In the ever-evolving landscape of artificial intelligence, China's DeepSeek model has emerged as a significant player, stirring both excitement and skepticism within the investment communityThe initial trading day of the year revealed the ripple effects of this development, with shares across the AI computing sector, including power chip manufacturers, optical communication firms, and AI server producers, experiencing considerable declinesFor instance, the stock of Cambricon Technology fell nearly 10%, while ZTE Corporation dropped over 6%. The impact on the optical communication sector was equally stark, as companies like Zhongji Xuchuang and Tianfu Communication both saw their stocks plunge by over 10%.
This downward trend in stock prices is reflective of the broader uncertainty in the market regarding the underlying assumptions driving investments in AI computingThe entry of DeepSeek, with its impressive capabilities, has cast doubt on the previous notion that merely ramping up computational power would naturally yield improved models and returns
Advertisements
As a result, investors are re-evaluating their strategies in light of DeepSeek’s novel and efficient approach to AI model development.
The impact on the AI computing industry is palpableCambricon Technology, recognized as a leader in the Chinese AI chip market, experienced a significant slide in share value, crashing from a closing price of 612.98 yuan to 551.68 yuan in a single trading dayMeanwhile, ZTE's foray into AI chip design—a field where it has positioned itself as a dominant player—suffered as its stock dropped more than 6%. Such sharp declines reflect a growing discontent among investors, who are grappling with a rapidly shifting landscape where traditional powerhouses in AI may no longer hold the coveted status they once did.
The optical communication sector plays a critical role in optimizing the efficiency of AI computations, especially in large-scale AI operations
Advertisements
Companies within this field have already begun supplying high-performance optical communication products to leading international AI chip manufacturersNotably, Zhongji Xuchuang has carved out a niche in catering to overseas clients, while also being a major supplier in ChinaHowever, even these firms are not immune to market turbulence; Zhongji Xuchuang’s stock recently plunged by over 10%, and Tianfu Communication took an even harder hit, resulting in an 11% decline at one point.
AI server manufacturing companies are also feeling the heat, with notable players like Inventec experiencing a more than 6% drop in stock pricesThis is particularly surprising given Inventec’s strong performance in both AI server and high-speed networking equipment sectors, executed in tandem with rising demands for AI infrastructureThis pattern underlines the interconnectedness of supply chains within the AI ecosystem, where the fate of one sector can heavily influence another.
Interestingly, various stocks in the sector had previously seen explosive growth
Advertisements
For example, Cambricon reached a historic peak of 777.77 yuan per share, while ZTE achieved its highest price since July 2024, hitting 43.75 yuanLikewise, Tianfu Communication had a surge in its stock, reaching a ceiling limit on January 16. That those stocks soared before this downturn is indicative of market volatility and investor expectations of aggressive returns through the continued evolution of AI technologies.
As concerns rise regarding the sustainability of traditional AI computing paradigms, the emergence of DeepSeek presents a compelling counter-narrativeDeveloped by a company known as "Quantitative Mirage," DeepSeek has captivated the attention of tech enthusiasts and professionals alike, presenting a cost-effective path toward achieving advanced AI capabilitiesSpecifically, the latest iteration, DeepSeek-V3, was trained with only 2.788 million H800 GPU hours at a cost of $557,000, yet it rivals the performance of OpenAI’s state-of-the-art GPT-4o model
- US Stocks Continue to Rise in January
- Nasdaq Surges Over 2%
- US-Canada Trade Spat: Will Oil Prices Surge?
- Bank of Japan Resumes Interest Rate Hikes
- FOF Asset Allocation: Strategies for Steady Progress
Notably, such capabilities are typically associated with resources that may demand upward of 16,000 GPUsIn contrast, deep learning giants recently launched networks involving close to 100,000 GPUs.
Furthermore, the pricing for DeepSeek’s API hints at a promising future that could reshape the dynamics within the AI communityThe costs resemble a fraction of those incurred with models like GPT-1, illustrating how DeepSeek can deliver competitive performances at a significantly reduced investment thresholdFor context, the pricing for each million tokens generated by DeepSeek-R1 is just $2.19, compared to GPT-1’s $60. This stark contrast poses pivotal questions: Is heavy financial investment truly necessary to produce higher-performing models, or can advancements in algorithmic innovation and software architecture pave the way for breakthroughs in AI development?
This development invites industry experts and analysts to reconsider how investments, computational power, and performance are intertwined
Wang Zhi, a partner at Weichuangxin, emphasized that while the market may be reacting negatively in the short term due to emotional trading sentiment, the fundamental role of AI computing resources will remain essential in the long runHe suggested that future AI model advancements will necessitate innovative algorithms and architectures, laying the groundwork for a robust ecosystem driven by potential breakthroughs.
In conclusion, the trajectory of DeepSeek could redefine the competitive landscape within the AI sector, challenging the norms surrounding investment in AI computational resourcesAs investors digest these outcomes, the pressing question looms: how will they adapt their strategies in response to this paradigm shift? It is vital for stakeholders to consider the implications of DeepSeek’s efficient modeling approach, and how it may set new benchmarks for future AI developments and investments