The History of Magnetic Confinement Fusion
A couple of weeks ago, The Fusion Report wrote an article titled “The History of Inertial Confinement Fusion: A Trip Down Memory Lane”, in which we explored the some of the seminal developments in Inertial Confinement Fusion (ICF), who were the key movers during ICF’s development, and where is the technology at today. Since we are not about “picking winners” at The Fusion Report (the market is far better at that, and their opinion is the only one that really matters anyway), it only seems fair that we write a similar article on Magnetic Confinement Fusion (MCF). And for hybrid fusion technologies (and the companies developing and commercializing those technologies), don’t worry – we will do a similar article on the history of hybrid fusion’s history as well in the month of May. Finally, for those who were wondering, the illustration above is of the International Thermonuclear Experimental Reactor (ITER) Tokamak being built in Toulouse, France by an international consortium. ITER, which is scheduled to go online in 2034, and begin “full plasma current operation” in 2036, and will start fusing deuterium and tritium (D-T) in 2039, is not a commercial fusion machine; rather it is a research facility to examine some of the techniques and technologies related to MCF.
Types of Magnetic Confinement Fusion (MCF) Machines and Technologies
Before we dive into the history of MCF, it would be worthwhile to list the various approaches to magnetic confinement. By far and away, most of the MCF machines today are based around toroidal machines, which are variations of “magnetic donuts” that contain, heat, and compress the plasma to achieve a fusion triple product, also known as the Lawson Criteria. The three “leading flavors” of toroidal fusion machines are:
Z-Pinch Machines: Z-pinch machines take a plasma flow, and “pinch” it by using magnetic fields to induce a Lorentz force on the plasma. This Lorentz force makes the electrically-conductive plasma squeeze towards itself, (hopefully) causing enough compression and heating to meet the Lawson Criteria and fuse. Because this force must be applied through a changing or alternating magnetic field, most Z-pinch machines use a large bank of capacitors and a spark gap (also known as a Marx Generator) to create the powerful magnetic “pinch force”.
The primary challenge that Z-pinch machines have had is what is known as “kink instability”, where the plasma hits the walls of the toroid. Interestingly enough, the first Z-pinch machines were utilized to generate high-power neutrons and X-rays for nuclear weapons research.
Stellarators: Stellarators were one of the earliest fusion machine designs, which attempted to overcome some of the issues with Z-pinch machines by using a ‘twisted” plasma path that pushed nuclei back to the inside of the toroid. One of the first stellarators was built at (what is now) the Princeton Plasma Physics Laboratory (PPPL). Unfortunately, circa-1960s stellarators were found to be leaking plasma beyond theoretical predictions.
Given Soviet advances on tokamaks in the late 1960s, most US researchers abandoned work on stellarators in favor of tokamaks. Recently, advances in high-tesla magnets have rekindled interest in stellarators, and there are currently ~10 stellarators in operation worldwide.
Tokamaks: Tokamaks are by far and away the most popular MCF machines today – there are estimated to be about 60 tokamaks currently in operation. Tokamaks combine a variety of different magnetic coils (toroidal and poloidal) into a doughnut-shaped vessel to achieve plasma stability, and which will hold the vacuum required for fusion to work.
The other important type of MCF machine is known as “Magnetic Mirrors”, a linear machine similar to an hourglass where plasmas at both ends of the linear vessel are accelerated towards each other. These magnetic mirror machines typically have a “pinch” or “waist”, formed either physically (making the vessel look like an hourglass) or magnetically (the fields become stronger in the center of the vessel), or both. The intention of magnetic mirrors is that the plasmas will meet at the “waist” of the hourglass; in a sense magnetic mirror machines utilize both magnetic confinement and inertia (the accelerated particles) to produce fusion, which is why they are sometimes considered hybrid fusion machines. Helion Energy’s Trenta and Polaris machines are much-advanced versions of magnetic mirror machines.
The Origins of MCF and the Tokamak Stampede
Tokamaks were the first MCF machines which showed promise as being viable for fusion energy. The first tokamaks, which were built by Soviet scientists in the 1950s (one of their earliest tokamaks is shown here).
The success of these early tokamaks, especially the Soviet T-3 tokamak at the I. V. Kurchatov Institute of Atomic Energy, was initially ignored in favor of stellarators. This changed in 1969 when a group of British scientists were invited by the Soviets to visit the T-3 tokamak site to validate the results. This resulted in a “tokamak stampede”, where large numbers of tokamaks were built in the early- and mid-1970s, and tokamaks became the preferred machines for MCF.
Scaling Up, the Race for Q=1, and The Birth of ITER and EAST
As the 1970s progressed, a number of tokamaks demonstrated the ability to achieve the conditions (temperature, plasma density, and containment time) required for fusion breakeven (“Q=1”), but not at the same time in a given reactor. At the same time, it became obvious that the deuterium-tritium (D-T) fuel combination was the way to go, as it required the least energy to achieve fusion in spite of its shortcoming of not being aneutronic (not outputting high-energy neutrons). There were two notable fusion machines that came out of these efforts.
The Joint European Torus (JET) in the UK began operation in 1983. JET, which was built at the Culham Centre for Fusion Energy in Oxfordshire UK, was the result of work of the member states of the European Atomic Energy Community (Euratom).
JET was built to utilize D-T fusion fuel and superconducting magnets, significantly increasing the strength of the magnetic fields that could be produced. While JET did not achieve break-even, largely due to new plasma instabilities which had not been seen at lower plasma densities and pressures, it did set a number of records. These included:
- The first controlled release of fusion energy in 1991
- Q=0.67 using 24MW of input energy to heat the D-T fuel in 1997, resulting in a power output of 16 megajoules (MJ) of fusion energy in a 5.2 second pulse
- The greatest fusion output energy (69MW) in 2024
The reactor was shut down after the last achievement in JET’s last experiment, operating for a total of 27 years.
The Tokamak Fusion Test Reactor (TFTR) in the U.S. began operation in 1982 (shown below).
The reactor, which was located at the PPPL and funded by the U.S. Department of Energy (DoE), was at the time of its construction the largest tokamak in the world, with a “plasma diameter” of 2.2 meters. Like JET, TFTR never achieved Q=1. However, it did achieve plasma temperatures of 510 million Kelvin in 1995 (a record at the time). While TFTR did not have the now-signature D-shaped plasma cross-section, a number of the experiments done on TFTR led specifically to the formulation of the D-shape as a way to avoid a number of the plasma instabilities seen at TFTR. It also pioneered the enhanced reversed shear mode of plasma confinement. TFTR continued operation until 1997 (15 years of operation), and was dismantled in September 2002.
Because of the issues encountered in both JET and TFTR, an international effort was undertaken to build an even-larger tokamak that incorporated the lessons learned in both JET and TFTR. This program, which was called ITER (International Thermonuclear Experimental Reactor), was to be funded and managed by China, the European Union, India, Japan, Russia, South Korea, and the United States (the UK remained part of ITER for a time after Brexit, but eventually left the program). Partner countries include Australia, Canada, Kazakhstan, and Thailand. The original goal for ITER was to build a fusion machine that could generate 500MW of thermal power from an electrical input of 320MW. ITER was designed as the largest fusion machine to be built to date, with an original budget of €6 billion (that budget has ballooned to between €18 billion and €22 billion, according to official estimates). Assembly of ITER began in 2020, with a goal of completion (“first plasma”) in 2025; this has now been pushed back to 2034.
Notably, ITER does not include superconducting magnets built of 2nd-generation High Temperature Superconductor (HTS) tapes, for its superconducting magnets, unlike follow-on commercial fusion machines. The illustration above from 2025 shows one of the ITER D-magnet sections being put into place.
At the same time ITER was being built, divisions within the ITER partner countries started to develop, particularly between China and the U.S.. These largely revolved around sharing specific fusion intellectual property. As a result, China built a superconducting tokamak named the Experimental Advanced Superconducting Tokamak (EAST) in Hefei, China. Originally a testbed for ITER technologies, EAST became China’s leading fusion research facility. The facility was commissioned in 2006 at a reported cost of $37 million (U.S.), well below the cost of ITER. EAST generated its first plasma in September 2006, with an electrical current of 200 kiloamperes (KA) that was contained for three seconds.
Over the next decade and a half, EAST would go on to break a number of records, including the first tokamak to sustain an H-mode plasma for over 1 minute at a temperature of ~50 million degrees Celsius in 2016, as well as achieving a 120 million degrees Celsius electron temperature for over 101 seconds in May 2021. EAST continues to operate today.
Venture-Funded MCF Machines Leading the Way on Commercialization
Around the late 2010s, something else happened in fusion – private venture capital funds started investing in fusion energy, a domain previously limited to governments. To be fair, the first private investments in fusion was into TAE Technologies (founded in 1998), General Fusion (founded in 2002), and Shine (founded in 2005). Also in the mid-2010s, Venture capital (VC) funds focused on fusion, clean energy, and similar technologies started popping up.
One of the first of these was Breakthrough Energy Ventures, which was founded in 2015 by Bill Gates. Other VC funds focused on “green energy” followed Breakthrough Energy Ventures, including Lowercarbon Capital, Khosla Ventures, and high-value individuals including Eric Schmidt, John Doerr, Sam Altman, and Jeff Bezos also began investing in fusion power. By 2020, over $1.5 billion (US) had been invested in fusion, which reached over $6.2 billion in 2025. The largest private investments were in Commonwealth Fusion Systems ($2.0 billion), TAE Technologies ($1.2 billion), and Helion ($1.0 billion).
For the nine fusion companies with investments over $100M, the majority (5 or 6, depending on how you count them) utilize MCF. These companies are focused on putting fusion power onto the grid by 2035 or sooner. For the good of our planet, let’s hope that “sooner” is when fusion energy becomes commercialized and is fulfilling at least some of our electricity needs.