{"id":20624,"date":"2025-01-27T08:43:00","date_gmt":"2025-01-27T11:43:00","guid":{"rendered":"https:\/\/lantyer.com.br\/?p=20624"},"modified":"2025-01-27T08:44:55","modified_gmt":"2025-01-27T11:44:55","slug":"deepseek","status":"publish","type":"post","link":"https:\/\/lantyer.com.br\/en\/deepseek\/","title":{"rendered":"How US Trade Sanctions Boosted Chinese AI Innovation: The Case of DeepSeek"},"content":{"rendered":"

Introduction<\/h3>\n\n\n\n

<\/p>\n\n\n\n

The restrictions imposed by the United States on the sale of cutting-edge chips to China have generated significant repercussions in the global technology landscape. From an international law perspective, the objective of these sanctions is usually to protect US interests and national security, while also attempting to curb the technological advancement of potential strategic competitors.<\/p>\n\n\n\n

However, history shows that commercial and technological obstacles often serve as catalysts for high-impact homegrown solutions. This is exactly what happened with the emergence of DeepSeek<\/strong>, a Chinese artificial intelligence that has been shaking up the market by delivering competitive results with more modest resources.<\/p>\n\n\n\n

\n\n\n\n

1. The Genesis of the Problem: Technological Constraints and Bottlenecks<\/h2>\n\n\n\n

In recent years, Washington has adopted a policy of increasingly restricting the export of advanced components to Chinese companies, especially high-performance processors and GPUs (such as Nvidia's H100 and A100 series). The legal rationale behind these measures is based on domestic and international concerns, aiming to prevent sensitive technologies from being used in projects that contradict US interests.<\/p>\n\n\n\n

However, such trade barriers ultimately spurred China's search for domestic solutions. In addition to the limitations imposed on cutting-edge chips, there was increased inspections of Chinese companies and even a ban on certain GPU models by the US government. This created a bottleneck that, rather than paralyzing innovation, motivated Chinese startups and universities to invest heavily in research and development of more efficient algorithms capable of running on less sophisticated hardware.<\/p>\n\n\n\n

<\/span>\"Deepseek
\n

<\/p>\n<\/div><\/div>\n\n\n\n

\n\n\n\n

2. DeepSeek: The \u201cChild\u201d of External Pressure<\/h2>\n\n\n\n

THE DeepSeek<\/strong>, a relatively unknown startup until recently, exemplifies this scenario of overcoming constraints. The company recently launched an AI model\u2014the DeepSeek-V3<\/strong> \u2014 which, according to the team itself, was trained at a total cost of less than US$1,400,000, using H800 chips, which are considered outdated compared to Nvidia's most advanced models.<\/p>\n\n\n\n

These H800s were initially designed as "hardened" versions for the Chinese market to comply with US export regulations, and were later banned by further rounds of sanctions. Even using less powerful hardware, the DeepSeek team developed an innovative training approach: parameter optimization, intelligent data selection and leaner algorithms<\/strong>.<\/p>\n\n\n\n

The result was an AI that surprised experts and consumers alike, achieving performance comparable to (or even superior to) established chatbots like ChatGPT and Claude. It's no coincidence that DeepSeek's assistant became the highest-rated free app on the US App Store.<\/p>\n\n\n\n

\n\n\n\n

3. Impacts on Markets and Geopolitics<\/h2>\n\n\n\n

3.1. Collapse in the Big Tech Euphoria<\/h3>\n\n\n\n

DeepSeek's sudden success shook the markets. Investors' distrust of the real "need" for ultra-advanced chips for AI sent shares of giants like Nvidia and companies related to the semiconductor supply chain tumbling.<\/p>\n\n\n\n

Analysts fear that if the idea that "less is more" takes hold, the pace of high-performance GPU sales and data center expansion could slow. This would undermine one of the technology sector's strongest growth theses in recent years.<\/p>\n\n\n\n

3.2. Pressure on Washington and Sanctions Reviews <\/h3>\n\n\n\n

From an international law perspective, the episode highlights a side effect of trade restrictions: the unintentional encouragement of self-sufficiency in the sanctioned market. If Chinese companies maintain a strong and sustainable pace of innovation, Washington could find itself facing an even greater geopolitical impasse.<\/p>\n\n\n\n

Not only have sanctions not stopped Chinese technological advancement, they have potentially encouraged it.<\/p>\n\n\n\n

3.3. Regulatory and Intellectual Property Issues<\/h3>\n\n\n\n

There are complex legal implications. The adoption of open-source development strategies by DeepSeek and other Chinese initiatives, in addition to reducing costs, enables international collaborations beyond the reach of sanctions.<\/p>\n\n\n\n

The prospect of an open, globalized ecosystem contrasts with heavily funded projects in the US, such as the partnership between OpenAI and Stargate (a data center joint venture that has attracted billions of dollars). In this scenario, patents, usage rights, and licensing agreements may be challenged in international arbitration forums and organizations such as the World Trade Organization (WTO).<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n
\n\n\n\n

4. Future Perspectives and Lessons Learned<\/h2>\n\n\n\n
    \n
  1. Innovation from Adversity<\/strong>
    Historically, restrictive policies tend to encourage domestic solutions in target countries. The "DeepSeek case" reinforces the idea that the more barriers imposed on the import of hardware or software, the greater the motivation to advance local research and creative optimizations.<\/li>\n\n\n\n
  2. Corporate Strategy Review<\/strong>
    Companies and investment funds around the world are looking at DeepSeek and wondering whether so much computing power is really needed to develop world-class AIs. This could trigger an "efficiency race," where algorithms become increasingly cleaner, more compact, and less dependent on expensive infrastructure.<\/li>\n\n\n\n
  3. Challenges for Global AI Governance<\/strong>
    The success of low-cost Chinese AI is sparking new debates in international forums about safety standards, ethics, and data sharing. While the US and Europe debate AI regulations, China is advancing technologies that can be rapidly adopted in other jurisdictions, including emerging countries lacking the resources to stay at the forefront of cutting-edge hardware.<\/li>\n\n\n\n
  4. Balance between National Security and Free Market<\/strong>
    From an international law perspective, there is a constant clash between the protection of strategic interests (national security, technological supremacy) and free trade obligations. Tensions between the US and China make it clear that, in a scenario of growing multipolarity, sanctions can have the opposite effect to those intended.<\/li>\n<\/ol>\n\n\n\n
    \n\n\n\n

    Conclusion <\/h2>\n\n\n\n

    The recent history of DeepSeek<\/strong> shows that trade sanctions can, paradoxically, act as a trigger for under-the-radar innovations. Deprived of cutting-edge chips, Chinese teams had to refine their algorithms and models to the maximum, producing solutions that were surprisingly efficient.<\/p>\n\n\n\n

    In the geopolitical chess game of the digital age, this is yet another chapter that illustrates how external pressures don't always contain competition\u2014they often end up strengthening it. In this scenario, the global AI market is likely to rebalance, forcing Western companies to reevaluate their investments and business models. From a legal perspective, new uncertainties loom over the true effectiveness and long-term impact of trade restrictions.<\/p>\n\n\n\n

    Meanwhile, DeepSeek, born of adversity, is taking its place and redefining what can be done with little\u2014perhaps ushering in a new phase of \u201clean\u201d AI that\u2019s accessible to the entire world.<\/p>\n\n\n\n

    <\/p>\n\n\n