Every energy source humanity uses is either burning something (coal, gas, oil), capturing something (wind, solar, hydro), or splitting something (nuclear fission). Fusion — the process that powers the sun — combines hydrogen atoms into helium, releasing enormous energy with no carbon emissions and minimal radioactive waste.
Fusion has been “30 years away” for fifty years. In 2022–2025, it got closer than ever. Here is what happened, what it means, and why you should pay attention now.
How fusion works (simply)
Fission (current nuclear): Split heavy atoms (uranium). Releases energy. Produces long-lived radioactive waste. Risk of meltdown.
Fusion: Combine light atoms (hydrogen isotopes: deuterium and tritium). Releases more energy per gram than fission. Produces helium and a neutron. Minimal waste (no long-lived radioactive material). No meltdown risk (reaction stops if containment fails).
The challenge: Fusion requires temperatures exceeding 100 million degrees Celsius — hotter than the sun’s core. At these temperatures, matter exists as plasma. Nothing on Earth can contain plasma except magnetic fields or inertia.
The fuel: Deuterium is abundant in seawater (one in every 5,000 water molecules). Tritium is rare but can be bred from lithium. Fuel supply is essentially unlimited.
The breakthrough timeline
December 2022 — NIF (National Ignition Facility), Lawrence Livermore, USA: For the first time, a fusion reaction produced more energy than the laser energy delivered to the fuel pellet. Net energy gain (Q > 1). Historic. The reaction lasted nanoseconds and the total energy was enough to boil a kettle — but the physics worked.
2023–2024 — Multiple milestones:
- NIF repeated ignition multiple times, improving yield
- China’s EAST tokamak sustained plasma for 1,066 seconds (18 minutes) — record duration
- South Korea’s KSTAR sustained 100-million-degree plasma for 48 seconds
- MIT’s SPARC project on track for 2026 net energy demonstration
2025–2026 — Commercial momentum:
- Over $7 billion in private fusion investment (2024 alone)
- 40+ private fusion companies globally
- First commercial contracts signed (Microsoft agreed to purchase fusion power from Helion Energy by 2028 — ambitious timeline)
The two main approaches
Magnetic confinement (tokamak)
How: Plasma contained in a donut-shaped magnetic field. Heat until fusion occurs.
Leaders:
- ITER (France) — the world’s largest tokamak, under construction since 2010. First plasma expected 2035. Full fusion expected 2039. International collaboration (35 nations). $25 billion project.
- SPARC (MIT/CFS) — compact tokamak using high-temperature superconducting magnets. Targeting net energy by 2026–2027. If successful, ARC (commercial prototype) follows.
- JET (UK) — record fusion energy output (59 megajoules, 2022). Proof of concept at scale.
Inertial confinement
How: Laser beams compress and heat a fuel pellet so rapidly that fusion occurs before the pellet disintegrates.
Leaders:
- NIF (USA) — achieved ignition 2022. Not designed for continuous power generation (pulsed approach — one shot per day currently).
- Laser Fusion (various startups) — attempting to increase pulse rate for commercial viability.
Alternative approaches
- Stellarator (Wendelstein 7-X, Germany) — twisted magnetic configuration, potentially more stable than tokamak
- Magnetized Target Fusion (General Fusion, Canada) — compress plasma with mechanical pistons
- Field-Reversed Configuration (Helion Energy, TAE Technologies) — simpler magnetic geometry, pulsed operation
- Muon-catalyzed fusion — exotic approach, early stage
Why this time might be different
1. Superconducting magnets — new high-temperature superconductors enable stronger magnetic fields in smaller, cheaper tokamaks (MIT’s SPARC is 1/65th the volume of ITER).
2. AI and simulation — machine learning optimizes plasma confinement in real time, solving instability problems that plagued earlier attempts.
3. Private investment — fusion is no longer only government-funded. Venture capital sees commercial potential and is funding aggressive timelines.
4. Material science — better materials for plasma-facing components that survive neutron bombardment.
5. Incremental progress compounding — 50 years of physics research reaching engineering readiness simultaneously.
The honest challenges
Timeline: Even optimistic projections place commercial fusion power plants in the 2030s–2040s. Your home will not run on fusion this decade.
Engineering scale: Achieving fusion in a lab is fundamentally different from operating a power plant that runs continuously for decades, feeding a grid, at competitive cost.
Tritium supply — tritium is rare and must be bred in the reactor. Breeding blankets are unproven at commercial scale.
Neutron damage — fusion produces high-energy neutrons that degrade reactor materials. Material lifetime limits economic viability.
Cost competition — solar and wind are already the cheapest energy sources. Fusion must compete not with coal but with renewables that improve annually.
Regulatory framework — no country has a fusion power plant licensing process. Building one requires creating the regulatory environment simultaneously.
Fusion vs. other clean energy
| Source | Status | Cost trend | Limitation |
|---|---|---|---|
| Solar | Mature, scaling | Falling rapidly | Intermittent, land use |
| Wind | Mature, scaling | Falling | Intermittent, location |
| Fission nuclear | Mature, stalled | High, stable | Waste, public fear, cost overruns |
| Fusion | Pre-commercial | Unknown | Not yet demonstrated at scale |
| Geothermal | Mature, limited | Stable | Location-dependent |
| Hydro | Mature | Stable | Ecosystem impact, limited sites |
Fusion’s role may not be replacing solar and wind but complementing them — providing baseload power when the sun is not shining and wind is not blowing, without carbon emissions or long-lived waste.
What to watch
- MIT SPARC results (2026–2027) — if net energy achieved in compact tokamak, the timeline accelerates dramatically
- Helion’s Microsoft contract (2028 target) — most aggressive commercial timeline; skepticism is warranted but worth monitoring
- ITER first plasma (2035) — the scientific gold standard, slow but authoritative
- China’s fusion program — investing heavily, results often underreported in Western media
Why fusion matters even if it is still decades away
Climate change requires decarbonization now — fusion does not solve the immediate crisis. Solar, wind, and battery storage are the present-tense solution.
But energy demand is growing — electrification of transport, heating, industry, and computing (AI data centers) will require multiples of current generation. Fusion offers the theoretical promise of unlimited, clean, baseload power — the energy source that could sustain civilization for millennia.
The sun has been doing fusion for 4.6 billion years. We have been trying to replicate it for 70. The gap is closing — not because of one breakthrough but because of thousands of incremental ones compounding into genuine momentum.
Fusion may still be 20 years away. But for the first time in those 20 years, the people closest to the physics believe it — and they have better evidence than ever before.
Lumen is edited by Leo Hartmann. Related: Solid-State Batteries · Climate Migration