MAST Upgrade Fusion Campaign Hits Record Plasma Pressure Stably

A pressure record, with a catch

Britain's main fusion experiment has finished its biggest run so far. The UK Atomic Energy Authority (UKAEA) says MAST Upgrade, the spherical tokamak at its Culham campus in Oxfordshire, reached the highest plasma pressure in its history during its fifth experimental campaign. That campaign ran through 2025 and 2026 and produced more than 1,100 plasmas.12 The authority published the results on August 6, and the final count of plasmas was well above the roughly 950 pulses planned when the run was announced. More than 200 researchers from over 40 institutions took part.1

The headline phrase, a "practical nuclear fusion milestone achieved with higher plasma pressure," is accurate but easy to misread. Outlets covering the campaign agree on the main points: a pressure record for the machine, stable plasma control, four different ways to suppress damaging edge bursts, and a new real-time method for controlling where the plasma sits.212 Where they differ is how much weight the pressure figure should carry. One detailed write-up points out that UKAEA has not given a number for the record, so it cannot be compared with any other machine. The same piece notes that nobody claims MAST Upgrade produces more energy than it uses, and it was never built to.1 Another account states it more directly: the number that matters from the campaign is four, not the pressure.5

That is the right reading. The pressure record is real, but it only counts for this machine. The bigger result is that MAST Upgrade pushed harder while staying under control.

Why pressure matters, and why it's dangerous

Plasma pressure combines temperature and density. Fusion reactions climb steeply as both rise, so a higher-pressure plasma can produce more fusion power from the same volume. That makes high-pressure operation central to any design that hopes to sell electricity.122 The problem is that higher pressure makes a plasma more likely to become unstable.12

The instability at the centre of this campaign is the Edge Localised Mode, or ELM: a sudden burst at the outer edge of the plasma. UKAEA says a single ELM can throw out up to about a tenth of the plasma's stored energy, sending it into the surrounding components.12 Over time these bursts wear down the inner walls and exhaust hardware.2 A more severe event is a full disruption, where the plasma loses confinement completely, dumps its energy within milliseconds and drives large currents through the vessel.2 Accounts based on UKAEA's statements say the campaign recorded no disruptions at all, which, if it holds up, is the strongest stability claim in the coverage.25

The economics are simple. If fusion machines keep damaging their own walls, maintenance costs alone could stop them competing on price with other power sources.1

Four ways to calm the edge

The team tackled ELMs four ways on one machine in one campaign. Two methods, Quasi-Continuous Exhaust (QCE) mode and Resonant Magnetic Perturbations (RMP), build on earlier MAST Upgrade work with coils that create 3D magnetic fields. RMP holds edge pressure below the level where a burst would happen. QCE lets the edge leak energy steadily instead of storing it up for a large release.1

The other two, Quiescent H-mode and I-mode, are high-confinement operating states rather than suppression techniques. UKAEA describes them as holding energy in better while keeping the edge below the threshold for large bursts.12

James Harrison, UKAEA's head of MAST Upgrade science, framed the four regimes and the new position-control method as leading-edge science. He said international demand for the data shows the UK's central role in fusion research.12

Getting all four on one device matters because no one yet knows which edge-management approach will work best in a power plant. Showing several options under conditions relevant to future machines gives designers choices instead of one bet.12

Finding the plasma by its glow

The most original result may be the control technique. Researchers watched visible light from deuterium in the machine's upper and lower outer divertors, the exhaust regions above and below the plasma. Comparing the two let them detect small shifts in the plasma's position and correct them in real time. UKAEA calls it a world first.21

This is a practical step. A commercial plant cannot rely on scientists adjusting settings by hand, so automatic real-time control will be essential.12 Reading position from exhaust light is the kind of robust signal such a system would need.

The campaign also tested the exhaust problem directly. Adding small amounts of nitrogen at the plasma edge caused much of the exhaust power to leave as light, spreading the heat out before it reached the walls.212 That ties into MAST Upgrade's main purpose: developing a divertor, the exhaust system, that can survive the heat loads of a commercial-scale reactor.11

How it compares

Other "plasma pressure" headlines from the past year show why the MAST Upgrade figure can't be ranked. In late 2025, Zap Energy reported that its compact FuZE-3 device reached electron pressures of 830 megapascals, an estimated 1.6 gigapascals in total. Those pressures lasted about a microsecond and were measured by Thomson scattering.320 Zap called it the highest pressure ever recorded in a sheared-flow-stabilized Z pinch, and said the results were preliminary and awaiting journal publication.3

These results don't compete with each other. Zap's design trades confinement time for extreme compression. A tokamak like MAST Upgrade aims for steady, controlled plasmas lasting far longer.3 A record on one design says little about another. With no number published for MAST Upgrade, it can't even be placed on the same scale.1

The more relevant comparison is China's EAST tokamak. In January, researchers there reported reaching a long-predicted "density-free regime": stable plasma at densities well above the usual empirical limit, achieved by carefully managing how the plasma interacts with the reactor walls during start-up.1319 Reported line-averaged densities were about 1.3 to 1.65, against EAST's usual range of 0.8 to 1.0.19

Set side by side, the EAST and MAST Upgrade results point to one trend in tokamak research. Progress now comes less from raw performance and more from removing the instabilities that have capped performance for decades. EAST's team still has to show the approach works in high-confinement operation, which it plans to do next.13

Public labs, private race

The timing reflects a busier fusion field. Private fusion companies raised $4.5 billion in the year to July 2026, bringing total investment since 2021 to $14.2 billion across 56 companies, according to the Fusion Industry Association.11 Helion reported a plasma temperature of 150 million degrees Celsius and measurable deuterium-tritium fusion in February, though not net energy gain. Tokamak Energy reported a peak magnetic field of 13.7 tesla in its Demo4 magnet tests.11

UKAEA's role goes beyond its own machines. This week, General Fusion said its LM26 device heated electrons in compressed plasma above 1 keV, about 12 million degrees Celsius. The measurement was made with a Thomson scattering system built jointly with UKAEA, including a polychromator, a light-analysing instrument, made at Culham.84 UKAEA chief executive Dennis Whyte called it tangible progress across a range of private-sector approaches.4 General Fusion's own next targets are 10 keV and then the Lawson criterion, the combined conditions needed for net fusion energy, which it hopes to reach by the end of 2028.7

MAST Upgrade fits that pattern. It is a public research machine producing data and methods that others, including commercial developers, can use. UKAEA links the campaign to its 2026–2030 strategy as groundwork for commercial fusion.2 The machine is also getting more powerful in stages, with neutral beam heating being raised from 5 to 12.5 MW.11

The bottom line

Calling this a "breakthrough" overstates it, and calling it routine understates it. MAST Upgrade did not produce net energy, made no claim to, and gave no figure for its pressure record.1 It did push to its highest pressure while running more than 1,100 plasmas, reportedly without a disruption. It also showed four ways to control the edge bursts that damage reactor walls, and introduced a light-based position control method suited to automated plants.2512

The headlines focus on the pressure figure, but stability, exhaust handling and automated control are what will decide whether fusion becomes affordable. Those are the problems this campaign made progress on.