We are witnessing the beginning of a new conflict: the technological cold war.
If those who hold the knowledge hold the power, should we be concerned when the world’s greatest powers compete for dominance in artificial intelligence?
By Paulo Cabral on February 2025
Whoever controls the Artificial Intelligence industry and manages to consolidate the most widely used language model and the standardization of today’s networks will undoubtedly be the most powerful. Since this industry has no cooperation or spirit of development, what is left then? What exists is an uphill battle supported by billions of dollars, whose sole objective is to consolidate the most standardized AI ecosystem possible. For this, it depends on the innovations and developments in technology, business, health, education, finance, and government.
While OpenAI, Nvidia, Google, and some of the main AI companies have built a discourse around the high cost and computing power needed to consolidate this technology, China has come to break everything with DeepSeek: it was created with a fraction of the money and a minimal part of the technology of ChatGPT, in addition to being open source and free to the world.
Round 1
DeepSeek marked the beginning of a technological cold war between China and the United States. Both nations are racing to develop general Artificial Intelligence that not only surpasses human intelligence but can also tackle humanity’s most pressing questions and problems. However, this pursuit comes with a significant risk: the possibility of the technology becoming uncontrollable.
What unfolds next is a significant conflict: OpenAI, Nvidia, and other leading technology companies attribute the chaos to DeepSeek. Taiwan, with its burgeoning microchip and processor industry, emerges as the crown jewel in this new historical chapter. The United States responds by blocking DeepSeek and any associated AI, imposing sanctions on countries that utilize or integrate the technology. Indeed, it triggers a domino effect.
Many are asking: why all the panic? The answer lies in the potential implications: if China has indeed discovered a more cost-effective and efficient method for developing artificial intelligence, Silicon Valley’s massive investments in hardware and supercomputers might be rendered obsolete. This scenario spells trouble for the American economy.
The crisis
Recently, a seismic event disrupted international markets and reshaped the geopolitical landscape of artificial intelligence. A Chinese company announced the development of an AI model for a mere US$6 million, capable of rivaling American giants such as ChatGPT, Meta, and Google—companies that have each invested over US$1 billion in their respective AI advancements. To add to the upheaval, DeepSeek made its model open source, enabling anyone to modify and adapt it freely.
The White House has swiftly responded to the news, with spokeswoman Karoline Leavitt announcing that the National Security Council is assessing the strategic implications of DeepSeek. For President Donald Trump’s administration, the emergence of this new Chinese competitor is a “wake-up call” for American industry. In a speech to Republican congressmen, President Trump emphasized the urgent need for the U.S. to reclaim its leadership and supremacy in the sector.
Experts have likened the rise of DeepSeek to a “Sputnik moment,” referencing the profound impact the launch of the first Soviet satellite had on the U.S. during the Cold War space race (1947-1991). The fact that a supposedly small Chinese startup managed to develop an efficient AI model without the vast resources available to large American companies has raised doubts among investors about the true technological advantage of the U.S. Consequently, this development has significantly affected the market value of American tech giants.
A global risk factor
Some global consequences have already been observed. The Australian government has banned the use of the Chinese AI model in government devices, citing an “unacceptable risk” to national security. Similarly, other government institutions in various countries, including Taiwan, Italy, and even the U.S. Navy, have restricted the installation and use of this technology.
Countries are increasingly concerned about information leaks and security breaches, which have already taken place. It is essential to understand what personal data is collected, the sources of this data, the purposes for its use, the legal basis for processing it, and whether the data is stored on servers in China. DeepSeek is under suspicion for using user data to train its AI and for lacking a defined legal basis, thereby violating principles of transparency and information obligations.
In light of China’s technological advancements and censorship, the West appears to face two options: the first, a purely technological approach accompanied by censorship; and the second, a technological path grounded in values, allowing for the free circulation of different versions in the West, including participation from China, but still accepting censorship.
Round 2
This would be an uneven struggle, a kind of “cultural battle,” waged differently in the West and China. Historical experiences with communist countries show that they do not easily yield on ideological and cultural grounds, viewing the defense of their principles as a dam that must not crack, because eventually it will break—but it always does. What will the West accept in this new scenario?
One of the primary concerns regarding DeepSeek is the potential use of AI for state control and surveillance. Developed in the Chinese city of Hangzhou, the technology is backed by a robust digital surveillance system aimed at sustaining the power of the communist regime. Additionally, there are apprehensions about the collection and storage of data by the AI, as the company’s privacy policy explicitly states that user information is stored on servers in China, under the jurisdiction of the Xi Jinping regime.
Although DeepSeek is open source, there are concerns about censorship mechanisms built into the technology. While Western models, such as ChatGPT, are criticized only for ideological biases and restrictions on certain topics, DeepSeek has even stricter prior censorship, which filters out information that goes against the interests of the Beijing regime (such as topics about the Tiananmen Square Massacre, Taiwan independence, and criticisms and nicknames made of that communist dictator).
Are the costs worth it?
Financial firms have warned that the emergence of DeepSeek could trigger a reassessment of billion-dollar investments in AI. In the short term, markets should brace for a period of volatility as investors reevaluate the value of AI-related companies, including giants such as Nvidia.
In 2024, U.S. companies like OpenAI, Google, and Meta spent hundreds of billions of dollars developing new AI models. Now, investors are questioning whether these high costs are still justifiable in the face of supposedly cheaper competitors.
Adding to the crisis, OpenAI has announced an investigation into whether the Chinese startup trained its new chatbot using the artificial intelligence model previously created by the Silicon Valley-based company. In an official statement, OpenAI revealed that in recent months, it has detected several attempts by Chinese entities to extract large volumes of data from its AI tools2. “It is vitally important that we work closely with the U.S. government to better protect our most capable models from efforts by rivals and competitors to seize American technology,” the document reads. It also directly accuses China-based companies of “constantly attempting to distill the models of leading U.S. AI companies”
The winner is…
In this way, the rivalry between the United States and China in the domain of AI reflects a new technological cold war, in which both nations seek to consolidate their supremacy in this strategic field. Trump’s recent policies and DeepSeek’s advances illustrate the intensification of this contest, with significant implications for the future of technology and the global economy.
The AI race is a battle for influence and power. AI has the potential to transform economies, military strategies, and even global narratives. Unlike the nuclear arms race, there is no clear end. The crucial question is not simply who will win, but how this competition will shape our future. Will it inspire unprecedented innovation and collaboration, or deepen existing divisions? The AI race is on now, and its outcome will define the 21st century.
Learn more: DeepSeek vs. ChatGPT
While DeepSeek and ChatGPT both represent state-of-the-art AI language models, they take different approaches to solving similar problems. Let’s break down their key differences in a way that matters.
Model Architecture
At their core, the intelligences work differently. DeepSeek activates only a subset of its parameters for each request, increasing efficiency and optimizing performance and resource usage by dynamically adapting to different types of queries. In contrast, ChatGPT uses a traditional transformer model, which is more consistent but potentially less efficient.
Performance Strengths
DeepSeek has demonstrated impressive capabilities in technical tasks. ChatGPT, however, demonstrates a greater ability to understand context and provide more nuanced responses across a wider range of topics.
Accessibility and Cost
DeepSeek takes an open-source approach, meaning it is freely available and can be modified by the community. ChatGPT operates on a freemium model, offering basic features for free but requiring a subscription for advanced features.
Customization and Ease of Use
DeepSeek offers more extensive customization options. However, this comes with a steeper learning curve and requires some technical knowledge. ChatGPT prioritizes ease of use, offering a more polished experience that is accessible to those beginning their data science journey.
Development Philosophy
DeepSeek’s development approach emphasizes efficiency, having been built using innovative training methods and less powerful hardware. This demonstrates how clever engineering can sometimes overcome resource constraints. ChatGPT, powered by substantial computational resources, takes a more traditional approach to its capabilities.
Writing Assistance
ChatGPT excels at producing engaging, conversational content with rich context—perfect for explaining complex data concepts to non-technical stakeholders. DeepSeek, on the other hand, shines in technical writing scenarios, producing precise, formal documentation that is particularly valuable for data design documentation and technical specifications.
Coding and Technical Consultations
ChatGPT offers comprehensive code assistance, providing detailed explanations alongside code suggestions. DeepSeek takes a more direct approach, with faster code generation, useful when one needs efficient solutions to specific coding challenges.
Brainstorming and Creativity
ChatGPT excels at generating multiple approaches to a problem, helping one explore multiple analytical possibilities. DeepSeek typically offers fewer but more developed solutions, delving deeper into a single approach.
Learning and Research
ChatGPT provides comprehensive, tutorial-style explanations that work well for learning new concepts. DeepSeek focuses more on accuracy and conciseness.
Cost and Efficiency
DeepSeek stands out for its cost-effectiveness, using energy-efficient hardware and edge implementations to keep operating costs low. Since it’s free to use, it’s a great resource if one is working on a budget. ChatGPT’s subscription model, while more expensive, offers consistent performance and advanced features that can be valuable for professional data work.
Privacy and Ethical Concerns
ChatGPT adheres to Western data protection standards, making it a safer option for projects that require strict data privacy compliance. DeepSeek’s data storage practices and content moderation policies may raise concerns for certain types of projects, especially those involving sensitive information or requiring unrestricted analytical discussions.