Meta Responds to DeepSeek’s AI Breakthrough with Four Strategic Teams

HomeTechnology

Meta Responds to DeepSeek’s AI Breakthrough with Four Strategic Teams

Zuckerberg Unveils S$87.9 Billion Investment to Enhance AI Capabilities Amidst Growing Competition

my-portfolio

Meta has reportedly established four dedicated teams, or "war rooms," in response to DeepSeek, a promising AI startup backed by the Chinese hedge fund

Affordable and Reliable Tyre Workshops in Singapore for Quality Used Tyres
Micron’s $7 Billion Move Reinforces Singapore as a Global Semiconductor Leader
Kusu Island Becomes Fully Self-Sufficient with Solar Power, Producing Its Own Water and Electricity

Meta has reportedly established four dedicated teams, or “war rooms,” in response to DeepSeek, a promising AI startup backed by the Chinese hedge fund High-Flyer Capital Management. DeepSeek recently launched its open-source AI chatbot, R1, at a significantly lower cost, sparking concerns within Meta about its competitive position in the AI market.

At the same time, Meta’s CEO, Mark Zuckerberg, announced plans for a substantial investment of up to US$65 billion (S$87.9 billion) to strengthen the company’s AI initiatives. This move comes just days after OpenAI unveiled its ambitious AI project, Stargate, which is backed by the White House and is set to involve new data centres across the United States, in partnership with major companies like SoftBank and Oracle.

According to The Information, Meta’s AI infrastructure director, Mathew Oldham, has expressed concerns that DeepSeek’s latest AI model might surpass Meta’s upcoming Llama AI. This has heightened pressure within the company, with Zuckerberg hinting that Llama’s new version could launch by early 2025.

In response, Meta has formed four specialised teams to tackle the challenge. Two of these teams will investigate how DeepSeek managed to significantly reduce the cost of training and running its AI models, with the aim of implementing similar strategies for Meta’s Llama. The remaining two teams will focus on identifying the data DeepSeek utilised for its model and examining how these insights could be incorporated into future versions of Llama.

A Meta spokesperson told The Information: “We regularly assess competitive AI models as part of our development process. Llama has been instrumental in establishing a strong foundation for open-source AI, and we’re excited to expand on that leadership with the upcoming launch of Llama 4.”

COMMENTS

WORDPRESS: 0
DISQUS: