SoftBank's $7.1 Billion Quarterly Loss Traces Back to Turbine Deposits
SoftBank has reported a $7.1 billion quarterly loss, more than half of it tied to power-generating equipment, according to TechCrunch. The equipment-related portion exceeds $3 billion.
The losses trace back to turbine orders. About two years ago, SoftBank began placing deposits on more than 13 gigawatts (GW) of gas turbines, almost 50% more than the capacity it said it needed. Those deposits were paid to turbine manufacturers and are nonrefundable. The company now considers much of that money worthless.
The turbines were meant to serve a large natural gas project. In February, SB Energy announced plans to spend $33 billion on a 9.2 GW natural gas power plant. That figure is a plan, and the reported losses do not change its announced scope. What the quarterly results show is that the equipment bought to support the buildout has already produced a material accounting hit, before a single megawatt of that plant has been commissioned.
Frame turbines, the large heavy-duty gas turbines used in utility-scale plants, are in short supply, and queues for new orders now extend into the early 2030s. A buyer that wants them delivered on a useful schedule has to commit capital long before it knows its load, its offtake contracts (the agreements committing someone to buy the power), or its financing terms. A nonrefundable deposit functions as the price of a place in that queue. If the project shrinks, slips or changes configuration, the deposit does not come back, and the place in line may not transfer cleanly to another project.
That is the mechanism behind the loss. Securing capacity early was a rational response to a long lead time. The cost arrives when the order book runs ahead of the demand it was sized for, and when the deposits cannot be redeployed. Turbine deposits are the least glamorous line item in an AI-era power strategy, and they are the first one to appear in the quarterly numbers.
In my view, the more useful question for the sector is not whether gas turbines remain a viable path to firm capacity. It is how much optionality a developer should buy, and at what price, when equipment queues are this long. Over-ordering by roughly half is a measurable cost of hedging against timing risk. Buyers who can match deposits to signed power contracts, or who can negotiate transferable slots, will carry less of that exposure. Those that cannot will carry more.
The broader context here is that gas generation can be built and brought online faster than many alternatives at utility scale, which is why it keeps appearing in large power plans. Turbine scarcity is a supply constraint the industry can address over time through new manufacturing capacity and better procurement practice. The companies that come through this cycle with clean balance sheets will be the ones that learned to commit capital to power in step with demand, not ahead of it.
For now, the numbers are the story. A $7.1 billion quarterly loss, more than half linked to power equipment, deposits on capacity beyond the plan, and a queue that runs into the 2030s. Each fact is narrow on its own. Together they describe what happens when the infrastructure for compute-heavy growth is ordered faster than the demand behind it can be confirmed.
