SoftBank's Ambitious AI Investment: $1 Billion Initial Impact Drive
SoftBank will invest an initial 150 billion yen ($960 million) to bolster its computing facilities by 2025 in preparation for development of its own generative AI
SoftBank, the Japanese tech conglomerate, is embarking on a mission to pioneer a cutting-edge Japanese-language-specific generative artificial intelligence model, aiming to attain global excellence in this field. A recent Nikkei report unveiled SoftBank's ambitious plan, disclosing its commitment to inject a staggering $960 million into enhancing its computing infrastructure over the next two years.
The development and training of large language models (LLMs), such as the renowned Chat GPT by OpenAI, demand top-tier graphics processing units (GPUs). To meet this requirement, SoftBank intends to procure GPUs from the prominent U.S. chip manufacturer, Nvidia, as revealed by sources cited in the Nikkei report on Monday.
This substantial investment of 150 billion yen ($960 million), slated for 2024 and 2025, supplements the 20 billion yen SoftBank allocated for computing infrastructure upgrades in the previous year. Nikkei's report highlights that this investment stands as the largest of its kind made by a Japanese entity to date, positioning SoftBank to wield unparalleled computing capabilities within the country upon completion.
Despite the burgeoning interest in technology, Japan currently grapples with a deficiency in private enterprises equipped with high-performance supercomputers essential for LLM development, as outlined in another report from Nikkei Asia. SoftBank's significant investments have the potential to rectify this shortfall, furnishing Japan with a formidable domestic contender in the realm of generative AI—a critical move amidst escalating efforts by international players to infiltrate this market.
Recent developments underscore the intensified competition in Japan's AI landscape. OpenAI's establishment of its inaugural Tokyo office and Microsoft's announcement of a $2.9 billion investment aimed at fortifying its cloud computing and AI infrastructure in the nation are indicative of this trend.
SoftBank has outlined its roadmap, anticipating the completion of its inaugural model boasting 390 billion parameters by fiscal year 2024—a testament to the model's complexity. Moreover, plans are underway to commence the development of a higher-performance model featuring a staggering 1 trillion parameters as early as 2025.
NTT, a prominent Japanese telecommunications firm, is also in the race, having disclosed its intentions to roll out an LLM within the fiscal year. Additionally, NTT has earmarked a substantial investment of 8 trillion yen ($51.7 billion) over the next five years for bolstering growth areas like data centers and AI.
Projections from Statista Market Insights paint a promising picture for Japan's AI market, forecasting a growth trajectory to approximately $13 billion by 2030—a seventeen-fold increase from 2023 figures.
The strategic pivot towards AI has borne fruit for SoftBank, evident in the favorable trajectory of its stock price, which has surged by approximately 20% year-to-date. Moreover, as the majority stakeholder in Arm, a leading chip company, SoftBank stands to reap significant benefits amid the burgeoning AI landscape.
SoftBank's foray into AI infrastructure extends beyond mere investments, with reports suggesting plans for the establishment of AI data centers across Japan. Notably, the conglomerate has joined forces in a venture to construct a 65 billion yen center in Hokkaido, underscoring its commitment to shaping Japan's AI ecosystem.
