SoftBank Group’s mobile unit is stepping into a corner of the AI race that has quietly become its most important: power. The Japanese telecom and technology conglomerate announced this week that it will begin large-scale battery cell manufacturing at its Sakai plant in Osaka, in a direct response to the surging electricity demand from artificial intelligence services and the data centers that run them. The move underscores a broader shift across the global technology industry, where the limiting factor in AI development is no longer access to GPUs or model architecture, but reliable, scalable electricity.
The Sakai initiative is structured around two newly launched ventures. The first, AX Factory, will focus on AI data center operations and hardware. The second, GX Factory, will produce next-generation batteries, solar panels, and related clean energy products. Production capacity at the Sakai plant could reach several gigawatt-hours, with SoftBank signaling plans for global expansion.
For SoftBank, the move integrates three previously separate businesses — telecom, energy, and computing — into a single AI-aligned strategy. For the wider industry, it is a clear signal that infrastructure now sits at the center of the global AI competition.
The Bottleneck Has Shifted From Chips to Power
For most of the past three years, the dominant AI infrastructure narrative was about access to GPUs from companies like Nvidia. That narrative is being rewritten in real time. Industry analysts increasingly describe the next phase as a race to bring compute online faster, which means securing electricity, grid connections, cooling systems, and battery storage at scale.
The shift has practical implications. Data centers built for AI workloads consume far more electricity than traditional cloud infrastructure, and they often need to operate at constant maximum capacity for extended periods. Renewable generation, while abundant in some regions, is intermittent. That intermittency makes large-scale battery storage one of the most valuable assets a hyperscaler or AI infrastructure builder can secure.
SoftBank’s positioning is unusual in that it integrates the storage, generation, and compute sides under one corporate umbrella. Few global technology companies have attempted to vertically build energy capacity alongside compute capacity at this scale.
A Global Race Playing Out Across Asia, Europe, and the US
The SoftBank announcement does not exist in isolation. It is one of several major AI-related infrastructure and commercial moves landing in a single 24-hour window across multiple regions.
In China, Alibaba is integrating its Qwen AI model with Taobao, the country’s largest e-commerce platform. The move is intended to reshape consumer shopping around AI agents that handle product search, recommendations, and checkout inside a conversational interface. It reflects a broader race among Chinese technology companies to embed AI agents directly into everyday consumer transactions rather than as separate add-on features. According to South China Morning Post coverage, the deployment is a platform-level strategy rather than a product feature, signaling that AI shopping agents are moving from experimental phase to commercial deployment.
China also ranked third in a recent global index measuring AI competitiveness in life sciences, trailing the United States and the United Kingdom. That ranking matters because life-sciences AI is one of the few areas where AI infrastructure spending is producing measurable revenue, drug pipeline progress, and clinical trial efficiency gains.
In Turkey, the global AI economy is producing its own ripple effects. Istanbul-based mobile gaming startup Grand Games raised $70 million in a Series B round led by Balderton Capital, demonstrating that investor appetite for category-specific consumer technology remains strong even in a cautious global venture environment. Turkey has produced a string of notable gaming exits in recent years, and the country’s startup pipeline continues to attract major European venture firms.
Across all three developments, the through-line is the same: AI competition is no longer happening only in research labs. It is happening in factories, e-commerce platforms, and energy grids.
Why Energy May Be the Real Moat
For startups and hyperscalers alike, energy access is becoming a competitive moat that is difficult to replicate quickly. Building a battery factory, securing land for a data center, signing power purchase agreements, and navigating grid interconnection processes can take years. Companies that can bring compute online faster than competitors stand to win the next phase of the AI race, regardless of whose models are technically best.
SoftBank’s structure positions it to do exactly that. With AX Factory handling data center operations and hardware, and GX Factory handling the energy and storage side, the company can theoretically build, power, and operate AI infrastructure within a single corporate ecosystem. The Sakai plant becomes a node in that system, with global expansion already on the roadmap.
The same logic is driving billion-dollar infrastructure deals elsewhere. Cloud providers, telecoms, and chip companies are signing power purchase agreements with utilities, investing in nuclear and renewable projects, and acquiring data center sites near abundant electricity supply.
The Sakai launch is just one step in what is likely to be a multi-year cycle of energy infrastructure investment tied to AI growth. Analysts expect more announcements from telecom-energy integrations, sovereign-backed AI infrastructure funds, and regional clean energy initiatives designed to support compute-intensive workloads.
For Japan, the SoftBank initiative is also a signal of how the country is positioning itself in the global AI economy: not as a model developer first, but as an infrastructure power. Whether that strategy pays off will depend on execution at scale, the pace of global AI adoption, and how quickly other nations and companies build their own competing infrastructure footprints.
What is no longer in question is the direction of travel. The AI economy is being rebuilt around energy, and the companies that secure power early will set the terms of the next decade of competition.
