The orders were arranged through Silver Bands 7 LLC, a SoftBank subsidiary responsible for securing manufacturing capacity ahead of competing buyers. With gas turbine manufacturers facing substantial demand and delivery schedules extending into the early 2030s, SoftBank sought to reserve equipment early rather than risk delays to its infrastructure projects.
The strategy was closely tied to CEO Masayoshi Son's broader investments in artificial intelligence. SoftBank had committed $33 billion through its SB Energy subsidiary to develop the Ohio power project, which is associated with the PORTS-Pike Technology Campus leased by OpenAI. The company was also expanding its financial exposure to OpenAI as part of its larger AI investment strategy.
However, changes to the Ohio project's specifications reduced the amount of equipment required. SoftBank subsequently canceled advance orders that were no longer necessary, resulting in a ¥597.2 billion writedown, equivalent to approximately $3.7 billion.
The financial consequences were substantial because SoftBank had already committed significant funds to turbine manufacturers. By September 30, 2026, cumulative advance payments for U.S. power generation equipment had reached $9.48 billion. Some of those deposits were nonrefundable, leaving the company to recognize losses on equipment reservations it no longer expected to use.
SoftBank retains rights to conditional rebates worth as much as $2.1 billion against future turbine purchases. Those potential credits could offset part of the financial impact, although their value depends on whether the company makes qualifying purchases.
The cancellations also drew attention after Elon Musk suggested that SoftBank had failed to meet its turbine blade contract obligations. SoftBank's chief financial officer rejected that characterization, stating, "there is no such default." The company has attributed the canceled orders to revised project requirements rather than an inability to fulfill its financial commitments.
The turbine writedown was a major factor in SoftBank's fiscal second-quarter results. For the period ending September 30, the company recorded a net loss of ¥1.12 trillion, or roughly $7 billion. Higher financing costs amid volatile market conditions also contributed to the decline.
The episode illustrates the financial risks associated with securing equipment far in advance of finalized infrastructure requirements. AI developers and data center operators have faced lengthy waits for power connections, prompting greater interest in dedicated natural gas generation. SoftBank's purchasing strategy reflected an attempt to secure scarce manufacturing capacity before competitors could obtain it.
Peter Gardett, CEO of energy technology company Noreva, told TechCrunch that the decision to reserve more equipment than necessary was partly a response to supply constraints.
"SB Energy did not have a choice except to overbook turbines in the future if they wanted some turbines right away," Gardett said.
The approach offered SoftBank a way to obtain equipment sooner, but it also exposed the company to losses if its eventual requirements changed. Rather than risk securing too little capacity, SoftBank committed to more turbines than its projects ultimately needed.
"This charge is just that price discovery finally happening," Gardett said.
The broader question is whether dedicated natural gas generation will remain as attractive to AI infrastructure developers as initially expected. While gas turbines can provide an alternative to waiting for grid connections, their use introduces operational considerations that may affect long-term planning.
Brent Nelson, senior managing director of markets and strategy at Ascend Analytics, noted that natural gas supplies can face restrictions during winter periods when demand increases. He suggested that data centers may benefit from maintaining grid connections while using large battery systems to manage peak electricity demand or temporary reductions in utility service.
That approach could reduce the amount of dedicated generation capacity required for some facilities. As operators gain more experience managing AI workloads and their electricity needs, their infrastructure decisions may also become more precise.
SoftBank's canceled orders do not necessarily indicate a broader retreat from natural gas power generation. Demand for electricity from AI data centers remains substantial, and the long manufacturing lead times for turbines continue to influence purchasing decisions. The company's losses instead highlight the consequences of committing billions of dollars before individual projects have settled on their final requirements.
Despite the writedown, SoftBank continues to make major investments in artificial intelligence. The company has acquired a 13% stake in OpenAI for $64.6 billion and announced a separate $1.85 billion commitment to AI infrastructure development in Malaysia.
The immediate financial impact is clear, but several elements of SoftBank's power strategy remain unresolved. Its ability to recover value through future turbine purchases will depend on the conditions attached to its potential rebates, while the Ohio project's development will determine how much generation equipment is ultimately required.
For SoftBank, the $3.7 billion charge represents a costly adjustment to its infrastructure procurement strategy rather than an announced withdrawal from AI development. The company remains heavily invested in the sector, but its turbine purchases demonstrate how rapidly changing project requirements can turn early access to critical equipment into a significant financial liability.
This analysis is based on reporting from TechCrunch, Tom's Hardware & CryptoBriefing.
Image courtesy of Dealroom.
This article was generated with AI assistance and reviewed for accuracy and quality.
