SoftBank's AI Energy Aspirations Hit a Snag with Major Losses

SoftBank has made headlines for its extensive investments in emerging technologies, and its recent moves into the energy sector are no exception. The company has shown a pattern of making large bets on market trends, which now has led to some challenges in the natural gas arena.

Big Bets on Natural Gas and AI

In early 2025, SoftBank invested $30 billion into OpenAI, demonstrating its commitment to the AI sector. Alongside this, the company's subsidiary, SB Energy, announced an ambitious plan in February to allocate $33 billion for constructing a 9.2-gigawatt natural gas power facility. This aggressive approach seems to be a part of a larger strategy to dominate sectors where they place investments.

Overordering Turbines

Approximately two years ago, SoftBank made a remarkable move by placing deposits for over 13 gigawatts of turbines—nearly 50% more than what was necessary for their projects. This strategic overordering was intended to secure supplies and accelerate project timeliness.

Peter Gardett, CEO of Noreva, pointed out that such a tactic was practically unavoidable if SB Energy wanted turbines ready in the near term. The energy sector was unprepared for the rapid surge in power demands spurred by AI technologies, leading to a bottleneck in turbine supply chains extending into the early 2030s.

The Fall of a Big Gamble

Unfortunately for SoftBank, this gamble has turned sour. The company revealed a shocking $7.1 billion loss in a recent quarterly report, with a substantial portion attributed to power-generating equipment investments. Much of the financial strain stems from nonrefundable deposits made to turbine manufacturers, which are now considered near-worthless.

According to Gardett, this situation represents a painful but necessary learning moment regarding prices in the energy market. SoftBank's strategy of overordering has backfired, costing them over $3 billion in excess spending.

Evaluating Energy Sources for AI

The enthusiasm for natural gas emerged from the pervasive uncertainties in the AI power landscape. Many companies in the AI sector are seeking reliable and accessible energy sources amid prolonged wait times for grid connections. Natural gas appeared to be a secure option, as it is a well-established energy resource.

However, risks still linger. Winter shortages can sharply raise demand while reducing supply, leading to potential crises, as noted by Brent Nelson, senior managing director at Ascend Analytics. In light of these factors, firm grid connections might prove more beneficial for most data centers, even if smaller than originally planned. Integrating large battery systems could also enhance energy reliability, providing an alternative to meet peak demands.

No Immediate Fallout for Natural Gas

Despite SoftBank’s staggering losses, experts believe that these challenges are unlikely to result in widespread cancellations of natural gas projects. The demand for energy within the AI sector remains too substantial for that. For the foreseeable future, natural gas continues to be a key player, even though SoftBank may have reached its limit in this area.