A defining challenge for fusion energy is the scarcity of tritium, the key fuel for most near‑term D‑T fusion concepts. Global tritium inventories are extremely limited, and current production rates fall far short of what would be required for large‑scale deployment of fusion power plants. As a result, tritium availability remains one of the most significant bottlenecks for the realization of commercial fusion energy.

New fusion startups change the landscape of fusion development and tritium demand. While the classical large research reactors and DEMO‑class conceptual designs demand significant quantities, the startups show a spectrum of approaches with very different tritium requirements. Startup‑scale D‑T concepts require significantly reduced inventories. And at the frontier, several advanced concepts aim to eliminate tritium entirely—although these are also the most technically challenging to achieve.

The figure compares the tritium inventory for all fusion concepts under consideration here. Since exact figures are in general not available, general technical considerations were used to estimate a lower and a higher end. The purpose of these crude estimates is to illustrate the fundamental differences between different fusion concepts. These fall into three categories.

1. Major Research Installations & Conceptual Reactors (Highest Tritium Demand)

These devices illustrate the scale of tritium required for high‑power D‑T fusion and why tritium supply is a bottleneck.

Systems and Fusion Concepts

The international ITER project in France is a magnetic confinement tokamak. The latest update foresees the start of the first Deuterium-Tritium (DT-1) phase in September 2039. It will burn the D-T plasma with a 500 MW fusion power goal. Tritium breeding will only be done on a small experimental scale. The tritium inventory needs to be supplied from outside. In contrast the European DEMO tokamak reactor will be equipped with full tritium breeding blankets and is planned to begin operation in the 2040s-2050s. China is already operating the CFETR tokamak. It will have full tritium breeding blankets in the mid-2030s. Several other countries are planning national DEMO reactors. These are India, Japan, Korea, the UK, and the USA. The Russian Federation is planning FNS-R as a large fusion neutron source using the D-T fusion reaction. These reactors will require multi‑kilogram tritium inventories and have even higher throughput. At DEMO scale this will be an annual tritium consumption of around one hundred kilograms. Their fuel‑cycle requirements will dominate global tritium demand. Therefore, the goal is to make them self-sufficient by producing at least as much tritium as they lose.

2. Tritium‑Fueled Startup and Commercial Concepts (Moderate Tritium Demand)

These designs use the D‑T reaction but operate with significantly smaller tritium inventories than large DEMO scale reactors. Many incorporate breeding blankets or rely on small on‑site tritium stores.

Systems and Fusion Concepts

• First Light Fusion (FLF) (UK) – Projectile‑driven inertial fusion (D‑T, early 2030s)
• General Fusion (GF) (Canada/UK) – Magnetized target fusion (D‑T, early 2030s)
• Focused Energy (FE) (Germany/USA) – Laser inertial fusion (D‑T, 2030s)
• Renaissance Fusion (RF) (France) – Stellarator (D‑T, 2030s)
• Commonwealth Fusion Systems ARC (CFS) (USA) – Tokamak (D‑T, mid-2030s)
• Tokamak Energy (TE) (UK) – Spherical tokamak (D‑T, mid-2030s)
• Proxima Fusion (PF) (Germany) – Stellarator (D‑T, mid-2030s)
• Type One Energy (TOE) (USA/UK) – Stellarator (D‑T, mid-2030s)
• Xcimer Energy (XE) (USA) – Laser inertial fusion (D‑T, mid-2030s)
• Novatron Fusion Group (NFG) (Sweden)  - Mirror machine (D-T, n/a)
• Realta Fusion (ReF) (USA) – Compact mirror (D-T, 2035)
• Zap Energy (ZE) (USA) – Sheared-flow Z-pinch (D-T, 2040)

These concepts typically require tritium inventories in the range of a few and up to tens of kilograms. Inertial fusion approaches tend to require smaller quantities than systems based on magnetic confinement.

3. Tritium‑Free Fusion Concepts (No External Tritium Fuel)

These approaches avoid the tritium bottleneck entirely by using aneutronic fuels or D‑D / D‑³He cycles. Some produce trace amounts of tritium as a by‑product, but none require external tritium supplies. They also represent the most technically challenging pathways, requiring breakthroughs in confinement, heating, or laser technology.

Systems and Fusion Concepts

• Helion (Microsoft Plant, USA) – FRC (D‑D / D‑³He, 2028)
• Avalanche Energy (USA) – Electrostatic confinement (D‑D / optional D‑T) (late 2020s)
• TAE Technologies (USA) – FRC (p‑¹¹B aneutronic, 2030s)
• Marvel Fusion (Germany) – Laser‑driven p‑¹¹B (early 2030s)
• HB11 Energy (Australia) – Laser‑driven p‑¹¹B (2030s)
• Helion (Nucor Plant, USA) – FRC (D‑D / D‑³He, early 2030s)
• First Light Fusion (UK) – (known for D-T, but some concepts explore non‑D-T)
• Focus Fusion / LPPFusion (USA) – (p–¹¹B)
• NearStar Fusion (USA) – (D–D)
• Magneto‑Inertial Fusion Technologies, MIFTI (USA) – (D–D → D-T)

These concepts offer the long‑term promise of fusion without tritium constraints, but they demand the greatest advances in plasma physics, materials science, and driver technology.

Overall Perspective

Tritium availability remains a key limiting factor for fusion energy. However, the diversity of fusion approaches provides a pathway forward:

• Large research reactors demonstrate the scale of tritium needed for high‑power D‑T fusion.
• Startup‑scale D‑T concepts require far smaller inventories, making early deployment more feasible.
• Tritium‑free concepts offer the most sustainable long‑term solution, though they are also the most technically demanding.

This spectrum of approaches highlights both the challenges and the opportunities in the global pursuit of fusion power with constraints in tritium supplies. In the 2030s, the global demand may be up to 5 kg/year, mostly as initial inventory. In the 2040s, the demand will drastically increase up to 100 kg/year burned. Even if tritium breeding becomes successful, the external need may still be around 10–20 kg/year.

How does this compare to current tritium production in the civilian domain? The production of around 2 kg/y is dominated by detritiating the heavy water moderator of CANDU reactors for radiation protection purposes. Since the current global market is small (estimated at 0.1-0.3 kg/y), there is a surplus of tritium on stock believed to be around 35 kg but decaying with the half-life of 12.3 years.