Today, DeepSeek, a rising star in China's AI field, has triggered an "earthquake" in the science and technology sector globally with its amazing speed and strength. This application, which is known as the "light of domestic AI", has not only pushed down the free list of the App Store in the U.S. region, but has also been recognized as the best AI application in the world. ChatGPT The rise of DeepSeek is not only the success of an app, but also symbolizes the strong rise of China's AI power in the global arena and heralds a profound change in the global AI competition pattern.
(coll.) fail (a student) DeepSeek and ChatGPT both ranked among the top three on the U.S. App Store's free charts, while Google's Gemini The fact that DeepSeek only ranked 33rd is not only a simple change in ranking figures, but also a profound reflection of the change in market and user attitudes towards different AI products. In the background of TikTok encountered the storm of shelving, DeepSeek can once again dominate the list of the U.S. application store, which is undoubtedly a strong proof of the influence of Chinese applications overseas. This is also confirmed by the rapid reaction of the market - Nasdaq 100 mini futures fell more than 300 points, reflecting the capital market's concern about the future competition in the U.S. AI industry, as well as the vigilance of DeepSeek as a strong rival.
The influence of DeepSeek is not only limited to overseas, but also set off a "screen" effect in China, which was quickly transmitted to the stock market. AI intelligent body, AI corpus and other conceptual stocks rose in response, and the capital market started to dig deep into the industry chain related to DeepSeek. From equity affiliates Zhejiang Dongfang and Huajin Capital, to computing power infrastructure providers Wave Information and Dawning of China Science and Technology, to vertical and technical partners Topsy, KDDI and Kingsoft Office, as well as data and ecological partners Daily Interactive and Zhuochuang Information, the ecological map of DeepSeek's partners is rapidly expanding, with huge commercial potential and social value behind it.
Feng Ji, the producer of Black Myth: Wukong, spoke highly of DeepSeek - "it could be a national technology achievement" - and elevated the significance of DeepSeek to the level of national strategy. He believes that DeepSeek can "make you get rid of the charm of the current state-of-the-art AI" and "take another solid step forward" in promoting knowledge and information affirmation. This is not an empty claim, but is based on DeepSeek's breakthrough innovations in technology, cost, openness, and other dimensions.
The six DeepSeek breakthroughs summarized by Feng Ji -Powerful, cheap, open source, free, connected, localThe "mysterious power of the East" is using these advantages to send an unprecedented shockwave through Silicon Valley, forcing top AI competitors to re-examine the market landscape. With these advantages, this "mysterious force from the East" is creating an unprecedented shakeup in Silicon Valley, forcing top AI competitors to reexamine the market landscape. The 40-minute feature story "How China's New AI Model DeepSeek Threatens U.S. Dominance" produced by the U.S. financial media CNBC, as well as more than 2 million views and tens of thousands of comments, fully confirm the huge attention and discussion DeepSeek has generated in the U.S.
In the comment section, there are reflections on the "threat theory", praise for DeepSeek's open source spirit, and reflections on OpenAI's closed mode. These comments reflect that in the new stage of global AI development, people have multiple thoughts about the path of technology development, business model and social responsibility, and the emergence of DeepSeek is not only a technological breakthrough, but also triggered a profound social discussion about the direction of AI development.
In the face of DeepSeek's strong attack, OpenAI and Meta and other giants have also reacted quickly: OpenAI founder Sam Altman announced that the ChatGPT free package will be available for o3-mini, and Meta was revealed to have urgently set up four combat teams to try to dismantle the technical secrets of DeepSeek's models. These actions show that DeepSeek has already posed substantial competitive pressure on the U.S. AI giants, forcing them to accelerate innovation and adjust their strategies.
Founded in May 2023, DeepSeek is known for its open source and low price strategy, and has been hailed as the "Poundland of AI" and "domestic AI price butcher". Its inference model R1, performance comparable to OpenAI o1 model, but the cost is greatly reduced, API price is much lower than o1, and provide free App and web version to use, but also support network search. This "universal AI" strategy allows more people to experience advanced AI technology at a low threshold, accelerating the popularization and application of AI technology.
- Open source is about much more than sharing technology: DeepSeek's open source means that users can download the largest version of the R1 model and run it on their local computers, which protects privacy and data security while providing unlimited possibilities for secondary development and customized applications. This spirit of openness, in the same vein as the early open source movement on the Internet, promises to build a more prosperous and healthy AI ecosystem.
- The Imagination of Mobile DeploymentDeepSeek R1 1.5B version can be deployed to run locally on cell phones, which signals that AI technology will accelerate the penetration of mobile, future mobile applications and services will be more intelligent, and the user experience will usher in a qualitative leap.
- The far-reaching implications of free accessThe free use of high-end models breaks the barriers of elitism and commercialization of AI technology, allowing a wider user group to access and use the most advanced AI technology, truly practicing the concept of "AI for everyone" and promoting the civilianization of AI technology.
DeepSeek has become the center of Silicon Valley's buzz in just a few days, and is seen as a "threat to American AI giants". Foreign media even put DeepSeek R1 's launch was compared to "China's ChatGPT moment," arguing that it "shatters the argument that the U.S. is invincible in the field of AI." Marc Andreeseen, founder of Silicon Valley's top venture capitalist a16z, hailed DeepSeek as "one of the most amazing and impressive breakthroughs I've ever seen" and "a far-reaching gift to the world as open source ". This is not just a commercial accolade, but a high recognition of DeepSeek's technical prowess and open source philosophy.
Technology leaders such as Stanford University's Department of Computer Science visiting professor Wu Enda and Microsoft Chairman Satya Nadella have also expressed concern and appreciation for DeepSeek's progress. Jim Fan, Senior Research Manager at NVIDIA, was more direct in stating that DeepSeek continues OpenAI's original mission of "truly open, cutting-edge research that empowers everyone." Likun Yang, Chief AI Scientist at Meta, also believes that Meta Chief AI Scientist Likun Yang also believes that DeepSeek's success demonstrates that open source models are surpassing proprietary models, and that open source will benefit everyone. These comments from the world's top tech people further подтверждает DeepSeek's status and influence in the global AI field.
DeepSeek R1 has not only made a breakthrough in technology, but its writing style has also quickly become popular on social media platforms. It can mimic different language styles and create well-written articles, showing strong content generation capabilities. Although there is still an illusion problem, it does not prevent it from becoming a new generation AI model with great potential, and DeepSeek R1's "poisonous comments" on competing products also show its humor and personalized side, which is closer to users and more humane.
DeepSeek's open source and ease of use has greatly energized the AI community. Using only the DeepSeek inference model, software composer Riley Brown created in under an hour the Perplexity AI The clone version of DeepSeek is not required to write a single line of code. This fully reflects the power and ease of use of DeepSeek, which lowers the threshold of AI application development and enables more people to participate in AI innovation.
Alex Cheema of AI lab EXO Labs connected multiple Mac mini and MacBook Pros to run DeepSeek R1 to build an "at-home AGI" system. This attempt demonstrates the flexibility and scalability of DeepSeek models, and signals that in the future, individual users will also be able to run more powerful AI models and realize more advanced AI applications through distributed computing.
Y Combinator CEO Garry Tan's experience with DeepSeek also confirms its reliability. He believes that DeepSeek searches are "more reliable" because "seeing the reasoning process greatly increases the user's trust". This "interpretability" and "transparency" is one of the key factors in DeepSeek's ability to win users' trust.
Sources on Blind, an anonymous Silicon Valley gossip platform, say Meta is in crisis over the success of DeepSeek, and The Information's report cites Meta employees who say Meta executives are concerned that the next-generation Llama model won't perform as well as DeepSeek, and that Meta has even launched four teams to dissect how DeepSeek works and plan to use the lessons learned to improve it. Meta even launched four battle groups to dissect DeepSeek's workings and plan to use the experience to improve the Llama model. It's clear that DeepSeek poses a serious challenge to a tech giant like Meta.
A statement from Meta spokesperson Jon Carvill, while emphasizing that Meta has been evaluating competing models and is excited about the prospects of Llama 4, side-stepped the impact of DeepSeek on Meta. Some AI companies are already considering a move to DeepSeek, such as Together AI and SuperFocus.Together AI has even doubled the ability to run DeepSeek models to meet customer demand, and many customers are also benefiting from Anthropic's Claude 3.5 Sonnet switches to DeepSeek model. superFocus has also indicated that it may switch to DeepSeek in the coming weeks to reduce costs and improve margins.
"The DeepSeek effect has not only shaken the U.S. stock market, but also given U.S. AI companies an unprecedented sense of urgency. China's AI companies have demonstrated that they can make top-level models with limited computing resources, which is certainly a powerful impact on the U.S. AI industry. OpenAI, with a valuation of $157 billion, will have its valuation and overhead questioned if it fails to maintain its leadership position and demonstrate a significant return on investment.
The U.S. is planning to invest heavily in AI infrastructure, and Trump's "Stargate Project" is planning to spend $3.6 trillion over the next four years to build AI data centers. Microsoft and Meta have also announced huge AI investment plans. However, the emergence of DeepSeek has raised questions about the utility of such huge capital expenditures. Journalist Holger Zschaepitz says that DeepSeek could be the biggest threat to the U.S. stock market, while Garry Tan argues that DeepSeek's success could be good for U.S. competitors, as it would boost the popularity and adoption of AI technology and accelerate the growth of computing demand.
DeepSeek founder Liang Wenfeng has said that DeepSeek bets on three directions: math and code, multimodal, and natural language, and hopes that China can gradually become a contributor to technological innovation instead of always hitching a ride.The emergence of DeepSeek R1 is undoubtedly an important breakthrough in China's AI technological innovation, which is rewriting the rules of the game of AI and, with the power of open source. The emergence of DeepSeek R1 is undoubtedly an important breakthrough in China's AI technological innovation, which is rewriting the rules of the AI game and, with the power of open source, stirring up a far-reaching hurricane in the global technology industry.
The volatility of the stock market may be short-lived, and the road to AGI is still long and full of challenges. However, the emergence of DeepSeek R1 and the Chinese open-source AI power it represents have injected new vitality and possibilities into global AI development. The hurricane kicked off by DeepSeek will have a profound impact on the future pattern of global AI and accelerate the pace of human society towards the age of intelligence.
English 
简体中文 
日本語 
Deutsch (Sie) 
Português do Brasil