, generating meg amps of current through a metallic cylinder enclosing the fusion material. The generated current induces an extremely powerful magnetic field which compresses the metal cylinder, resulting in pressures sufficient for initiating fusion reactions. Although this type of pulsed-power configuration hasn’t yet achieved net energy output, studies indicate that it could potentially do so in the future.

While the scientific principles underlying Pacific, Commonwealth, and similar ventures appear solid, their projected timetables raise questions. Many of these fusion firms pledge to deliver grid-connected power within approximately ten years or even sooner. However, numerous external specialists view such predictions as unduly optimistic. "I believe it will be challenging for anyone to generate electricity for the grid by then," states Ryan McBride, a nuclear engineering professor from the University of Michigan. "Since people aren’t keen on waiting for ITER, they’re attempting to accelerate this process through alternate approaches; should any prove successful, that would indeed be remarkable.... Yet it remains" not It hasn’t been demonstrated yet. The sole genuine fusion achievement that has actually been shown is the NIF ignition outcome."

Even with its achievements, NIF remains far from showcasing economically feasible fusion power. While NIF proved capable of extracting more energy through fusion reactions than the amount supplied to the target via lasers, the electricity drawn by these lasers from the grid is approximately 100 times greater than the output—significantly higher than the energy produced per fusion event. For a functional fusion reactor modelled after NIF to be practical, this issue needs resolution along with increasing the frequency of laser pulses considerably.

Power plants generally produce around 500 megawatts of power, give or take a factor of two. To achieve that figure, a power plant based on NIF would need to reliably implode a fuel target around once a second, every second, for days, months, and years on end. But in the three years since NIF first achieved ignition, it’s repeated the feat exactly five more times. “Sometimes you'll hear folks say that the science behind fusion is solved, it's all just an engineering problem. I disagree with that statement,” says Tammy Ma, one of the lead NIF scientists. “The science of fusion is not solved….NIF is the most successful fusion experiment to date, but we don't get ignition every time.”

Pacific Fusion encounters comparable hurdles with a technology that hasn't yet demonstrated its ability to attain net energy gain, although computer models indicate it might. Should their device manage to generate net energy, it’s quite possible this achievement would occur later than anticipated, pushing it beyond the projected timeframe into the early 2030s. "From our experience in fusion research, once we begin, achieving conditions close to optimal usually requires between 10 to 15 years," explains Alberto Loarte, who leads the Science Division at ITER. He adds, "Since fusion was first successful around 1960 up until now in 2020, things have remained largely consistent, making significant changes unlikely anytime soon." This lengthy development phase is not factored into the timeframes provided by companies like Pacific, Commonwealth, or others within today's cohort of emerging fusion enterprises.

The Commonwealth also faces its share of challenges. While it's accurate that the Commonwealth’s infrastructure mirrors ITER, the scale of ITER serves an important purpose. Both the Commonwealth’s SPARC system and ITER rely on superconducting magnets that must remain extremely cold, just slightly warmer than absolute zero. However, directly across from these frigid magnets within the tokamak chamber lies a scorching-hot 100 million-degree plasma. In SPARC, which offers limited space compared to ITER, the interior walls of the tokamak will likely experience temperatures up to five to six times higher than those found in ITER. Consequently, SPARC can only operate for around 30-second intervals before excessive heat threatens to disable its magnetic components. As stated by Loarte: "Even if they achieve fusion power generation under such conditions, extending operation beyond short bursts remains unresolved." This implies achieving sustained reactions leading toward practical reactors still poses significant hurdles.

The neutron challenge

The Pacific, Commonwealth, along with almost every other fusion company, encounters yet another significant challenge. They utilise the same kind of hydrogen fuel as what was used at NIF for their successful fusion experiments: a mixture known as deuterium-tritium, or "DT." While deuterium can be readily obtained, generating tritium—a necessary component—poses an additional hurdle for a future fusion power facility. Approximately 75% of the output from DT fusion manifests as neutrons; these are uncharged subatomic entities making it complicated to harness their energy effectively. Over seven decades of research and development into fusion technology, no one has managed to produce even a small quantity of electrical power from this process. When I say 'a little,' I mean not considering"net" electricity generation, merely some level of electric power production. any Electricity production," states plasma physicist Daniel Jassby, "has always eluded us as we've never managed to transform neutron bombardments into electrical energy.

To address this issue, a functional DT fusion power plant requires a component called a "blanket." This blanket absorbs neutrons, transforming their energy into heat before transferring that heat away from the reactor to spin a turbine and generate electricity. However, constructing such a blanket at scale remains uncharted territory; current prototypes like those planned for demonstration by firms including Commonwealth and Pacific within the coming half-decade—such as SPARC—are merely meant to prove that these companies' respective approaches to fusion can produce more energy than they consume. They project having fully equipped reactors incorporating blankets approximately ten years down the line, though building a blanket entails significant technical hurdles due to its complexity. Various design proposals exist, yet they universally entail encasing the fusion chamber in advanced cooling systems, frequently involving costly materials such as liquid lithium salts or lead, both presenting considerable technological challenges.

Neutrons pose additional challenges as they exit the reactor. Nuclear physicist Paul Springer explains, "The harm caused by neutron radiation is considerable; you cannot afford for critical structural elements like walls to suffer from this kind of degradation. It compromises their stability." He adds, "These structures turn brittle and spongy over time due to irradiation effects." Using unsuitable materials could render an otherwise efficient fusion reactor unreliable since frequent component replacements would be necessary owing to neutron-induced wear and tear. Solving this issue necessitates extensive trials demonstrating durability against constant neutron bombardment within reactors—a capability which current facilities do not offer.

The list of issues continues. Creating and providing a fresh objective each second poses significant challenges for pulsed power systems. The complexity of the design complicates upkeep for tokamak devices. Additionally, handling tritium, establishing supply networks, and educating the required personnel present obstacles for nearly every fusion enterprise.

The extensive array of hurdles casts doubt on the ambitious 10-year projections set forth by these startup ventures aiming for commercial fusion. Upon voicing my doubts, representatives from these firms admitted that the magnitude of difficulties is indeed intimidating. However, they maintained confidence in overcoming them. "Our history shows us that achieving this timeframe is feasible with sufficient resources," Sorbom stated in a written communication facilitated through Commonwealth, highlighting the swift development of the initial prototype element for their tokamak’s magnetic system. He added, “Fundamentally, both scientifically and based on foundational principles, there isn’t anything preventing commercial fusion energy from being realised within ten years.” Will Regan, who serves as President and co-founder at Pacific Fusion, proposed that advancements in pulsed power systems underpinning their method make quicker progress attainable: “Technologies associated with pulsers are currently rendering nuclear fusion not only more cost-effective but also easier to manage, implement, and expand,” he commented. vtrik via email.

Meanwhile, another firm is aiming even higher with plans beyond competitively priced fusion energy in the 2030s: Helion. The company has secured over $1 billion in funding, including more than $375 million contributed by Sam Altman; notable backers also include Dustin Moskovitz, Peter Thiel, and SoftBank. Although Helion missed its target of demonstrating net electrical output from fusion by 2024, they assert that they remain committed to fulfilling their obligation under an agreement made in 2023 to provide fusion power. Microsoft By 2028, at competitive market rates. In response to inquiries about why this deadline was considered more reliable compared to earlier projections, David Kirtley, Helion’s CEO and co-founder, stated, "We're undertaking an immensely complicated task that has never been accomplished before."

This appears to add another cause for doubt regarding their schedule, instead of fostering trust. However, Kirtley’s comment holds merit as it aligns with reality: Helion has not disclosed extensive details about its fusion method; yet what little they’ve shared suggests that their technology lacks a substantial background in research and testing. The firm opts out of utilizing DT fuel and asserts it won’t require a blanket since it plans to generate power directly from the expansion of its plasma. According to Helion, their "system recovers all unused and novel electromagnetic energy effectively" from the plasma without needing to first capture neutrons.

Since Helion has not released specifics about this system, no experts felt comfortable commenting on whether their approach seems feasible. However, considering the hurdles faced by traditional methods for achieving fusion, it’s tough to imagine that Helion’s tech will meet their stated timeframe. Fortunately, we shouldn’t have too much longer to speculate before getting answers.

Calibrating expectations

Even with all these hurdles, commercial fusion power could realistically be achieved within the coming decades. Although the scientific and technical obstacles are significant, there isn’t much cause to believe they cannot ultimately be overcome. Numerous effective solutions have already been suggested for most of these issues.

Nevertheless, almost all these solutions remain purely hypothetical, and developing them into prototypes—or even more so, commercially viable products—would demand considerable research efforts. "Should we secure sufficient funds, with global support—not limited to the US—I believe we could make significant progress. world He said, "This poses an existential risk. We require fusion technology. Everyone needs to get involved; let’s adopt a Manhattan Project or Apollo Program approach, truly focusing our efforts," notes Ma, the NIF scientist. He believes that accelerating fusion energy integration into the power grid might be possible. However, without such acceleration and considering past funding trends, he acknowledges it may exceed a decade-long timeline."

In the interim, as we await the advent of fusion power, there's a risk that the deceptive allure of imminent fusion might be presented as a cure-all solution. This could serve as justification for neglecting quicker paths towards reducing carbon emissions and instead encourage increased energy consumption at this very moment. Altman, Who else invests in nuclear energy aside from Helion? , he had already indicated this—during his January interview, he stated that "rapidly approving fusion reactors" was the most effective method for addressing climate objectives without impeding the progress of AI firms.

Advancements in future technology might further develop this storyline. There's a strong possibility—perhaps even likelihood—that companies such as Commonwealth or Pacific could achieve net power generation through fusion within the next five to ten years. However, reaching the point where fusion becomes commercially viable and cost-effective for widespread use remains quite distant. Without additional research solving numerous scientific and engineering challenges, we cannot rely solely on fusion power emerging in time to mitigate climate change, despite significant increases in funding over present amounts—it would be perilous to consider otherwise. As stated by Loarte: "There are individuals who wish to believe... that fusion holds the potential to become an energy supply capable of replacing other forms of energy production, including nuclear, thermal, coal, or natural gas, all within the span of one decade." He adds, "I do not find this outlook practical."

We’re all supporting each other," states Ma. "I would be thrilled if any of these fusion companies achieve their goals within five years, ten years—whichever timeframe they set…. However, at the same time, the enormity of the challenges ahead indicates that it will require substantial effort and more time to overcome them.

Adam Becker is a science journalist holding a PhD in physics. Recently, he authored More Infinity Always: AI Sovereigns, Cosmic Kingdoms, and Silicon Valley's Quest to Shape Humankind’s Destiny .

This tale was initially showcased on vtrik