Live Markets, Charts & Financial News

SoftBank in talks to get its hands on large volumes of Nvidia’s GPUs: Report By Investing.com

0 7

Japanese technology investor SoftBank (TYO:) is in discussions to find a way to secure a large amount of graphics processing units (GPUs) from Nvidia that are essential for the development of artificial intelligence (AI).

The latest move appears to be part of a broader plan by CEO Masayoshi Son to position SoftBank as a leader in AI technology, as it has begun discussions with banks about borrowing money to finance up to $10 billion in energy-related projects driven by demand for AI, The Information reported Thursday.

SoftBank, a backer of companies like Tencent, T Mobile, 10x Genomics (NASDAQ:), Automation Anywhere and others, may look to create a special purpose vehicle to buy GPUs with borrowed money, then lease the GPUs back to SoftBank, allowing the company to keep debt off the balance sheet.

The ongoing discussions show SoftBank’s intention to increase its exposure to AI infrastructure. Its plans to secure Nvidia’s (NASDAQ:) GPUs demonstrate the growing importance of these chips in the development of AI technologies, especially in their role in processing complex AI algorithms.

Meanwhile, SoftBank is also looking into energy investments to meet the power demands of AI data centers. Innovations in renewable energy and nuclear technology are being considered to ensure adequate power supply for these centers.

By contrast, Son has so far shied away from investing in generative AI companies, which have recently become a favorite among investors across geographies. He blocked SoftBank’s Vision Fund from investing in Mistral, a Paris-based open-source modeling company, The Information reported.

Instead, Son is pinning all his hopes on Arm, the U.K.-based chip designer acquired by SoftBank in 2016. Son sees Arm as the linchpin of his AI ambitions, with its designs crucial to the future development of AI chips. Arm’s IPO and growing market value have boosted SoftBank’s position in the AI ​​hardware space.

Leave A Reply

Your email address will not be published.