To The Moon

The AI investment boom has created enormous opportunities for semiconductor companies, but another constraint is becoming increasingly difficult to ignore: getting enough electricity to data centres quickly enough.

According to Reuters, investment in fuel cells for data centres could rise from $2.8 billion in 2025 to approximately $30 billion by 2030, based on projections from Rystad Energy.

That is a potentially significant opportunity for an industry that has traditionally received less attention than AI chips.

⚡ 1. Why fuel cells are attracting attention

AI data centres require reliable electricity around the clock to power computing equipment and cooling systems.

The problem is that connecting large new facilities to the US electricity grid can take three to six years, according to Rystad estimates.

Fuel cells offer an alternative by generating electricity on-site through an electrochemical reaction rather than conventional combustion within the cell.

For data-centre developers, the attraction is straightforward: if on-site power can be deployed faster than a grid connection, it could help bring expensive computing capacity online sooner.

Rystad estimates that around 40% of projected US data-centre capacity in 2030 could pursue dedicated on-site power.

🚀 2. Bloom Energy (BE): The company to watch

Bloom Energy is emerging as a key player in this market.

According to the Reuters report:

• Third-quarter revenue surged 138% to $1.24 billion.

• The company reported net income of $250.3 million, compared with a loss of $23.1 million a year earlier.

• Its 2026 revenue forecast was raised to $4.4 billion–$4.6 billion.

• Oracle intends to procure up to 2.8 gigawatts of Bloom fuel-cell systems, with 1.2 gigawatts contracted.

These figures point to strong commercial momentum. However, investors should distinguish between a customer’s announced procurement plans, contracted capacity and systems actually delivered.

Bloom’s opportunity depends on converting growing demand into sustainable revenue, profits and cash flow.

🔋 3. FuelCell Energy (FCEL): Another potential beneficiary

FuelCell Energy is also targeting the data-centre market, although it uses a different fuel-cell technology.

The company expects data centres and related business to account for 75%–85% of its business within two to three years, up from 25%–30% currently, according to Reuters.

That ambition creates potential upside if the market expands as expected.

However, FCEL should not automatically be treated as equivalent to Bloom Energy. Their technologies, financial positions, commercial progress and ability to scale differ. Investors need to examine each company’s cash requirements, margins and project pipeline separately.

📈 4. Could this become the next major AI investment theme?

The opportunity extends beyond fuel-cell manufacturers.

Companies involved in electricity generation, grid infrastructure, power management and electrical equipment could also benefit as data-centre developers seek faster ways to secure power.

But there are important limitations.

Fuel-cell projects still require permits, fuel connections, transformers and other supporting infrastructure. Manufacturing capacity must also expand to meet potential demand.

Furthermore, fuel cells are not automatically emissions-free. Their environmental impact depends partly on the fuel used, and systems relying on natural gas still involve upstream emissions.

The industry’s projected growth will ultimately depend on whether suppliers can deliver reliable electricity at competitive costs.

🎯 My takeaway

The AI trade may be expanding beyond the familiar chipmakers.

As the industry builds more computing capacity, electricity availability could become an increasingly important factor determining which projects proceed and how quickly they generate revenue.

Bloom Energy offers exposure to this emerging opportunity, while FuelCell Energy represents a different, potentially higher-risk route into the sector.

The key is to separate genuine commercial progress from projections that still depend on future contracts, manufacturing expansion and successful project execution.

AI needs more than computing power. It needs electricity, and the companies that can deliver it reliably may become an increasingly important part of the AI infrastructure story.

For discussion purposes only.