← Energy Energy · Series

Fusion 101: A Plain-English Guide to Fusion Energy

Updated October 9, 2026

Fusion is the process that powers the Sun: light atomic nuclei, usually the hydrogen isotopes deuterium and tritium, are forced together and release energy. Making it work on Earth means creating a plasma that is hot enough, dense enough and held together long enough. This series explains the field in plain English and ties our news coverage to a stable reference, so a headline like "12.6 million degrees" can be read for what it does and doesn't show.

Reading Order

1. Start here: the primer below.
2. The tokamak bet: SPARC and the race to net energy
3. The compression bet: General Fusion's 1.1 keV milestone
4. Where AI fits: a digital twin of SPARC

Zooming out to the power system fusion would plug into: nuclear power and the AI boom and our guide to small modular reactors.

The Three Numbers That Matter

A fusion plasma has to clear three hurdles at once, and progress on one does not guarantee progress on the others.

• Temperature. Plasma temperatures are usually quoted in keV, thousands of electron-volts. One keV is about 11.6 million degrees Celsius. General Fusion's stated next target is 10 keV, which it describes as roughly 100 million degrees.
• Density. How many fuel particles are packed into the plasma, since more of them means more collisions.
• Confinement time. How long the plasma holds its heat before losing it.

Together these make up the Lawson criterion, the threshold for a plasma to give out more fusion energy than it needs to keep going. A plasma that is hot but thin, or hot but short-lived, does not meet it.

What "Breakeven" Means

Headlines use "breakeven" loosely. Physicists use several distinct terms, summarized here from the standard definitions of the fusion energy gain factor, known as Q:

• Scientific breakeven (Q = 1): fusion power equals the heating power put into the plasma. For laser systems, the count is the driver energy that reaches the fuel.
• Ignition: the fusion reactions' own heating supplies all the heat the plasma needs, so outside heating can drop to zero.
• Engineering breakeven: the plant recovers enough electricity, after conversion losses, to power its own heating systems. This is the one that matters for a power plant.

The best-known example is the U.S. National Ignition Facility. On Dec. 5, 2022 it delivered 2.05 megajoules of laser energy to a target and got 3.15 megajoules of fusion energy out, a target gain of about 1.5. The facility's lasers, however, drew more than 400 megajoules of electricity to do it, so it was scientific breakeven, not a net-electricity result. Wikipedia lists a later record of 8.6 megajoules in 2025, a figure we have not checked against a primary source.

Three Ways to Build It

• Magnetic confinement. Powerful magnets hold the plasma in a ring-shaped chamber, most often a tokamak. Commonwealth Fusion Systems' SPARC and the international ITER project are tokamaks. Our milestones guide covers SPARC.
• Compression. A plasma is squeezed quickly to raise its temperature and density. General Fusion's Magnetized Target Fusion compresses a magnetized plasma mechanically over milliseconds. See the LM26 article.
• Laser inertial fusion. Powerful lasers blast a tiny fuel capsule so it implodes. This is the NIF approach.

Scoreboard, as of Oct. 9, 2026

Each line shows the latest figure we have verified, with the source in the "Go Deeper" list. Company targets are labeled as targets.

• Commonwealth Fusion Systems, SPARC (tokamak): about 75% built as of April 2026, with a company target of net energy in 2027. In October 2026 DOE picked CFS to lead an AI digital twin project for the machine, covered in our report.
• General Fusion, LM26 (compression): 1.1 keV electron heating reported Oct. 8, 2026, up from 0.72 keV in June. The paper is submitted for peer review. Next target 10 keV; first-of-a-kind plant around 2035, which is a company target.
• National Ignition Facility (laser): ignition on Dec. 5, 2022, as described above.
• ITER (tokamak, international): under its 2024 baseline, operation starts in 2034 with a deuterium-only plasma and deuterium-tritium operation in 2039, at an extra cost of about 5 billion euros.

How to Read a Fusion Headline

This section is analysis, not reporting. Five questions sort most announcements:

• Which temperature? Electron and ion temperatures differ, and ions are what fuse.
• Which Q? Scientific breakeven, ignition and engineering breakeven are different claims.
• Is it peer reviewed? A paper posted on a company site is not the same as a published one.
• Is it a result or a target? "Aims to operate a plant by 2035" is a goal, not an achievement.
• For how long? A burst of milliseconds is different from steady operation.

We apply this checklist in each article in the series, and we update this page when milestones change.