A magnetic mirror device extracted electricity directly from plasma on June 19, proving that a historic design flaw can slash fusion plant costs by 10 to 20 percent. The tokamak-centric R&D pipeline now faces a hard question about its thermal-only dogma.
The oldest dead end in fusion just lit a light bulb, and every thermal-only reactor design suddenly looks too expensive.
On June 19, at the University of Wisconsin–Madison, a small team of engineers pulled multiple amps of current at roughly 100 volts from the Wisconsin HTS Axisymmetric Mirror (WHAM) device. They powered several light bulbs. WHAM does not burn deuterium-tritium fuel, so this was not fusion electricity. The converter harvested input power injected into the plasma, not alpha particles from a burn. But it was the first time a private company extracted electricity directly from a plasma machine, bypassing the steam turbine entirely.
The fusion industry has a $6 billion-plus R&D pipeline built on one assumption: a reactor’s only path to the grid runs through a thermal cycle. Realta Fusion just made that assumption optional. The magnetic mirror, an architecture dismissed for decades because it leaks charged particles like a sieve, now holds the first structural cost advantage in the sector.
The Amps That Came From the Leak
The mechanism is brutally simple. Realta’s direct energy converter (DEC) sits at one end of the mirror, where magnetic field lines open and plasma escapes. Charged particles streaming out hit an electrostatic field that slows them down, building electrical potential that drives current. The setup echoes the Venetian blind converter theorized by Richard Post at Lawrence Livermore National Laboratory and demonstrated in the 1970s. The physics sat on a shelf for fifty years.
Realta CEO Kieran Furlong put it plainly: “People have been talking a big game about DEC for years, so we just went out and did it.” Chief Scientific Officer Dr. Derek Sutherland added the necessary asterisk: this is not yet net-electricity or large-scale conversion of fusion power. The demo was small. It was also real, making Realta the first commercial fusion company to convert plasma kinetic energy into electricity on a working machine.
Realta is backed by Khosla Ventures and Future Ventures and was selected for the US Department of Energy’s Milestone-Based Fusion Development Program. That fact alone resets the competitive landscape.
The Half-Century Theory That Finally Got a Private-Scale Demo
Magnetic mirrors were the fusion field’s brightest dead end. The geometry is simple: a linear tube with magnetic coils pinched at both ends, reflecting particles back toward the center. The ends always leaked. Plasma escaped too fast for net fusion gain, and the architecture was largely abandoned by the 1980s in favor of tokamaks, which confine plasma in a closed toroidal loop.
Realta’s insight: stop fighting the leak. “We can consider the leakiness of magnetic mirrors as a compelling feature rather than a hurdle,” Sutherland told ANS Nuclear Newswire. Escaping particles carry kinetic energy. Instead of slamming the door, you put a turbine in the current—in this case, a direct converter that turns ion velocity into voltage without a thermal intermediary.
This is the first private-sector DEC demonstration on a functioning machine. The contrast with tokamaks is the key tension. A tokamak’s closed magnetic geometry traps charged particles by design. There is no convenient exhaust stream to tap. To extract electricity, a tokamak must thermalize everything: neutrons heat a blanket, which boils water, which spins a turbine. Maximum theoretical efficiency caps at roughly 45 percent. A direct converter can exceed 90 percent efficiency on the charged-particle output fraction.
A 10 to 20 Percent Cost Gap That Cannot Be Ignored
In Realta’s planned first-generation plants, targeted for the mid-2030s, 80 percent of fusion power would still flow through a thermal cycle at up to 45 percent efficiency. The remaining 20 percent would route through DEC at over 90 percent efficiency. That 20 percent slice does outsized work: it offsets all energy injected to start and sustain the plasma, raising net energy gain and lowering cost per kilowatt-hour by at least 10 to 20 percent.
This is not incremental. A 10 to 20 percent cost reduction at the busbar is the difference between a fusion plant that competes with natural gas and one that does not. The sector has spent decades chasing physics gain (Q). Realta just introduced an engineering gain (dollars per kilowatt-hour) that requires no breakthrough in confinement.
The mirror’s geometry enables it. A tokamak cannot easily retrofit DEC without a fundamental redesign of the divertor and exhaust handling, and no major tokamak developer has a DEC program on its public roadmap. Commonwealth Fusion Systems, TAE Technologies, and others have converged on a thermal-cycle architecture that locks in the steam turbine as the sole revenue path. Adding DEC now would mean re-opening reactor-core designs frozen years ago. The bottleneck is not physics. It is sunk cost.
Here Comes the R&D Reset
Realta’s demo establishes that DEC is engineerable on a real machine, not a laboratory curiosity. The next step is to scale voltage and current, then harvest alpha particles from an actual deuterium-tritium burn. The engineering path is visible. The consequences will unfold in sequence.
The financial forcing function is immediate. Any confinement scheme with open field lines—stellarators, field-reversed configurations—has a natural exhaust channel that could accommodate a converter. A 10 to 20 percent cost gap is fatal in a commodity power market. Within 18 months, expect at least two other private fusion firms to announce DEC experiments. They cannot afford not to try.
Policy and funding will follow. The Milestone Program, pressured by congressional interest in non-tokamak paths, is a near-certain candidate for a DEC-specific track by late 2027. The DOE has precedent for funding technology-agnostic advances as public goods, and a working private-sector demo with light bulbs is the kind of tangible milestone that unlocks appropriations.
Investor behavior will shift before policy does. Khosla Ventures has a history of doubling down on hardware companies that weaponize a physics insight into a unit-economic advantage. DEC is the first real cost differentiator in the sector. A follow-on round led by Khosla at a valuation exceeding every non-CFS private fusion company is a high-probability event. Expect a patent land grab around Venetian blind and traveling wave converter geometries in the next 12 months—the claims will be filed before the funding rounds close.
The second-order effect is a narrative shift. The industry’s question moves from “when does Q exceed one” to “when does cost per kilowatt-hour beat natural gas.” The third-order effect is architectural: magnetic mirrors, long dismissed as dead ends, re-enter the conversation as the dark horse that solves the wrong problem—leakage—better than tokamaks solve the right one—confinement.
The New Due Diligence Question
For fusion CEOs, the question is no longer “can you confine plasma.” It becomes “can you extract energy cheaper than a mirror.” A company with no DEC path is carrying a cost-structure assumption that may not survive the decade.
For venture investors, DEC capability is now a due diligence filter: if a startup’s reactor architecture has no natural exhaust stream, its levelized cost of electricity is permanently capped by the Carnot cycle. A competitor that can bolt on a 90-percent-efficient converter has a structural margin advantage that no amount of operational excellence can close.
For grid operators, the long-game implication is that fusion plants may have two revenue streams—thermal power sold into baseload markets and direct electricity dispatched on demand. That changes the capacity-credit calculation that determines whether a plant gets financed.
For policymakers, the takeaway is straightforward: DEC R&D is a public good. The physics was proven in the 1970s. The engineering was proven on June 19. Scaling it across multiple magnetic confinement schemes is classic public infrastructure, and the Milestone Program is the obvious vehicle.
The Feature That Was Always a Bug
The light bulbs at UW–Madison drew a few hundred watts. They proved nothing about fusion gain. They proved everything about cost. The magnetic mirror’s leakiness was always framed as a fatal flaw, a reason to abandon the architecture. Realta reframed it as a revenue stream.
The central question now is whether the tokamak-centric establishment can adapt fast enough to a competitor that turned a physics bug into a balance-sheet advantage. The mirror that bleeds plasma just started bleeding money from every thermal-only design on the drawing board.
