Data centre construction, a key component of the expanding artificial intelligence (AI) infrastructure, faces significant challenges as global investment in these facilities is projected to reach $31.6 trillion by 2050. While the sector rushes to meet growing demand, industry experts warn that major building projects frequently encounter delays and cost overruns, a pattern that may repeat in the AI data centre boom.
SB Energy, a subsidiary of Japan’s SoftBank, exemplifies the ambitious push in this space. The company has filed for an initial public offering in New York, aiming to attract investors with a portfolio that includes the sizable PORTS-Pike development in Ohio. Unlike companies engaged in manufacturing chips and servers, SB Energy focuses on constructing the physical infrastructure, potentially insulating itself from the volatility of hardware prices. The firm has secured long-term contracts, including deals that span more than a decade with OpenAI, positioning itself to meet clear demand for AI-related data processing facilities.
However, building data centres carries notable risks. Challenges such as shortages of materials, labor constraints, planning delays, and securing power supply approvals are common. In the United States, gaining “interconnection” permissions—approval to connect the facility to the electrical grid—can extend from two years to as long as a decade depending on the state, complicating construction timelines. SB Energy’s strategy includes co-locating data centres with new power plants, adding layers of complexity due to the regulatory and construction demands of energy infrastructure.
Public opposition also presents a growing hurdle. Despite statements from former President Donald Trump criticizing local resistance to data centre projects, recent polling indicates strong community pushback. A Gallup survey found that 71 percent of Americans somewhat or strongly oppose the construction of large AI data centres near their homes, reflecting broader concerns about environmental and social impacts.
SB Energy acknowledges evolving public resistance as a material risk in its regulatory filings. While the company benefits from early filings for more grid connections than it currently requires, early applications do not guarantee timely project approvals. Unlike some developers who find tenants late in the construction process, SB Energy has the advantage of pre-secured customers, but this also raises pressure to meet deadlines. Delays could force the company into discounted rates or compel clients to buy out projects at reduced prices, potentially straining cash flow.
Financially, SB Energy’s ambitions are substantial. The company reportedly seeks around $50 billion in enterprise value through its offering, which would imply a valuation roughly 400 times its earnings before interest, taxes, depreciation, and amortization (EBITDA) based on its first half of 2026 results. OpenAI’s acquisition of warrants in SB Energy, which could yield returns if the company’s valuation reaches $80 billion, signals confidence from key technology players.
The data centre sector’s expansion is widely seen as critical to supporting future AI and digital infrastructure. While the U.S. has demonstrated capacity for large-scale construction projects, as seen in the shale energy boom, not all megaprojects meet expectations. The path ahead for data centres involves balancing ambitious growth with navigating regulatory, technical, financial, and community acceptance challenges.