DeepSeek’s Threat to OpenAI Remains Limited, Experts Say Amidst China-US AI Competition
Tech advancements demonstrated by DeepSeek are evidence that China-US AI competition is on. According to CNBC, major tech companies hold that the emergence of DeepSeek means that China cannot be ignored in the AI innovation space.
China – US AI Rivalry
Last month, the Chinese AI startup triggered turmoil in the global tech markets after its technical paper claimed that it spent less than $6 million to develop its latest AI model. This amount is far less than the billions western AI startups like Anthropic and OpenAI are spending.
But OpenAI’s Chief Global Affairs Officer, Chris Lehane says DeepSeek’s AI models confirm that there is real competition between the US-led democratic AI and Chinese-led authoritarian AI. DeepSeek critics also claim that the AI model censors sensitive topics. For instance, the AI assistant responded with “Sorry, that’s beyond my current scope. Let’s talk about something else” when it was asked about the Tiananmen Square massacre that happened in 1989.
“There’s two countries in the world that can build this at scale. Imagine if there were only two countries in the world that could build electricity at scale. That’s sort of how you have to think about it. For us, what DeepSeek really reinforces and reaffirms is that there is this very real competition with very real stakes,” Lehane said.
Limited Threat
Tech leaders agree that DeepSeek shows that China has made considerable progress in the global AI race than initially thought. However, tech executives think that the Chinese AI model does not pose significant risk to OpenAI’s competitiveness. Although there is a general agreement that advances made by the Chinese company are impressive, questions around the cost of developing its AI models abound.
“Most of the market fear around DeepSeek is in fact misplaced. It still requires large models – it was distilled from large models. I think the short answer everyone should take is, game on, but large models still really matter,” LinkedIn Co-Founder Reid Hoffman said.
SemiAnalysis, a semiconductor research firm, estimated DeepSeek’s hardware spending throughout its history to be more than $500 million. The report by SemiAnalysis also found that research and development spending and costs related to ownership are significantly high. Generating synthetic data to train its AI model would also cost a considerable amount.
In light of these cost concerns, some tech leaders hold that Chinese AI startups may have achieved high performance by training its AI models on large US AI systems- a technique called distillation. This technique involves using more powerful models to evaluate the responses that a newer model generates. OpenAI has alluded to this claim too. Last month, the AI startup said it was reviewing allegations that DeepSeek may have used output data from its AI models inappropriately to develop its latest model.
Challenging Status Quo
DeepSeek has claimed that its R1 model rival’s OpenAI’s o1 model while using less costly and energy efficient processes. This claim has caused experts to question the decision by the West to restrict AI chip exports to countries such as China. AI companies need graphic processing units to train and operate AI applications.
But DeepSeek’s R1 model has challenged this. It is “a big deal in showing that the game is on. The competition is afoot with China. It’s a credible, actionable model,” Hoffman said
Geopolitical Business Advisory Firm founder Abishur Prakash says DeepSeek has shown that the West has a limited understanding of China.
“America’s assumed place as the technological captain of the world is no longer the acceptable belief. That is the new status quo now, that the space between the U.S. and China has narrowed almost overnight- but it hasn’t narrowed overnight, it’s been years of progress. If there’s one takeaway for the West, it’s that their understanding of China is incredibly limited- and we don’t know what’s coming next,” Prakash said.
Early this month, DeepSeek announced plans to restrict access to its app interface service. Due to shortages in its server capacity.
