SoftBank is taking a significant step into the battery production sector by repurposing a closed Sharp LCD factory in Sakai, Osaka. This large-scale manufacturing facility is set to begin operations as soon as fiscal 2027. Unlike traditional battery producers such as CATL and Panasonic, SoftBank's primary objective is not to compete in the consumer market but rather to create batteries that will power its own AI data centers.
The demand for data center power is forecast to triple by 2030 due to the rapid advancement of AI technologies. This increase in energy needs compels companies reliant on AI infrastructure to innovate in how they source electricity. Rather than wait for existing energy supplies to adapt, SoftBank is proactively establishing its power supply chain.
By converting an existing industrial space, SoftBank can launch this initiative more quickly and cost-effectively than building a new facility. The anticipated timeline for production is roughly two years, which is striking in the context of battery manufacturing.
A noteworthy aspect of SoftBank's battery strategy is its decision to forgo lithium and cobalt in its cell compositions. These materials have complicated battery supply chains, primarily due to their geopolitical implications and ethical sourcing issues linked to cobalt mining in the Democratic Republic of Congo. Instead, SoftBank aims to develop batteries that are affordable, durable, and scalable, vital characteristics for stationary storage applications, which do not require the high energy density demanded by electric vehicles.
What does this mean for the broader AI energy landscape? Larger technology firms like Microsoft and Amazon are taking steps to secure energy through various means, such as investing in nuclear power and modular reactors. They recognize that current electricity grids cannot meet their AI operational needs. SoftBank's approach diverges from this trend by focusing on energy storage solutions rather than generation. Through battery systems, the company can stockpile energy during low-usage periods and deploy it during spikes in AI demands, offering flexibility in Japan’s electricity market, which is marked by high costs and limited grid capacity.
The Osaka facility could become one of Japan's largest battery production lines. However, executing this plan carries inherent risks. Battery manufacturing is notoriously challenging, particularly when it comes to scaling production. New material chemistries can yield unpredictable results, and transitioning from pilot testing to large-scale manufacturing can present challenges even for seasoned manufacturers. The timeline for SoftBank's fiscal 2027 goal is ambitious, considering the company's lack of experience in battery cell production.
By owning critical components of the AI ecosystem—from Arm's chip designs to various AI enterprises—SoftBank's initiative to secure a dedicated and cost-effective power source is essential. This investment aims to relieve one of the most significant bottlenecks in expanding AI operations.