Softbank’s Acquisition of Graphcore Confirmed

Softbank, a prominent technology conglomerate in Japan, has recently made headlines by confirming its acquisition of the UK-based AI chip startup, Graphcore. This strategic move solidifies Graphcore as a wholly owned subsidiary of SoftBank, allowing it to maintain its operations under the familiar Graphcore brand. The acquisition positions Graphcore alongside another UK chip design giant, ARM, under the SoftBank umbrella.

Expressing his enthusiasm for the acquisition, Graphcore's co-founder and CEO, Nigel Toon, highlighted the significance of the deal. Toon, who recently authored a book on AI, praised his team's ability to develop groundbreaking AI technologies at scale. He described the acquisition as a validation of their efforts and a pivotal moment for the company.

Despite its promising trajectory, Graphcore faced challenges in the competitive AI chip market. The company had secured substantial investments totaling over $710 million, positioning itself as one of the most well-funded AI chip startups. However, the reported $400 million acquisition by SoftBank marked a significant loss for investors, including industry players like Dell.

Graphcore's innovative approach centered around its Colossus chip, featuring on-chip SRAM memory designed to handle complex AI frameworks efficiently. To complement its hardware offerings, the company developed proprietary servers, racks (referred to as pods), and a specialized development tool named Poplar. While Graphcore made strides in the AI sector, it faced setbacks such as job cuts and challenges in scaling its operations.

SoftBank's investment arm, led by Managing Partner Vikas J. Parekh, emphasized the importance of next-generation semiconductors and computing systems in advancing artificial general intelligence (AGI). Parekh expressed optimism about collaborating with Graphcore to drive innovation in AI technology. With headquarters in Bristol and additional offices in key locations like Cambridge and London, Graphcore remains poised to shape the future of AI.