Despite market concerns over DeepSeek’s AI advancements, Meta remains committed to significant AI investments.
On Wednesday, during Meta’s first-quarter earnings call, CEO Mark Zuckerberg reaffirmed that the company will continue spending heavily on AI, even suggesting investments could reach hundreds of billions of dollars in the long run. This comes after he announced plans to allocate over $60 billion in 2025, primarily for expanding Meta’s data centers.
When asked about DeepSeek’s potential effect on Meta’s AI spending, Zuckerberg stated that investing in AI infrastructure remains a strategic advantage. He acknowledged DeepSeek as a competitor but emphasized that it is too soon to determine whether demand for GPUs will decline, as these chips are still essential for AI processing.
With billions of users, Meta sees AI infrastructure as a key asset for delivering high-quality services at scale.
Zuckerberg also highlighted Meta’s ambitions with its next AI model, Llama 4. The company aims to make it the most competitive AI model available, surpassing both open-source and proprietary alternatives.
He noted that Llama 4 will incorporate advanced features, including agentic capabilities—similar to developments by OpenAI and Anthropic—as well as multimodal capabilities for enhanced functionality.
With Meta’s continued push into AI, Zuckerberg remains confident that scaling AI infrastructure will provide a competitive edge. While DeepSeek has stirred discussions about AI efficiency and chip demand, Meta is staying the course, focusing on developing models that can lead the industry.
