The rapid advancements in artificial intelligence (AI) technology have resulted in numerous generative AI models being introduced by different companies, such as CHATGPT and Gemini. Until recently, the majority of online users were increasingly dependent on the ChatGPT model for reasoning tasks, but this shifted with the launch of the DeepSeek model. In January, a Chinese artificial intelligence (AI) company based in Hangzhou developed an AI model known as DeepSeek R1, which claimed that its software contended with ChatGPT-maker OpenAI in its reasoning capabilities and performance with a lesser cost to create.
The launch of DeepSeek R1 technology sent shockwaves throughout the investment world. The United States tech stocks fell rapidly, and the chip giant Nvidia's stock plummeted, which resulted in a loss of approximately $593 billion in market capitalisation. This disruption by the Chinese in the American market contests the firmly ingrained belief that American companies would retain the most dominance in the AI sector. This shift in market positions raises uncertainties about the future of innovation, data sovereignty, and the geopolitical implications of AI leadership.
Inside DeepSeek’s Innovation
DeepSeek technology primarily centres around generating open-source large language models (LLMs). It launched its first model back in November 2023, which since then has been improved, leading towards new variations being released with time. Deepseek would not gain its global recognition until the release of its R1 reasoning model in January 2025, which advances AI’s problem-solving and logical reasoning. DeepSeek's training process for its R1 models notably consumed lesser time with reduced AI accelerators and reduced cost structure to develop as compared to OpenAI. The company claims to have developed the technology in two months with an estimated cost of $6 million (£4.8m) as compared to it taking OpenAI more than $100m (£80m) to train GPT-4. Principal AI analyst at Counterpoint Research Wei Sun remarks that "DeepSeek has proven that cutting-edge AI models can be developed with limited compute resources." This achievement has placed the company in the global tech market.
DeepSeek has emerged as a strong industry competitor for OpenAI, which established itself as an industry leader after its groundbreaking success in 2022 for its ChatGPT model. In its closed-source GPT model series, OpenAI launched a proprietary language model created for general-purpose conversational AI. It is a subscription-based model with a structured Application Programming Interface (API ) pricing system. This closed-source system is expensive, and it restricts the customisation alternatives for developers over the granted API access. In comparison, DeepSeeks offers more flexibility for developers with its open-source system and its API pricing being more economically viable. This is because processing one million input tokens with GPT-4o costs $2.50, while for DeepSeek-R1, it costs $0.55 for the same number of tokens.
Most investors sold off their stocks after reevaluating future growth prospects and capital expenditure for Nvidia, which caused the company to lose billions of dollars from its market valuation
This substantial difference in the expenses being incurred startled the tech industry and spurred their interest to invest in DeepSeek rather than GPT or other major tech companies.
Why were the US stocks affected?
There’s long been a widely held notion that American AI companies such as OpenAI, Nvidia, and Microsoft would continue to reign over the tech market for years. This prevailing assumption was soon to be contested with the launch of China’s DeepSeek’s R1 model into the market, posing as a new competitor to the U.S. tech giants in the AI industry. China's expansion into the tech market signaled concerns about market dominance and the AI framework being transformed for American tech companies.
These apprehensions were affirmed when Nvidia's market value saw a significant decline as DeepSeek advanced its technology. The recent optimisation to its models enabled them to operate on China-made AI chips and work towards reducing reliance on massive computing power with lesser cost. DeepSeek trained its models using approximately 2,000 Nvidia H800 GPUs, a modified version of the H100 chip created specifically to comply with U.S. export regulations for China. These new developments ignited concerns about China gaining the technology to create its own AI hardware ecosystem, which could potentially restrict Nvidia’s future sales of their top-tier graphics processing units (GPUs) and growth in tech markets.
Moreover, this fear, along with DeepSeek having a low-cost approach, led to financial markets being disrupted on 27 January, making the stocks of chip makers and data centres fall in leading companies. Most investors sold off their stocks after reevaluating future growth prospects and capital expenditure for Nvidia, which caused the company to lose billions of dollars from its market valuation. According to the Bloomberg report, Nvidia’s shares dropped by 17%, wiping off $589 billion from the company’s market capitalisation. Major tech heavy companies suffered from Nvidia’s sell off, including Nasdaq’s index fell more than 3%, Dell lost $12.4 billion, Google cofounders Larry Page suffered a $6.3 billion reduction, and Sergey Brin $5.9 billion. These financial losses demonstrated the beginning of a new era for the AI industry where the United States may no longer have market leadership.
Kathleen Brooks, research director at XTB, noted that the impact of the concerns on the growing AI skills of the Chinese tech companies hindering U.S. tech dominance was also illustrated in semiconductor stocks, with Broadcom dropping 16% and ASML in Netherlands falling 6.7%, signaling the market’s reaction to the rising expansion of China’s AI.
There is a rise in global competition between AI companies, particularly American tech giants and emerging Chinese, that is contesting the established dominance of the Americans in the tech market
Nevertheless, given the heightening tensions between the U.S and China on Trump’s tariffs policy and the U.S ban on exporting AI chips to China, DeepSeek's growth in Western tech markets could be very restricted. Moreover, most countries and organisations worldwide have prohibited the usage of DeepSeek by referring to it as a security and privacy risk due to its user data being stored in China, giving rise to a potential risk of data leak to the Chinese government. For instance, the New York State has banned the use of DeepSeek on government electronics for protecting data privacy and ensuring security against the Chinese government. Michael Wooldridge, a professor of the foundations of AI at the University of Oxford, remarked that “ it was not unreasonable to assume data inputted into the chatbot could be shared with the Chinese state.” With these rising concerns of the Chinese government’s influence in DeepSeek’s program, it will be a challenge for the company to maintain its new competitor position.
Conclusion
With the growth of impressive options of AI, such as DeepSeek, the global AI industry’s landscape is transforming. There is a rise in global competition between AI companies, particularly American tech giants and emerging Chinese, that is contesting the established dominance of the Americans in the tech market. The formidable models of DeepSeek introduce a new cutting-edge open source technology that includes features such as Mixture-of-Experts models and advanced training techniques, which generate quality reasoning for various tasks at a reduced cost and with limited resources required. The Chinese tech market continues to benefit from a new surge in AI trading, while other countries remain adamant about laying down measures to stop Chinese expansion based on security concerns. Despite worldwide efforts to ban DeepSeek's expansion, the impact it has had on the global AI industry’s leadership and financial markets will certainly remain a turning point for the industry.