In a matter of months, a relatively unknown Chinese AI startup has begun to challenge the status quo, raising doubts about the future of US leadership in the field. DeepSeek, a newcomer to the artificial intelligence scene, has outperformed some of the most established names in the industry, including OpenAI’s ChatGPT, despite having limited resources and a development timeline that seems almost unbelievable.
While large-cap US tech companies like OpenAI have raised billions in capital and employ thousands, DeepSeek has made its mark with a fraction of that investment and workforce. The company’s development costs were reported to be under $10 million, using chips that are considered less advanced than those powering US AI systems. Yet, users are already comparing DeepSeek favourably to ChatGPT, with DeepSeek surpassing its competitor in several key areas. This rapid growth has surprised many, especially considering that tech analysts had once assumed that companies like OpenAI, with their $17.9 billion in funding, were untouchable leaders in AI.
The success of DeepSeek is more than just impressive performance benchmarks. It’s about how swiftly the company has scaled. In categories such as AIME, MATH-500, and GPQA, DeepSeek is proving to be a strong contender, often outperforming ChatGPT. While ChatGPT still holds an advantage in coding-related tasks, DeepSeek is quickly closing the gap in ways that are hard to ignore. For a product that took only months to develop, this is a remarkable achievement—and one that has captured the attention of the industry.
The market’s reaction has been swift and telling. After DeepSeek became the top free app on the App Store, the Nasdaq 100 futures experienced a sharp decline, losing 330 points in just a few hours. Investors in US tech stocks are clearly concerned, as this unexpected shift suggests DeepSeek could soon pose a serious threat to established giants. As the app’s popularity continues to grow, concerns about the future of US AI companies are becoming louder.
DeepSeek’s rapid success highlights the vulnerabilities in the US tech sector. The company’s cost-efficiency is a significant factor in this shift. DeepSeek’s R1 model is available at a fraction of the cost of OpenAI’s services. With DeepSeek charging just $2.19 per million output tokens, compared to OpenAI’s $60 per million tokens, investors are rethinking the value of US AI companies. DeepSeek’s ability to offer comparable services at such a much lower price could have wide-ranging effects on the competitive environment.
But it’s not just the cost that’s making an impact. DeepSeek’s R1 model is fully open-source, giving it a clear advantage over proprietary systems like ChatGPT. Open-source software has long been a driving force in the tech industry, and DeepSeek’s decision to follow this path means developers and researchers around the world can access, modify, and build upon the technology. This openness could accelerate its development and application in ways a closed system can’t match.
This openness, combined with DeepSeek’s affordability, places it in a strong position. Its API is reported to be user-friendly, and rate limits aren’t an issue for those seeking to integrate it into agentic AI systems. Nvidia has already stated that agentic AI is the next big thing, and with DeepSeek set to be integrated into these systems, its growth potential seems nearly limitless.
The mood has shifted, especially among those invested in the “Magnificent 7” tech stocks—the seven largest US technology companies. These stocks, which have been central to the bull market over the past two years, are now trading at levels around two standard deviations higher than what was seen in 2001. Much of this increase has been driven by expectations around AI hardware and software, but with DeepSeek’s emergence, those assumptions are being called into question.
The question on many minds now is whether DeepSeek’s rise signals the end of US tech dominance. DeepSeek’s rapid growth and cost-efficiency have taken many by surprise, and as the company continues to make headway in the AI sector, it’s clear that the competition for supremacy in this field is far from settled.
Liang Wenfeng, the founder of DeepSeek, has been central to this success. In 2021, Liang started buying thousands of Nvidia graphic processing units for his AI side project while running his quant trading fund, High-Flyer. Industry insiders initially dismissed it as the actions of a billionaire looking for a new hobby. “When we first met him, he was this very nerdy guy with a terrible hairstyle talking about building a 10,000-chip cluster to train his own models. We didn’t take him seriously,” said one of Liang’s business partners.
“He couldn’t articulate his vision other than saying: I want to build this, and it will be a game-changer. We thought this was only possible from giants like ByteDance and Alibaba,” the person added.
Liang’s outsider status in the AI space turned out to be an unexpected asset. At High-Flyer, he built a fortune by using AI and algorithms to spot patterns that could influence stock prices. His team became highly skilled at using Nvidia chips to trade stocks. In 2023, he launched DeepSeek, with the goal of developing human-level AI.
As investors and industry leaders continue to monitor DeepSeek’s growth, one thing is clear: the industry is changing. With its performance, pricing, and open-source approach, DeepSeek is positioning itself as a disruptive force—one that could alter the balance of power in AI and beyond. Whether the US tech sector can rise to meet this challenge remains to be seen, but DeepSeek’s rapid rise is a stark reminder that in the fast-evolving tech space, nothing is guaranteed.