Meta Scrambles to Compete with Chinese AI Startup DeepSeek’s Rapid Success
DeepSeek AI has shaken up the AI landscape in the United States. Just weeks after launching its AI model, DeepSeek surpassed ChatGPT to become the number one free app on the App Store. But the impact didn’t stop there — DeepSeek’s rise sent shockwaves through the tech industry, contributing to a staggering $400 billion market cap loss for NVIDIA in the US.
To solidify its momentum, DeepSeek recently introduced Janus-Pro 7B, an innovative image generation model that has been making waves. This new model has outperformed competitors like OpenAI’s DALL-E, Stability AI’s Stable Diffusion, and others in multiple benchmarks, further establishing DeepSeek as a leader in the space.
DeepSeek’s Rise Grabs Meta’s Attention
DeepSeek’s rapid ascent has caught the attention of Meta. In response, Meta has reportedly mobilized four specialized teams, dubbed “war rooms,” to analyze how this Chinese AI startup, backed by High-Flyer Capital Management, has been able to rival or surpass top competitors like ChatGPT — all while operating at a fraction of the cost.
The tipping point for DeepSeek’s popularity was the launch of its R1 model, an AI chatbot that outperformed ChatGPT. Remarkably, DeepSeek claims to have invested less than $6 million in training its model, compared to OpenAI’s reported $100 million investment for ChatGPT. Meta’s war rooms are now working to assess the potential threat posed by this disruptive breakthrough.
Meta’s Strategy: Learn and Compete
Two of Meta’s war rooms are focused on understanding how DeepSeek achieved such low costs in developing and running its R1 models, with hopes of applying similar strategies to Meta’s AI model, Llama. A third team is investigating the training data DeepSeek used, and the fourth is exploring ways to redesign Llama’s architecture to compete with Chinese AI technology.
Meta’s spokesperson declined to comment directly on the internal efforts but shared this statement with The Information:
“We regularly evaluate all competitive models in our development process and have done so since [the company’s] Gen AI [group] was formed. Llama has been foundational in establishing the ecosystem for open-source AI models, and we couldn’t be more excited to extend this leadership with the upcoming release of Llama 4.”
AI Industry on High Alert
Meta’s leadership is particularly concerned because, according to internal discussions, DeepSeek’s latest model could potentially outperform the upcoming Llama AI, set to launch in early 2025. Even OpenAI CEO Sam Altman has publicly acknowledged DeepSeek’s rise, calling it “impressive.” Meanwhile, NVIDIA, one of the biggest tech players impacted by DeepSeek’s ascent, also commended the Chinese AI startup while pointing out that NVIDIA GPUs were instrumental in powering DeepSeek’s software.
With its groundbreaking models and cost-efficient approach, DeepSeek is rapidly reshaping the AI landscape, leaving Silicon Valley giants like Meta and OpenAI scrambling to adapt.
via The Information
Research Snipers is currently covering all technology news including Google, Apple, Android, Xiaomi, Huawei, Samsung News, and More. Research Snipers has decade of experience in breaking technology news, covering latest trends in tech news, and recent developments.
