Ceres Power: Licensing fuel cell tech as AI demands explode power

Ceres licenses technology instead of building plants, keeping gross margin at 79%. Partners like Doosan and Delta fund all manufacturing while Ceres collects fees and royalties. No capex drag, no inventory risk, no factory debt. Revenue scales through partner capacity, not Ceres balance sheet. The model converts intellectual property into recurring cash without dilution or operational complexity.

Doosan launched the world's first Ceres-licensed fuel cell factory in South Korea with 50MW annual capacity. First commercial sales expected before year-end, triggering royalty payments. Delta committed £170 million for a Taiwan facility targeting data centres. Weichai signed a manufacturing licence in November. Development phase is over, partners are building and selling, shifting Ceres from engineering revenue to royalty income.

AI workloads push data centre power demand from 3% to 12% of US electricity by 2030. Grid can't expand fast enough, transmission takes decades. Fuel cells deploy on-site in months, not years. Solid oxide cells offer higher efficiency and lower cost than alternatives. Ceres technology sits at the centre of urgent infrastructure bottleneck with validated product and proven manufacturing partners already scaling.

• Delta factory construction progress: Delta committed £170 million to build a Taiwan facility. Groundbreaking or timeline updates signal serious capital deployment by a tier-one partner, reducing execution risk and proving data centre demand is real enough to justify major manufacturing investment.

• Weichai production ramp in China: Manufacturing licence signed November 2025. First production milestones or capacity targets would unlock the world's largest stationary fuel cell market. China's energy policy and data centre growth could drive volume faster than Western markets, multiplying royalty streams.

• Second-generation stack cost reduction: Ceres developing next-gen technology targeting 30% lower manufacturing cost. Commercial release and partner adoption would widen margin advantage over Bloom Energy, accelerate payback periods for customers, and expand addressable market by improving economics for smaller deployments.

Hydrogen electrolyser market emergence: Ceres technology works in reverse to produce green hydrogen. As hydrogen infrastructure builds out for industrial decarbonisation and transport, electrolyser licensing could open a second royalty stream independent of power generation, diversifying revenue beyond data centres into energy transition megatrend.

Doosan secured only one 9MW contract in four months post-launch, potentially generating just £1.35 million in royalties for Ceres. Market uptake slower than projected, partners may struggle to convert factory capacity into actual sales volume. If commercialization stalls, royalty income disappoints and licensing model credibility weakens. Past partnerships with Bosch, Nissan, and Honda failed without generating meaningful revenue, raising execution risk.

Bloom Energy controls 44% of the global stationary fuel cell market with 400MW already deployed to data centres and proven technology since 2010. FuelCell Energy and Plug Power also compete directly. Ceres depends entirely on partners to manufacture and sell, no direct market presence. If established players capture AI data centre demand, Ceres partners lose pricing power and volume, compressing royalty streams before they scale.

Top three partners, Doosan, Delta, Weichai, drive nearly all revenue. Bosch exited in 2024, terminating development with no revenue generated after years of collaboration. If any major partner fails, delays production, or shifts technology, Ceres loses income stream with limited control. Licensing model transfers all commercialization risk to partners who control timing, pricing, and market success.