Deepseek Marks a 'Turning Point' in the Global AI Race

AI deepseek Google

Eric Schmidt, the former CEO of Google, in a column called on the United States to increase its efforts to open source AI in response to the rise of DeepSeek, a big AI model in China.

Schmidt also changed the rhetoric of “America leading” last year and stated in his column that the rise of DeepSeek marks a “turning point” in the global artificial intelligence race, proving that China can compete with large technology companies with fewer resources.

Schmidt pointed out that in order to compete with DeepSeek, the United States must develop more open-source models, invest in artificial intelligence infrastructure like Stargate, and encourage leading laboratories to share their training methods.

Stargate is an artificial intelligence project. It was jointly developed by OpenAI, SoftBank, and Oracle, with an initial investment of $100 billion. And it plans to expand to $500 billion over the next four years.

Schmidt has previously urged the United States to invest in artificial intelligence multiple times, as his personal investments may also benefit from it. Schmidt’s White Stork startup may be able to provide artificial intelligence drones for the US military. He is also an investor in Systematic Artificial Intelligence (Holistic AI), which helps companies comply with AI regulations, and Augment, an AI programming aid tool that competes with GitHub Copilot.

The global artificial intelligence industry is currently experiencing an “Eastern force”, namely the DeepSeek R1 model released by Hangzhou DeepSeek last week. According to the company, compared to the billions of dollars invested by large American technology companies in chatbots, the company has built a large model called DeepSeek that can rival OpenAI’s strongest inference model o1 at a small cost.

The cost of training the flagship v3 model behind DeepSeek AI assistant is reportedly only $5.6 million.

The launch of DeepSeek R1 has also caused a shock wave in the technology industry, leading to significant declines in the stock prices of major tech companies including Microsoft, Meta, and Nvidia this week.

Justin Post, an analyst at Bank of America Securities, wrote in a report, “If the cost of model training drops significantly, we expect advertising, travel, and other consumer application companies using cloud AI services to gain cost-effectiveness in the short term, while long-term revenue and costs related to hyper scale AI may decrease.”