DeepSeek claims its R1 model cost under $6 million to make|Tim Reckmann|CC-BY-NC 2.0
After learning of DeepSeek’s expenditure on artificial intelligence, tech CEOs felt the need to reassure investors that spending big on AI is necessary for innovation and to maintain the US’s lead in the technology.
On their respective earnings calls yesterday, Nadella and Zuckerberg defended their $80 billion and $65 billion AI investments earmarked for 2025, arguing that large-scale infrastructure is crucial for long-term success.
Zuckerberg emphasized that it was too early to comment on how DeepSeek’s model will impact the AI ecosystem and his company. Meta’s shares rose almost 2% on Monday after investors appeared calm.
In his call with investors, Microsoft CEO Satya Nadella praised DeepSeek but stuck to its spending plan, which CFO Amy Hood says will grow next year.
Shares declined 6% on Thursday after reporting weaker growth in its Azure and other cloud services.
Both CEOs appear optimistic that DeepSeek will further increase demand for AI infrastructure.
However, some investors are skeptical of the large spending and the lack of big payoffs.
The most-used AI app, ChatGPT, and its maker, OpenAI, are not profitable yet. The Sam Altman company projected losses of $5 billion and revenue of $3.7 billion for 2024, per the New York Times September article.
Analysts warn that US companies need to show a clear path to monetizing AI rather than just increasing spending.