- Meta commits to massive AI investment despite the market panic over DeepSeek's impact.
- Zuckerberg announced plans for $60 billion capital expenditure in 2025, that will be mainly for data centers.
- Meta also aims for Llama 4 to surpass closed models with new agentic and multimodal capabilities.
In the face of market uncertainty triggered by DeepSeek's AI developments, Meta CEO Mark Zuckerberg has reinforced his company's commitment to artificial intelligence.
Zuckerberg promised "hundreds of billions of dollars" in long-term investment during Meta's first-quarter earnings call on Wednesday.
The announcement comes amid market turbulence that saw Nvidia's stock plummet nearly 20% on Monday, driven by speculation that DeepSeek's AI models could significantly reduce GPU demand.
Despite these concerns, Zuckerberg remains steadfast in his AI strategy, having already committed to over $60 billion in capital expenditures for 2025, primarily directed toward data center development.
While addressing analyst concerns about DeepSeek's potential impact, Zuckerberg highlighted that substantial AI infrastructure investment remains a "strategic advantage" for Meta.
While Meta views DeepSeek as a new competitor worth learning from, it's "way too early" to predict any decrease in chip demand, particularly given Meta's billions of users and ongoing inference needs.
The tech giant's ambitious AI plans extend to its model development, and Zuckerberg has set aggressive goals for Llama 4.
While Llama 3 aimed to make open-source models competitive with closed alternatives, Llama 4 is being positioned to lead the field outright, even against established closed models like ChatGPT.
The upcoming model is expected to feature both agentic capabilities - similar to recent developments from OpenAI and Anthropic - and multimodal functionalities.
"Our goal with Llama 3 was to make open source competitive with closed models," Zuckerberg stated, adding emphatically, "And our goal for Llama 4 is to lead."
Edited By Annette George
