Meta CEO Mark Zuckerberg has confirmed that the company will continue to invest heavy in Artificial Intelligence (AI) despite the recent market concerns due to the advances of Dipsek. On Monday, Nvidia’s stock fell to about 20 percent as it was feared that Deepsek’s AI model graphics could reduce the demand for graphics processing units (GPU). However, during the Meta’s earnings on Wednesday, Zuckerberg assured investors that the Meta’s plan to invest billions of dollars in AI in the long term is unchanged.

Meta has already planned to spend more than $ 60 billion in 2025, mainly on the construction and upgradation of data centers to support its AI infrastructure. This step is seen as the key to the long -term strategy of the meta. Asked if the progress of Deepsac can affect Meta’s AI spending, Zuckerberg said that the company will continue to invest in AI, seeing it as a major benefit, especially when Meta serves billions of users around the world. Does. He also suggested that Deepsek helped him restore his trust in investing heavily in AI space.

Deepsek is a competitor but it is not an immediate danger

Zuckerberg considers Dipsek a new contestant in AI Space, but said it would be too early to know if it will slow down the GPU’s demand. He believes that GPU will be important to run AI works like data processing. The mass of the meta of the meta requires a strong AI infrastructure to handle the scale of its operations, and Zuckerberg believes that the investment in the region will give the meta a competitive edge.

He explained, “At this point, I will bet that the ability to create the kind of infrastructure is going to be a major benefit for both and is able to serve the scale we want.” He also said that what Deepsek has achieved with less money, he “only (his) strong confidence has been correct.”

Zuckerberg also praised US President Donald Trump. He said, “We now have an American administration that proud of our major companies, prioritizes winning American technology and will defend our values ​​and interests.”

Lama 4: Meta’s next big AI leap that can reduce human jobs

An important part of Meta’s AI strategy is the development of its next AI model, Lama 4. Zuckerberg revealed that this model would be designed to compete with top AI models such as OpenAII chats. LLAma 4 is expected to have advanced facilities, such as the ability to work on functions by self (agent capabilities) and handling many types of inputs (multimodal abilities).

“Lama 4 is making a lot of progress in training. Lama is done with 4 mini pre-training and our logic models and big models also look good,” Zuckerberg said.

“Our goal with Lama 3 was to make the open source AI model competitive with a closed model, and with Lama 4, we want to lead,” Zuckerberg said. He said that Lama 4AI can set new standards in the world and help Meta to become a leader in Open-SOS AI.

Zuckerberg recently talked about his vision for 2025, calling it an important year for AI development. He said that by 2025, AI in meta and other technical companies could effectively change middle-level engineers who currently write codes. This will represent a big innings how technical companies look at software development.

He said, “We will reach a point where all the codes and AIs that produce it in our apps will also be written by AI engineers instead of people’s engineers.” For reference, Business Insider reported that the middle-level software engineers in Meta currently earns salary in mid-six data-a cost may be less than a large margin for AI company.

Mark Zuckerberg aims to reach a billion users with Meta AI

In addition to its big AI goals, Meta is also focusing on expanding its AI Assistant, Meta AI, which Zuckerberg hopes to reach more than one billion users by the end of the year. The meta approach to AI focuses on personalization, giving users an experience according to their personal preferences, interests and culture. Zuckerberg believes that this personal touch is one that will separate the meta from its rivals.

“We believe that people do not all want to use the same AI. People want their AI to be personal for their references, interests, personality and culture,” he said.