U.S. markets took a hit on Monday, spooked by whispers that DeepSeek’s AI models might cut into the demand for GPUs. Nvidia’s stock tumbled nearly 20% as investors panicked. But Meta’s CEO Mark Zuckerberg isn’t sweating it. During the company’s first-quarter earnings call, he doubled down on Meta’s commitment to AI, vowing to pour “hundreds of billions of dollars” into the technology over the long haul.

Just last week, Zuckerberg revealed Meta’s plan to spend over $60 billion in 2025 on capital expenses, with a big chunk going toward data centers. When asked about DeepSeek’s potential impact, he brushed it off, saying heavy investment in AI infrastructure will remain a “strategic advantage” for Meta.

DeepSeek is definitely on Meta’s radar as a new competitor, but Zuckerberg stressed it’s “way too early” to predict whether chip demand will shrink. He pointed out that chips are still essential for AI inference, and with billions of users, Meta needs to keep scaling up.

“Building that kind of infrastructure is going to be a major advantage for both the quality of our service and our ability to handle the scale we’re aiming for,” Zuckerberg said.

Meta’s next AI model, Llama 4, is designed to be a game-changer. Zuckerberg wants it to outshine even closed models like ChatGPT, with plans to roll out agentic and multimodal capabilities.

“Our goal with Llama 3 was to make open-source competitive with closed models,” he explained. “With Llama 4, we’re aiming to lead.”

Despite market jitters, Meta seems all-in on AI, making it clear they’re not backing down anytime soon.