SoftBank deploys world’s largest AI system with 4,000 Nvidia GPUs

SoftBank has deployed what it claims is the world’s largest artificial intelligence computing system, installing more than 4,000 of Nvidia’s most advanced Blackwell graphics processing units (GPUs) in a massive infrastructure expansion that signals Japan’s aggressive push into the global AI race.

The Japanese telecommunications giant’s new system delivers 13.7 exaflops of computing power—roughly equivalent to performing 13.7 quintillion calculations per second. To put that in perspective, this represents more raw computational horsepower than most national supercomputing centers, positioning SoftBank as a major player in the infrastructure arms race powering today’s AI revolution.

This deployment isn’t just about impressive numbers. SoftBank’s massive GPU cluster, built using Nvidia’s DGX SuperPOD architecture, represents a strategic bet on developing artificial intelligence systems specifically tailored for Japanese language and culture—a market largely dominated by Western AI companies like OpenAI and Google.

The technical foundation

SoftBank’s new AI platform runs on DGX B200 systems, Nvidia’s latest generation of specialized AI computers designed for training large language models. These systems are connected through Nvidia’s Quantum-2 InfiniBand networking technology, which enables thousands of processors to work together seamlessly on complex AI training tasks.

The deployment brings SoftBank’s total AI infrastructure to more than 10,000 GPUs, making it one of the largest private AI computing facilities globally. This scale matters because training sophisticated AI models requires enormous computational resources—often thousands of processors working in parallel for weeks or months.

SoftBank began building this infrastructure in 2023 with Nvidia’s earlier Ampere processors, then expanded in 2024 with Hopper-based systems. The recent Blackwell upgrade represents the most powerful and scalable configuration to date, reflecting the company’s commitment to staying at the cutting edge of AI hardware.

Domestic AI development strategy

The platform will be operated by SB Intuitions, SoftBank’s dedicated AI subsidiary focused on developing large language models optimized for Japanese users. This strategic focus addresses a significant gap in the current AI landscape, where most advanced systems are primarily designed for English-language interactions.

SB Intuitions previously built AI models with roughly 460 billion parameters—the variables that determine how an AI system processes and generates responses. The company now plans to launch a commercial model called Sarashina mini with 70 billion parameters by March 2026, using the expanded infrastructure to train more sophisticated Japanese-language AI systems.

This domestic focus carries both economic and strategic importance. By developing AI capabilities within Japan, SoftBank aims to reduce dependence on foreign AI systems while creating solutions better suited to Japanese business practices, cultural nuances, and regulatory requirements.

Japan’s Ministry of Economy, Trade and Industry (METI) has recognized this strategic value, certifying SoftBank’s platform under the country’s “Cloud Program” as critical national economic security infrastructure. Beyond internal development, SoftBank plans to offer access to Japanese enterprises and research institutions, supporting broader AI innovation across the country.

The artificial super intelligence vision

SoftBank’s infrastructure investment connects to CEO Masayoshi Son’s ambitious vision of achieving artificial super intelligence (ASI) within the next decade. ASI represents hypothetical AI systems that would surpass human cognitive abilities across all domains—from scientific research to creative problem-solving.

“We want to become the organizer of the industry in the artificial super intelligence era,” Son told shareholders at the company’s recent annual meeting. He compared the ASI opportunity to the “winner takes all” dynamics that have defined other technology sectors, pointing to how Microsoft dominated personal computing software, Amazon conquered e-commerce, and Google captured online search.

This vision explains SoftBank’s willingness to invest tens of billions in AI infrastructure and partnerships. Son believes that companies controlling the most advanced AI systems will capture disproportionate value across multiple industries, similar to how cloud computing giants now power much of the digital economy.

The Stargate connection

SoftBank’s domestic AI push runs parallel to its participation in Stargate, an ambitious $500 billion AI infrastructure project backed by President Donald Trump. The initiative involves SoftBank alongside OpenAI, Oracle, Microsoft, Arm, Nvidia, and technology investor MGX in what could become the largest AI buildout in history.

Reports suggest SoftBank may invest up to $25 billion directly in OpenAI, potentially becoming the AI company’s largest financial backer. Combined with over $15 billion committed to Stargate, SoftBank’s total AI investment could exceed $40 billion—a massive bet on controlling key pieces of the AI value chain.

This dual strategy—building domestic Japanese AI capabilities while partnering with leading Western AI companies—positions SoftBank to benefit regardless of which approach ultimately dominates the global market.

Industry implications

SoftBank’s deployment highlights the enormous capital requirements for competing in advanced AI development. The company’s investment in cutting-edge hardware demonstrates how the AI race increasingly favors organizations with access to massive computational resources and deep financial backing.

For the broader AI industry, SoftBank’s focus on Japanese-language models could accelerate development of more culturally and linguistically diverse AI systems. Most current AI development has concentrated on English-language applications, potentially creating opportunities for companies that successfully serve other major language markets.

The deployment also signals Japan’s determination to maintain technological sovereignty in AI development rather than relying entirely on American or Chinese systems. As AI becomes increasingly central to economic competitiveness and national security, more countries may pursue similar domestic capability-building strategies.

Looking ahead

SoftBank’s massive GPU deployment represents more than just impressive technical specifications—it’s a concrete step toward reshaping the global AI landscape. By combining domestic Japanese AI development with strategic international partnerships, the company is positioning itself to capture value from multiple AI development paths.

Whether SoftBank’s vision of artificial super intelligence materializes remains to be seen. However, the company’s willingness to commit tens of billions to AI infrastructure demonstrates the scale of investment that industry leaders believe will be necessary to compete in the next phase of artificial intelligence development.