AMD is refining its AI strategy, beating industry leader Nvidia to the acquisition of an AI startup.
AMD revealed on Wednesday it announced it could acquire Silo AI, a Finnish startup, connections “the largest private AI lab in Europe.” The total money transaction is price roughly $665 million and is expected to close in the second half of this yr.
The deal is not going to impact AMD’s financial results because it “unlocks a significant portion of the business for the future,” AMD senior vp Vamsi Boppana said. Reuters Agency.
The purchase for over half a billion dollars overshadows $125 million that AMD has invested in AI corporations in the past 12 months. AMD also brought in a French company Mypsology and based in Santa Clara Nod.ai joined the group last yr for an undisclosed amount.
Nvidia also could have been acquired by AMD if things had turned out otherwise nearly 20 years ago: According to industry insiders, AMD almost bought Nvidia in 2006. The merger allegedly fell through because Nvidia CEO Jensen Huang wanted to run the combined company.
AMD CEO Lisa Su. Photographer: Annabelle Chih/Bloomberg via Getty Images
Currently, the AI chips market is dominated by Nvidia, which has, among others, 70% and 95% market share, although competition is growing from all sides. Established chipmakers like Intel and AMD are stepping up their efforts, while startups like Tenstorrent are also joining the race, with AMD being considered for second place for Nvidia at the moment.
AMD’s recent acquisition, Silo AI, employs over 300 AI experts, including over 125 AI PhD researchers. The startup operates in North America and Europe, creating each open-source AI models and custom AI solutions for clients comparable to Unilever AND Philips.
Boppana highlighted the wealth of AI talent at Silo AI, in addition to the team’s experience building AI on AMD platforms in press release. He said the acquisition would “further accelerate” AMD’s AI development, AMD CEO Lisa Su also said. it echoed.
AI development costs thousands and thousands, if not billions, of dollars. Anthropic CEO Dario Amodei said in June that training an AI model currently costs about $100 million, but “there are models currently being trained that cost more like a billion.”
Nvidia the biggest Customers Microsoft, Meta, Amazon and Alphabet’s Google are also believed to be building their very own AI chips, which could eventually reduce their reliance on Nvidia. Google’s environmental report last week showed the tech giant is expecting higher emissions because it continues to invest in AI development.