'The Power of Open Source': Meta's Chief AI Exec Opens Up on DeepSeek's Disruption in the Space

The sudden release of DeepSeek's R1 model, the latest AI innovation from the Chinese firm, has taken the industry by surprise with its impressive performance, recently outpacing popular offerings from companies such as OpenAI, Meta, and Anthropic.

One notable figure to comment on DeepSeek's performance is Yann LeCun, Meta's Vice President and Chief AI Scientist. He emphasised how the model's success owes much to open-source research and projects that have enabled such advancements.

Surpassing Proprietary Models

In a post on his LinkedIn account, LeCun pointed out that the discussion around DeepSeek's popularity should not focus on how China is now leading the United States in AI research and development, but rather on how open-source models can surpass proprietary alternatives.

'DeepSeek has profited from open research and open source (e.g. PyTorch and Llama from Meta). They came up with new ideas and built them on top of other people's work. Because their work is published and open source, everyone can profit from it. That is the power of open research and open source,' LeCun explained.

His statement echoes Mark Zuckerberg's intention to make Meta's AI models—specifically Llama—open source. The tech giant's founder recently announced that Meta plans to invest more than £48.23 billion ($60 billion) in 2025 as it doubles down on AI.

'Part of my goal for the next 10 to 15 years, the next generation of platforms, is to build the next generation of open platforms and have the open platforms win. I think that will lead to a much more vibrant tech industry,' he stated.

A Threat to AI Giants?

The recent launch of DeepSeek's R1 model caused shockwaves across Western AI firms, with many experiencing notable share price declines following the announcement.

Nvidia took a particularly large hit, dropping 16.9%, while its rival Broadcom slumped by 17.4%. Meanwhile, tech giants Microsoft and Google's Alphabet were down 2.14% and over 4%, respectively, and Oracle and Cisco slipped 14% and 5%.

The BBC also reports that in Europe, the Dutch chip equipment company ASML saw its share price drop by over 7% by the end of Monday's trading. Siemens Energy, a manufacturer of AI-related hardware, experienced a sharp decline of 20%.

Analysts have noted that DeepSeek's success lies not only in its open-source technology but also in how, despite a relatively modest £4.5 million ($5.6 million) investment, it managed performance on par with OpenAI's offerings.

It is also worth noting that DeepSeek's achievement challenges the common assumption that AI model performance necessarily improves with greater numbers of parameters, larger training datasets, and more computing power.

'If DeepSeek's model is confirmed and were to be widely adopted in the industry, a large share of the massive AI infrastructure currently being built could end up in excess capacity, causing prices to fall and AI investment strategies to be reassessed. For the semiconductor, cloud services and data centre industries, the stakes are high,' wrote Aurélien Duthoit in a piece for Coface.

Heightened AI Investment in China

DeepSeek's announcement coincides with increasing levels of AI investment in China. Observers note that Chinese large language models (LLMs) are rapidly closing the performance gap with US models, with some outperforming their American counterparts in bilingual benchmarks.

In January, the state-owned Bank of China revealed plans to provide at least £110 billion (1 trillion yuan) in funding to companies in the AI sector over the next five years. This includes 'special comprehensive financial support' for all entities in the AI industry chain, with at least £33.30 billion (300 billion yuan) allocated to stocks and bonds. The bank also intends to establish 'special institutional guarantees' tailored to support AI innovation and cater to the industry's financial requirements.

DeepSeek's latest AI model underlines China's expanding prominence in global AI innovation, marking a significant milestone in the nation's technological ambitions. Amid ongoing geopolitical tensions with the United States, this accomplishment highlights China's capacity to compete at the cutting edge of AI research and reinforces its determination to reduce reliance on Western technology.

By demonstrating its potential to disrupt established norms in AI development, China positions itself as a formidable player in shaping the sector's future. In doing so, it could reshape the global AI ecosystem and alter the dynamics of technological leadership.