Big tech firms like Microsoft (MSFT) and Meta (META) have largely shrugged off concerns about Chinese-made AI firm DeepSeek’s claims of training its AI model for $5 million using less expensive chips. During their earnings calls, both CEOs acknowledged the efficiency improvements brought by DeepSeek’s AI technology but expressed optimism that it will ultimately benefit their companies as well.
Microsoft CEO Satya Nadella stated that the optimization introduced by DeepSeek means AI will become more ubiquitous, making it a good news story for Microsoft. Meta CEO Mark Zuckerberg downplayed the threat posed by DeepSeek, saying that the industry is constantly changing and new advancements are to be expected from every company.
However, analysts have questioned DeepSeek’s claims, noting that the $5 million figure does not account for costs associated with prior research and ablation experiments. Some have also criticized the technique used by DeepSeek to develop its model, which involves distilling an AI model using questions and answers provided by a larger model.
Despite these concerns, Meta plans to spend over $65 billion on AI infrastructure this year, citing the need for accurate answers to complex questions depending on computing power. Microsoft’s Nadella believes that its investments in data centers will help the company scale long-term business, with the ability to build out such infrastructure being a major advantage.
Other big tech companies are set to report their earnings later this week, and each is expected to have something to say about the latest AI shake-up.
Source: https://finance.yahoo.com/news/meta-microsoft-downplay-deepseek-threat-190639250.html