Meta CEO Mark Zuckerberg addressed the potential implications of DeepSeek’s groundbreaking AI model during a call with analysts on Wednesday, asserting that it is still too early to determine how the new technology will impact his company or the tech industry at large. DeepSeek, a Hangzhou-based AI lab linked to the Chinese hedge fund High-Flyer, made waves this week by unveiling its R1 large language model, which was developed in just two months with a budget of less than $6 million. The revelation raised questions about the resources required to build cutting-edge AI models, particularly in terms of processing power and infrastructure costs.
Despite the growing excitement over DeepSeek’s cost-efficiency, Mark Zuckerberg downplayed the notion that Meta’s AI spending would soon decrease as a result. He emphasized that Meta’s AI strategy remains focused on long-term growth and infrastructure investments. While acknowledging that it is too early to draw definitive conclusions, Mark Zuckerberg remained confident in the importance of Meta’s AI initiatives. “There are a bunch of trends that are happening here all at once,” he said, referring to the various developments in the AI space.
DeepSeek’s breakthrough shook technology markets this week, causing shares of Nvidia, a major supplier of graphics processing units (GPUs), to fall by 17% on Monday. The market reacted negatively to the idea that businesses may need fewer Nvidia chips for AI projects if DeepSeek’s model is an indication of where the industry is headed. In response, Meta’s stock rose nearly 2% on Monday, suggesting that some investors believe Meta could benefit from these advancements by building powerful AI systems more cost-effectively.
During Meta’s fourth-quarter earnings call, Mark Zuckerberg reaffirmed the company’s commitment to AI, announcing that Meta’s sales rose by 21% year over year, reaching $48.39 billion, surpassing Wall Street’s expectations. Despite the excitement surrounding DeepSeek’s cost-efficient model, Meta has made clear that it plans to continue its ambitious AI investments. The company revealed plans to allocate between $60 billion and $65 billion to its AI initiatives in 2025. Meta also projected its total expenses for the year to fall between $114 billion and $119 billion.
Mark Zuckerberg emphasized that Meta is still analyzing DeepSeek’s achievements and intends to apply some of these advancements to its own AI projects in the future. “That’s part of the nature of how this works, whether it’s a Chinese competitor or not,” he said, referencing how technological advancements in one company can benefit others across the industry. Although DeepSeek’s technology may result in more efficient models, Zuckerberg cautioned that running such models at scale may still require significant computational resources to ensure high-level intelligence and service quality.
In his comments, Mark Zuckerberg explained that serving the massive number of users on Meta’s platform, which reached 3.35 billion daily active people in the fourth quarter, will likely necessitate maintaining a substantial infrastructure. “It’s going to be expensive for us to serve all of these people,” he acknowledged. Mark Zuckerberg reiterated his belief that Meta’s ongoing investments in infrastructure and capital expenditures (CapEx) would ultimately serve as a strategic advantage in the AI space, even if current trends may suggest otherwise.
Meta’s focus on learning from technologies like DeepSeek’s and releasing its own AI models remains a key part of its strategy. Mark Zuckerberg pointed out that Meta’s open-source Llama AI models are part of the company’s broader push to contribute to the AI industry while gaining a competitive edge. “I continue to think that investing very heavily in CapEx and infra is going to be a strategic advantage over time,” Mark Zuckerberg said. While it’s possible that Meta’s approach may evolve as AI technology progresses, Mark Zuckerberg believes that building out infrastructure will continue to be an important factor in maintaining a competitive position.
Mark Zuckerberg also expressed his support for an open-source approach to AI, stating that DeepSeek’s emergence has validated Meta’s commitment to this direction. He emphasized the importance of establishing an open-source AI standard globally, particularly from an American perspective. “For our own national advantage, it’s important that it’s an American standard,” he said, highlighting the strategic value of the U.S. leading the way in AI development.
As Meta continues to push forward with its AI investments, the company’s ability to adapt to the evolving landscape and leverage cost-effective innovations like DeepSeek’s will be critical in shaping its future growth in the tech industry.
