Note: This is a research note supplementing the book Unscarcity, now available for purchase. These notes expand on concepts from the main text. Start here or get the book.

The Fusion Revolution: The Fuel of Unscarcity

Summary: On December 5, 2022, humanity crossed a threshold that had eluded scientists for seven decades: scientific ignition. By late 2025, this physics breakthrough had evolved into a $10+ billion commercial race. This article outlines the roadmap from the NIF breakthrough to the deployment of commercial fusion energy, the “Fuel” engine of the Unscarcity framework.

The Breakthrough Moment: When We Lit Our Own Star

At 1:03 AM on December 5, 2022, a team of sleep-deprived scientists at the National Ignition Facility (NIF) in California achieved something that critics had called “always 30 years away” for… about 60 years: Scientific Breakeven.

192 lasers, the world’s most powerful, delivered 2.05 megajoules of energy to a fuel pellet the size of a peppercorn. The fusion reaction released 3.15 megajoules back. For the first time in human history, we got more energy out of a fusion reaction than we put in.

The Gain (Q): 1.54 (54% surplus). In fusion science, “Q” is the ratio of energy out to energy in. Q=1 means you get back exactly what you put in. Q=1.54 means you’re now producing more than you consume, the holy grail of fusion research.

The Significance: The physics works. Everything else is engineering.

If you’re thinking “but 54% doesn’t sound like much,” you’re missing the point. The first transistor in 1947 was objectively worse than vacuum tubes. The first solar panels in 1954 converted light to electricity at about 6% efficiency. What matters is crossing the threshold. Once you’ve proven the physics, you’ve entered the domain of engineering improvement curves.

And those curves have been brutal.

By April 2025, NIF had achieved 8.6 megajoules of fusion output, nearly tripling the original breakthrough in under three years. The target gain hit 4.13. That’s 313% more energy out than in. This is the signature of the 100x Future: once a physical threshold is crossed, improvement becomes exponential, not linear.

The Commercial Race: From Labs to Contracts (2026 Status)

The race has shifted from government laboratories to commercial deployment. By late 2025, over 50 private fusion companies had raised more than $10 billion in cumulative investment, a fivefold increase since 2021. The fusion industry raised $2.64 billion in private and public funding in the 12 months leading to July 2025 alone. The next edition of the same survey, the Fusion Industry Association’s 2026 report, counted a record $4.48 billion raised in the twelve months to July 2026 and $14.24 billion in total, across 56 companies employing more than 16,000 people.

The timeline has shifted from “decades away” to “deployment imminent.” And the contracts are signed.

1. Commonwealth Fusion Systems (CFS) - The Google Bet

The Technology: High-temperature superconducting (HTS) tokamaks. A tokamak is a donut-shaped chamber that uses powerful magnetic fields to contain plasma (super-heated gas) hot enough to fuse atoms. It’s the “classic” approach to fusion, but with magnets so powerful they can shrink the entire reactor to a fraction of traditional designs.

The Milestone: In June 2025, Google signed a 200 MW Power Purchase Agreement (PPA) with CFS for their first commercial reactor in Virginia. This wasn’t a press release. It was a legally binding contract for electricity from a fusion reactor that doesn’t exist yet.

Google, whose lawyers probably have lawyers, signed a deal to buy power from a technology that hasn’t reached net energy. That isn’t optimism. It’s due diligence pointing toward confidence.

Rick Needham, CFS’s Chief Commercial Officer, called it “the first true bilateral fusion PPA and the largest fusion deal in history - so far.”

Building SPARC. The first of 18 superconducting magnets went into SPARC around CES 2026, alongside a digital twin built with NVIDIA and Siemens to accelerate testing in simulation. By April the machine was past its hardest construction milestones — magnet installation and vacuum vessel assembly — and by May CFS disclosed SPARC was 75% complete, on track to be substantially finished by end of 2026.

Building the business around it. CFS announced ARC, its full commercial plant near Richmond, Virginia, designed for grid connection from day one, and formalized the SPARC-to-ARC playbook: a 400 MWe first-of-a-kind plant powering roughly 300,000 homes. The Virginia siting isn’t sentiment. CFS is working with PJM Interconnection, the grid operator for thirteen states and D.C., whose territory has the fastest-growing AI data center demand in the country and a queue of over 250 gigawatts of generation projects waiting on interconnection studies. Grid operators are starving for dispatchable baseload, and a firm fusion plant jumps the strategic queue. In a side bet, CFS is commercializing its HTS magnets independently — selling to particle accelerators, medical device makers, and industrial users — turning magnet R&D into a revenue stream that funds the reactor.

Another billion dollars (July 2026). On July 30, CFS raised a further $1 billion, its largest round since the $1.8 billion of 2021, bringing total funding to $4 billion (about $3.94 billion). The investor list is the tell: pension funds, sovereign wealth funds, and infrastructure and industrial corporate partners. That is not venture capital buying a lottery ticket; that is the asset class that funds toll roads and transmission lines deciding fusion has become an infrastructure investment. CEO Bob Mumgaard has signaled more raises ahead, and the capital is earmarked for the SPARC-to-ARC transition. By late August he put SPARC at “about 80% complete,” with the new capital meant to carry the demonstration reactor in Devens, Massachusetts all the way to the 2027 breakeven attempt.

The buying has started for the plant after that. On September 30, CFS placed what it calls the largest single order of HTS tape ever made, more than 10,000 kilometers from Japan’s Fujikura, for ARC’s magnets, and gave the Virginia site a name: the Fall Line Fusion Power Station. Fujikura is expanding its factories to fill the order. Mumgaard’s framing is that CFS is “actively placing orders” so that ARC can proceed “once our SPARC fusion machine achieves Q>1.” Nobody orders ten thousand kilometers of superconductor for a science fair.

The offtake side firmed up too. Italian energy major Eni committed to over $1 billion worth of electricity from ARC, joining Google — whose 200 MW represents roughly half of ARC’s total output. Between them, the plant is substantially pre-sold before it is built.

Timeline:

SPARC ~80% complete as of late August 2026 (75% in May); substantially complete by end of 2026
Scientific breakeven (Q>1) targeted for 2027
Commercial power to the grid via ARC reactor (400 MWe) in early 2030s, near Richmond, Virginia
Total funding: $4 billion (July 2026), with roughly half of ARC’s output sold to Google and over $1 billion of it to Eni
ARC supply chain: 10,000+ km of HTS magnet tape ordered from Fujikura (September 30, 2026) for the Fall Line Fusion Power Station

2. Helion Energy - Microsoft’s Bet

The Technology: Field-Reversed Configuration (FRC) with pulsed magnetic fusion. (FRC is an alternative to tokamaks: instead of a donut, imagine two magnetic bubbles colliding and fusing.) The technical innovation is that Helion captures electricity directly from the fusion reaction, skipping the steam turbine entirely. Most power plants, even nuclear ones, use heat to boil water into steam that spins turbines. Helion skips all of that. It’s like going from combustion engines to electric motors, eliminating an entire layer of inefficiency.

The Milestone: Helion’s seventh-generation prototype, Polaris, became operational in late 2024 in their 27,000-square-foot facility in Everett, Washington. It took three years to build, fast by fusion standards. By mid-2025, Polaris was creating the largest FRC plasmas the company had ever achieved.

The Deal: Microsoft’s 2023 binding agreement to buy 50 MW by 2028 remains the most aggressive commercial timeline in the industry, and as of September 2026 Helion itself is hedging it (see the caveat below).

Status: In July 2025, Helion broke ground on Orion, their first commercial fusion power plant, in Malaga, Washington. This is not a prototype or a demonstration facility. It’s a power plant designed to plug into the grid and deliver electricity to Microsoft.

In January 2025, Helion closed a $425 million Series F round, pushing their valuation to $5.245 billion. Investors included Sam Altman, Lightspeed, SoftBank, and Nucor, the steel company. (When a steel manufacturer bets on fusion, they’re not thinking about press releases. They’re thinking about energy costs.)

What Polaris has actually done. In January 2026 it became the first — and still the only — privately developed fusion machine to run on deuterium-tritium fuel, and it has since hit plasma temperatures of 150 million degrees Celsius. It now runs nearly every day, validating the direct-electricity-recovery approach that Orion will use at commercial scale. Alongside it, Helion is building its Omega manufacturing facility to produce critical fusion components at volume.

Why OpenAI came knocking. In March 2026, POWER Magazine reported OpenAI was in advanced talks to secure dedicated fusion power for its AI data centers. The logic is vertical integration plus political survival: AI data centers are already pushing residential electricity bills up 8-25% in some markets, and community backlash is mounting. Building your own power source beats competing with households for grid electricity, and beats explaining to voters why their bill went up so a model could train.

The OpenAI numbers (May 2026). The OpenAI–Helion talks crystallized into specific figures. OpenAI is negotiating to buy 5 gigawatts of fusion power from Helion by 2030, scaling to 50 gigawatts by 2035. Fifty gigawatts is roughly the electrical output of the entire UK nuclear fleet at peak, earmarked for one company’s AI training runs. To clear the conflict-of-interest, Sam Altman stepped down from Helion’s board on March 23, 2026. Construction on Helion’s Orion commercial plant has kicked off in Malaga, Washington, with Microsoft still on contract for first 50 MW in 2028. Polaris is now considered three-quarters of the way to the temperatures needed for commercial operation.

The capital caught up to the ambition (June 2026). Helion closed a $465 million Series G led by Thrive Capital, valuing the company at $15.5 billion — roughly triple the $5.245 billion it carried a year earlier — and bringing total committed capital to about $3.2 billion. In September the round grew to $500 million, with what CEO David Kirtley described as crossover, pension, and sovereign wealth investors. Backers now include SoftBank Vision Fund 2, KKR, BlackRock, Peter Thiel’s Mithril, and Reid Hoffman alongside Altman. The offtake list already ran past Microsoft: steelmaker Nucor agreed in 2023 to a 500 MW Helion plant at one of its mills, targeted for 2030, and put in $35 million of its own — a company whose largest variable cost is electricity betting real money on the machine.

The first license to run a fusion plant (June 2026). Helion cleared a regulatory milestone no fusion company had reached before. On June 16 it secured a Radioactive Materials License and a Radioactive Air Emissions License from the Washington State Department of Health for Orion - making it the first company anywhere to hold the regulatory licenses needed to operate a fusion power plant, and confirmation that it has the facilities, trained personnel, and safety programs in place. With Orion’s assembly and office buildings already complete, Helion can now break ground on the generator building itself. It sounds like paperwork, but it moves the bottleneck: the open question on fusion is shifting from can the physics work to can you get licensed to sell the power, and Helion is the first to answer the second one at the state level.

The company’s own caveat, which is worth more than its press releases. Through the summer Helion said it remained on schedule for first electrons to the grid in 2028, while its CEO called that milestone aggressive and hard. In September the hedge became specific. Axios reported that Helion had removed from its website the prediction that Orion will begin producing electricity by 2028; the page now says the plant will “begin initial operations in 2028” and does not define the phrase. Its public affairs director told a Boston conference, “We will not, in no universe, have 50 megawatts hit full operation in 2028,” and said the Microsoft contract requires Orion to “be connected to the grid and begin operations by the end of 2028.” Full power, by the company’s own account, comes in 2029 or 2030 at the earliest. The goal for Polaris was reworded too, from “net electricity” to “demonstrating electricity from fusion,” which is a lower bar.

Take the concession seriously. As of October 2026, no fusion company anywhere has delivered a watt of commercial electricity, and none is within twelve months of doing so. Everything in this section is a construction schedule, not a generation record.

3. TAE Technologies - The Shortcut

The Technology: Field-Reversed Configuration (FRC) with neutral beams (particle accelerators that heat the plasma). TAE’s approach uses hydrogen-boron fuel, which produces almost no neutron radiation, making it potentially cleaner and safer than other fusion approaches. (Neutron radiation is the main reason nuclear fission plants become radioactive; avoiding it means reactors could potentially be decommissioned without centuries of contamination.)

The Breakthrough: In April 2025, TAE published results in Nature Communications that stunned the fusion community. They demonstrated the first-ever successful formation of stable plasma using only neutral beam injection, a goal fusion scientists had pursued for over three decades.

Why does this matter? Traditional FRC machines need intricate plasma formation sections with quartz tubes and supersonic collisions. TAE’s new approach creates, heats, and stabilizes plasma directly in the center of the machine. They eliminated an entire section of the machine.

What it changes: The streamlined machine, called Norm, did not let TAE skip a generation, whatever some early coverage (this page included) suggested. The company’s own announcement keeps the sequence: Copernicus, the next prototype, is meant to show that net energy is viable before the end of the decade, followed by Da Vinci, the first prototype power plant, in the early 2030s. What Norm did, in CEO Michl Binderbauer’s words, is that TAE “has materially de-risked Copernicus,” with a design that is shorter, simpler, and cheaper to build.

In June 2025, TAE raised another $150 million from Google, Chevron, and other investors, bringing their total funding to over $1.3 billion.

The Trump Media merger. On December 18, 2025, Trump Media & Technology Group, the parent of Truth Social, and TAE announced an all-stock merger valuing the combination at more than $6 billion, with each company’s shareholders owning about half. Read the structure before the headline number. Six billion dollars is a share-swap valuation, and the cash that reaches TAE is $300 million: $200 million paid at signing and another $100 million released by the filing of the merger’s registration statement, which the companies submitted to the SEC on September 30, 2026. Closing, first promised for mid-2026, is now expected in the fourth quarter. TAE would become a subsidiary of the listed company, President Trump would hold a little over 20% of the merged firm, and the combined company says it will site and start building a first plant of about 50 MWe, to operate in the early 2030s.

An earlier version of this article called the deal a $6 billion investment that made TAE the best-capitalized fusion company on Earth. That was wrong. $300 million is under a quarter of the $1.3 billion TAE had already raised privately, and CFS has raised $4 billion. What the merger buys is a public listing, access to Trump Media’s balance sheet, and political proximity.

Why would a social media company merge with a fusion developer? The stated logic runs through AI: compute needs power the current grid can’t supply, and the merged company’s pitch is that fusion delivers “a decisive edge in the global race for A.I. dominance.” Whether this is shrewd vertical integration or elaborate branding, it makes TAE one of the first fusion companies an ordinary investor can buy.

The deal raises questions. TAE is arguably the furthest from the finish line among the top three fusion companies. CFS is ~80% through building SPARC, and Helion has broken ground on a commercial power plant, while TAE’s Copernicus has yet to be built and, as of its June update, its first plant was still in site selection. The company with the longest road took the most unusual route to capital.

There’s also a regulatory question. The Inflation Reduction Act and DOE programs are already channeling billions into clean energy. If TAE’s political connections translate into preferential treatment for permits or subsidies, that distorts the competitive landscape for every other fusion company. Could it also accelerate fusion-specific regulatory frameworks that everyone needs? Yes. The question is whether the rising tide lifts all boats or just the one with the biggest anchor in Washington.

Da Vinci development continues, targeting a first-of-kind prototype power plant in the early 2030s.

4. The DOE Roadmap: Government Catches Up

The U.S. Department of Energy released its Fusion Science & Technology Roadmap on October 16, 2025, and finalized it on June 9, 2026. The stated goal is fusion power on the grid by the mid-2030s. Not a research grant or a vague aspiration, but a roadmap with milestones.

This matters because the bottleneck has shifted. The physics is proven and the money is in. What was missing was regulatory clarity, and here the United States moved earlier than most people noticed. Fusion machines don’t produce long-lived radioactive waste and can’t melt down in the traditional sense, so in 2023 the Nuclear Regulatory Commission decided to license them like particle accelerators, under its rules for radioactive materials, and not like fission reactors. Congress wrote that choice into law in 2024, and in February 2026 the NRC issued the proposed rule. That is why Helion’s licenses for Orion came from Washington State’s health department and not from a federal reactor proceeding.

The roadmap formally acknowledges what private capital already figured out: over fifty companies, $14 billion raised, binding PPAs from Google and Microsoft. The government is clearing the path for something already coming, not dreaming about something that might.

But the timing creates a painful collision. The Labor Cliff hits the 2030-2035 window, roughly the same decade the DOE is targeting for fusion commercialization. The jobs disappear first, because AI is displacing workers now. Commercial fusion plants connected to the grid won’t arrive until 2035-2045 at earliest. That creates a gap of five to fifteen years where millions of people are economically displaced but the abundance technologies that could support them aren’t mature yet.

The fusion roadmap is the supply side. Nobody in government is publishing the demand-side roadmap: what happens to workers in the gap. That silence is the story.

The Unscarcity Implication: Why Fusion Changes Everything

So we’re building fancy power plants. Why does this matter for The Foundation and the Abundant Foundation?

Because fusion is the only energy source capable of powering a post-scarcity civilization. Not solar, not wind, not fission. Only fusion.

1. The Energy-Matter Conversion

Economists don’t like to talk about this: the cost of any physical good is the cost of the energy required to manipulate atoms.

Water scarcity is really energy scarcity. With cheap fusion, we can desalinate oceans at planetary scale. The water is there. We just can’t afford to filter it yet.
Food scarcity is really energy scarcity. Vertical farming works beautifully; it just uses a lot of electricity. With fusion, we can grow food anywhere, anytime, decoupling agriculture from geography, weather, and seasons.
Material scarcity is really energy scarcity. Recycling is expensive because breaking waste down to atomic components takes massive energy. With fusion, we can build true circular economies where nothing is ever “used up,” just temporarily borrowed and returned.

When energy becomes effectively free, the economics of everything change.

2. The Collapse of Marginal Cost

Solar and wind are cheap but intermittent. You can’t run a hospital on “mostly available” power. You can’t smelt aluminum when the sun sets.

Fusion provides baseload density: reliable, always-on power (unlike solar that stops at night), with millions of times more energy per kilogram of fuel than coal. A coffee cup of fusion fuel contains as much energy as 10,000 barrels of oil. The fuel is deuterium from seawater and lithium, and we have enough for billions of years.

The Result: Electricity becomes a utility like water or air, flat-rate or free for basic needs. This is the precondition for The Foundation, where survival needs are met unconditionally.

3. Powering the Cognitive Revolution

The AI revolution is an energy revolution in disguise.

U.S. data centers consumed 183 terawatt-hours of electricity in 2024, more than 4% of the country’s total electricity consumption, roughly equivalent to Pakistan’s annual demand. By 2030, global data center consumption is projected to hit 945 TWh, a doubling. In the United States alone, data centers will account for 8.6% of all electricity demand by 2035.

AI servers use up to 10 times the power of standard servers, and companies are deploying them at unprecedented scale. By 2028, more than half of data center electricity will go to AI alone, consuming as much power annually as 22% of all U.S. households.

Electricity bills are already rising. In some markets, data centers could increase average residential bills by 8-25% by 2030.

Without fusion, we face a choice between AI progress and climate goals. We simply cannot power the cognitive revolution with fossil fuels without catastrophic consequences, and renewables alone can’t scale fast enough while maintaining reliability.

Fusion is the “Fuel” that runs the “Brain” without cooking the planet.

Timeline to Abundance (2026-2040)

What’s coming:

2026-2027 (Demonstration Phase) (We are here)

CFS SPARC: ~80% complete (late August 2026), magnets installed, digital twin operational, ARC commercial plant in Virginia, HTS magnets being commercialized as side revenue stream
CFS funding: $4 billion total after a further $1B raise on July 30, 2026 from pension funds, sovereign wealth funds, and industrial partners; ARC output pre-sold to Google (200 MW, ~half the plant) and Eni (over $1B of electricity); 10,000+ km of HTS magnet tape ordered from Fujikura on September 30
Helion: Orion commercial plant under construction in Malaga WA (first-in-the-world operating licenses secured June 2026), Polaris hitting 150 million degrees on D-T fuel, Series G of $500M at a $15.5B valuation ($3.2B committed total), OpenAI talks: 5 GW by 2030 scaling to 50 GW by 2035 (Altman stepped off board March 23, 2026), a 2023 agreement with Nucor for a 500 MW plant; as of September 2026 the company says Orion will “begin initial operations” in 2028, with full power no earlier than 2029-2030
TAE: all-stock merger with Trump Media valued at $6B+ (announced December 2025, registration filed September 30, 2026, closing expected in Q4), worth $300M in cash to TAE; Copernicus is the next machine, with the Da Vinci prototype plant in the early 2030s
DOE Fusion Science & Technology Roadmap released October 2025 and finalized June 2026, targeting power on the grid by the mid-2030s
Scientific breakeven (Q > 1) targeted for 2027; zero commercial watts delivered anywhere as of October 2026

2028-2030 (Early Commercial Phase)

Helion’s Orion is contracted to connect to the grid and begin operations by the end of 2028, with full output in 2029-2030 at the earliest; Google’s ARC power follows in the early 2030s
Costs are high, but the learning curve begins
More PPAs signed as utilities recognize the inevitable
NIF-style inertial fusion potentially reaches 30+ MJ yields with planned upgrades

2030-2035 (Scaling Phase)

Factory production of compact fusion reactors begins
Multiple competing designs reach commercial viability
Costs plummet along typical energy technology learning curves
First developing-world deployments begin

2040+ (The Abundant Era)

Energy costs approach pure maintenance and distribution costs
The “energy constraint” on human civilization is lifted
Desalination, vertical farming, materials recycling become economically trivial
The Foundation becomes physically possible

The Hard Truth About Timelines

There’s an obvious objection: fusion has been “30 years away” since the 1950s. Why believe it’s different now?

Three reasons:

1. The physics is proven. NIF achieved ignition. Repeatedly. With improving yields. This isn’t theoretical anymore.

2. The money is unprecedented, and it changed character. About $15 billion by the Fusion Industry Association’s count ($14.24 billion through mid-2026, before CFS’s July billion), from hard-nosed investors who don’t fund fantasies. Google and Microsoft signing binding PPAs. Eni committing over $1 billion for ARC electricity. OpenAI negotiating a 5-GW deal scaling to 50 GW. A listed company merging itself into a fusion developer. The U.S. government publishing a commercialization roadmap. And in July 2026, pension funds and sovereign wealth funds putting a billion dollars into CFS — the capital that finances toll roads and transmission lines, not moonshots. When infrastructure money shows up, it means someone’s actuaries have run the numbers.

3. The alternative is unacceptable. The AI energy crisis isn’t hypothetical. The climate crisis isn’t hypothetical. The major players have done the math. They need fusion to work. That kind of institutional commitment changes everything.

Could timelines slip? Of course. Engineering is hard, and fusion is harder. But we’re not waiting for a physics miracle anymore. We’re waiting for engineering execution, and that’s a different kind of problem, one humans are very good at solving when motivated.

Conclusion: We’re Not Waiting for a Miracle

The story of fusion has been one of overpromising and underdelivering for seven decades. But something has changed. The physics is proven. The contracts are signed. The concrete is being poured.

Google, Microsoft, OpenAI, Eni, Chevron, and a constellation of pension funds, sovereign wealth funds, and industrial giants have put roughly $15 billion into fusion companies, and signed contracts for power that doesn’t exist yet. The DOE has published a commercialization roadmap. CFS is ~80% through building SPARC, has raised $4 billion, and is ordering magnet tape for the plant after it. Helion is pouring concrete on Orion in Malaga, Washington, holding the world’s first fusion operating licenses, and quietly moving its full-power date to 2029 or later. TAE is merging its way onto the stock market. These aren’t dreamers. They’re institutions that have seen something in the engineering data that convinced them the physics is solved and the timeline is real.

Keep the ledger honest, though: none of them has sold a single commercial watt, and none will this year. We are not waiting for a miracle. We are waiting for the concrete to dry.

When the first commercial fusion plant comes online, probably in the early 2030s, the era of fossil scarcity ends. And the era of Unscarcity begins.

The Fuel is being built. The question is no longer if, but when.

Fusion Energy 2024-2030: The $15B Commercial Race