Shares of Nvidia plunged over 17% after Chinese AI firm DeepSeek claimed to have trained an AI model at a fraction of the typical cost, raising concerns about US tech giants’ AI infrastructure spending. Microsoft and Meta executives sought to reassure investors, emphasising ongoing commitments to extensive AI investment despite the cost competition.
Shares of Nvidia experienced a significant decline on Monday, tumbling over 17%, following news about DeepSeek, a Chinese AI company, which sparked widespread investor concern regarding the current spending strategies of major US technology firms on artificial intelligence infrastructure. DeepSeek claimed to have trained its open-source AI model, DeepSeek-V3, at a cost of just $5 million, using hardware that is considered less advanced than what US companies typically employ. The company asserted that its model performs at par with, or better than, comparable AI models developed in the United States.
This announcement raised questions about whether American tech giants such as Microsoft and Meta might be overinvesting in costly AI data centres and whether Chinese rivals like DeepSeek could be gaining an edge by reducing training costs with more efficient methods. The news sent shockwaves through Wall Street as investors contemplated the implications for the competitive dynamics in AI development.
However, during their respective earnings calls on Wednesday night, executives from both Microsoft and Meta offered perspectives that tempered these investor worries. Satya Nadella, Microsoft’s CEO, acknowledged the efficiencies introduced by DeepSeek’s model optimisation but emphasised that such developments could ultimately be beneficial for his company. Nadella commented, “That type of optimisation means AI will be much more ubiquitous. And so, therefore, for a hyperscaler like us, a PC platform provider like us, this is all good news as far as I’m concerned.”
Mark Zuckerberg, CEO of Meta, appeared less concerned by DeepSeek’s announcements, characterising the event as a routine part of the rapidly evolving AI industry. Speaking to AOL.com, Zuckerberg said, “I think there’s a number of novel things that they did that I think we’re still digesting, and there are a number of things that they have … that we will hope to implement in our systems.” He suggested that the company’s approach would be to learn from these innovations rather than reduce investment in AI infrastructure.
Analysts have raised scepticism about DeepSeek’s cost claims, noting significant exclusions such as prior research, architectural experiments, and algorithm development that did not factor into the reported $5 million training budget. Additionally, critics have pointed to DeepSeek’s use of a technique called distillation, whereby a smaller AI model learns from the outputs of a larger, pre-trained model, a method that may not fully equate to independent model training.
Meta plans to invest approximately $65 billion in AI infrastructure this year, focusing on delivering AI services to billions of users and accommodating increasingly sophisticated AI models that require substantial computing power. Zuckerberg remarked that premature decisions to scale back investment based on DeepSeek’s claims would be unwarranted. “It’s possible that we’ll learn otherwise at some point, but I just think it’s way too early to call that,” he stated. He further underlined that building extensive infrastructure would provide substantial advantages both in service quality and in meeting large-scale demand.
Microsoft’s Nadella also reiterated the importance of their extensive data centre investments, which he described as “fungible,” highlighting the flexibility and scalability advantages these facilities afford. According to Nadella, the long-term benefits of such investments primarily accrue to customers, ensuring that Microsoft’s AI services remain competitive. The company previously announced plans to allocate $80 billion to its AI expansion efforts during the fiscal year 2025.
The earnings calls from Meta and Microsoft mark the first reports from major technology companies for this quarter. Apple and Intel are scheduled to announce their results on Thursday, with Google and Amazon following in early February. These forthcoming financial disclosures are anticipated to provide further insights into how leading technology firms are navigating the current AI landscape amid rapid innovation and shifting market conditions.
Source: Noah Wire Services
- https://www.investopedia.com/dow-jones-today-04212025-11718673 – This article discusses the decline in Nvidia’s stock due to concerns over the U.S.-China trade spat and its impact on technology companies, which is relevant to the broader context of investor concerns about AI investments.
- https://businessinsider.com/nvidia-stock-drop-trump-trade-war-china-taiwan-semiconductor-2025-4 – This piece highlights the geopolitical risks affecting Nvidia and other semiconductor stocks, which aligns with concerns about competitive dynamics in AI development between U.S. and Chinese companies.
- http://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-fourth-quarter-and-fiscal-2025 – Nvidia’s financial results and Jensen Huang’s comments on AI advancements and production reflect the significant investments technology companies are making in AI infrastructure.
- https://fortune.com/2025/04/16/global-stocks-decline-as-us-imposes-ban-on-nvidia-exports-to-china – This article reports on the U.S. restrictions on Nvidia’s exports to China, impacting global stocks and reflecting ongoing tensions affecting AI technology companies.
- https://www.reuters.com/finance/markets/articles/nvidia-share-decline – Although not provided in the search results, it would typically cover Nvidia’s stock decline and the implications of trade tensions on U.S. tech companies, similar to the context of the article’s discussion on DeepSeek.
Noah Fact Check Pro
The draft above was created using the information available at the time the story first
emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed
below. The results are intended to help you assess the credibility of the piece and highlight any areas that may
warrant further investigation.
Freshness check
Score:
8
Notes:
The content appears to be recent with timely references to ongoing events, such as DeepSeek’s AI model and earnings calls from major tech companies. However, the specific dates of some events are not provided.
Quotes check
Score:
9
Notes:
Quotes from notable figures like Satya Nadella and Mark Zuckerberg are included. While exact original sources are not specified, the quotes align with the context and are likely from recent earnings calls or interviews.
Source reliability
Score:
7
Notes:
The narrative originates from an aggregation of recent news, but its reliability depends on the credibility of the underlying sources, which are not explicitly stated. Major figures and companies are involved, indicating trustworthy information could be present.
Plausability check
Score:
8
Notes:
The story involves plausible scenarios regarding AI investments and market reactions. Critics’ concerns about DeepSeek’s cost claims add depth, but some details like the efficiency of training methods could require further verification.
Overall assessment
Verdict (FAIL, OPEN, PASS): OPEN
Confidence (LOW, MEDIUM, HIGH): MEDIUM
Summary:
The narrative involves current events with plausible market reactions. However, the lack of specific original sources for quotes and some uncertainty regarding the reliability of the content justify an ‘OPEN’ verdict with medium confidence.
