zeeshan khan
Mark Zuckerberg, the chief executive officer (CEO) of Meta, stated on Wednesday that it is too soon to tell how DeepSeek’s innovative AI approach would affect his business and the larger IT sector.
The idea that Meta’s total AI budget will soon decrease, however, was shot down by Zuckerberg, who said that DeepSeek had demonstrated that advanced AI needs significantly less funding and processing power than originally believed.
Zuckerberg stated that it is likely too early to get an opinion about the implications of this for the direction of facilities and capital expenditures. “A number of trends are taking place here concurrently.”
DeepSeek AI Model Drives Market Shifts and Meta’s Outlook
This week, a model of AI developed by DeepSeek, a Hangzhou-based AI lab connected to the Chinese hedge fund firm High-Flyer, sent the technology markets into a downward spiral. According to DeepSeek, the development of its R1 big language model took less than $6 million and took two months to complete. Because of the inference that businesses no longer require every one of the chip manufacturer’s graphic processing units for their own artificial intelligence initiatives, Nvidia’s stock fell 17% on Monday.
During a conference call with investors on Wednesday to announce Meta’s fourth-quarter profits, Zuckerberg talked about DeepSeek. Sales increased 21% from last year to $48.39 billion, the business stated, above Wall Street’s projections.
The fact that Meta’s stock increased by over 2% on Monday suggests that some investors think DeepSeek’s expense-effectiveness may allow the social networking site to develop strong AI systems at a lower cost. The business announced that the company will spend between $114 billion and $119 billion in 2025.
Meta to Leverage DeepSeek’s AI Advancements
According to Zuckerberg, Meta is still taking in a few of DeepSeek’s achievements, and his team plans to use several of these developments in their own AI initiatives in the future.
Zuckerberg stated, “That’s a component of the reason how this operates, regardless if it’s a Chinese competitor or not,” implying to the way that advancements made by one business ultimately help others.
In the future, businesses might need fewer servers to teach these powerful models, but Zuckerberg stated that once the program is really operating and carrying out tasks, a significant amount of the server power might be required, leading to “a greater degree of intellect and a greater level of service.”
According to Zuckerberg, who’s business had 3.35 billion active users per day in the final quarter of the year, “it’s likely to be costly for me to reach all of these individuals since we help a lot of users.”
Meta’s Open-Source AI and DeepSeek’s Impact
Meta will keep learning from DeepSeek’s technology and provide its own open-source, free Llama artificial intelligence models that can advance the sector. According to Zuckerberg,
“I still believe that making significant investments in property and infrastructure is going to give a competitive edge,” He stated. “I think it’s far too soon to call that, but at this point, I guess maybe being able to expand out this sort of network will prove to be an important benefit.” “It’s likely that we’ll discover differently at some stage.”
Additionally, Zuckerberg stated that the rise of DeepSeek confirms Meta’s dedication to an open-source AI strategy.
“A global open-source framework will emerge,” He declared. “It’s critical to establish an American standard for its own national benefit.”
