DeepSeek: Tech Giants Unfazed, Continue AI Infrastructure Spending
DeepSeek's recent announcement sent shockwaves through Wall Street, raising concerns about the competitiveness of Big Tech companies like Microsoft (MSFT) and Meta (META) in the AI realm.
However, earnings calls held by Microsoft and Meta executives this week suggest otherwise. Despite acknowledging DeepSeek's efficiency gains, CEOs Satya Nadella (Microsoft) and Mark Zuckerberg (Meta) expressed confidence in their companies' AI strategies.
"AI optimization means it will be more ubiquitous," said Nadella. "This is good news for a hyperscaler like Microsoft." Zuckerberg downplayed the threat, stating that the industry is constantly evolving.
DeepSeek's low training cost of $5 million using less advanced chips has raised questions about the necessity of massive AI infrastructure spending by US tech companies. However, analysts have scrutinized DeepSeek's claims, noting omitted costs and the use of distillation techniques.
Zuckerberg emphasized that it's premature to consider scaling back AI infrastructure investments based solely on DeepSeek's data. Meta plans to allocate $65 billion for AI infrastructure this year, driven by the demand for accurate AI models and the company's vast user base.
Nadella echoed this sentiment, highlighting the importance of Microsoft's data center investments for long-term scalability. The company has earmarked $80 billion for its AI buildout by fiscal 2025.
With earnings reports looming for other Big Tech players like Apple (AAPL), Intel (INTC), Google (GOOG), and Amazon (AMZN), it is expected that the ongoing AI landscape and its implications will be key discussion points.