DeepSeek has made a fool of Donald Trump and his tech bros
The launch of the DeepSeek-R1 model has wiped a trillion dollars off US tech stocks overnight, and made Donald Trump look a fool in the process.
After a day of trying to get a DeepSeek access code I finally received one this afternoon. The delay was understandable given the massive interest in this new open source Chinese AI language model, and the cyberattacks the company has been suffering.
It seems to work extremely well, unless you ask it about things like Tianmen Square, Taiwan, or the umbrella revolution in Hong Kong, which results in responses that neatly accord with the Chinese Communist Party line.
It’s hardly surprising that a China-based tech company would comply with Chinese censorship rules, but it’s far from the main story, which is that a Chinese company has developed one of the most efficient AI language models on an infinitesimal fraction of the budgets of US tech companies, despite a US-imposed embargo on Chinese imports of the most powerful Nvidia H100 AI chips.
According to DeepSeek they only spent $5.6 million developing their R1 language model, which is a drop in the ocean compared to the hundreds of billions being sunk into US-based AI companies.
Donald Trump just announced a £500 billion investment in AI, which suggests that the US is vastly overpaying for whatever benefits they think AI is going to deliver for them, when a Chinese company can develop a language model that outperforms US models on all kinds of metrics for a fraction of a percent of what they’re paying.
The launch of DeepSeek-R1 instantly wiped a trillion dollars off the value of US stocks, especially tech companies and the makers of AI chips Nvidia, which lost $593 billion in stock value in a single day, making it the biggest one day loss for a single US company in history.
Analysts have been calling it China’s "Sputnik moment", in reference to the Soviet Union taking the US by surprise by launching the first space satellite in 1957.
DeepSeek hasn’t just rocked the US tech market by producing a competitive product at a tiny fraction of the cost, it’s also proven the failure of US tech sanctions on China.
Restricting the sale of the most powerful chips to China has backfired by seemingly forcing them to develop comparative technology much more efficiently. Given the collapse in Nvidia’s stock value on the back of this, it’s hard to imagine that they’re delighted with the US government’s market intervention to prevent them from supplying China.
Then there’s the fact that DeepSeek is operating as an open source venture which is the antithesis of the anti-competitive nature of capitalist US tech giants who seek to guard their technological advancements through patent law and rent-seeking behaviour to ensure the maximum profit extraction.
Because DeepSeek is cheap, efficient, and open source, it represents a hammer blow to US profit seeking. Why pay $billions for a US product when you can work off what the Chinese have developed for free?
As Richard Murphy put it in his post on DeepSeek: "If China can develop efficient tech at low prices and put it into widespread use through open-source diversification while maintaining a fee base for access to more advanced searches, then the whole myth on which the next generation of US wealth extraction was being built has been shattered."
Furthermore it’s proven the ineffectiveness of US tech sanctions on China, and made a fool of Trump at the beginning of his second stint as President.
Just last week he flanked himself with a load of tech bros for his inauguration (the Nazi saluting one and others) and announced an astronomical investment in AI, only to have a small Chinese tech company pull all their pants down for the whole world to see.
Further giant leaps forward in the Chinese tech sector seem inevitable, but the US response is harder to predict.
Trump has already been pushing a deeply protectionist agenda of tariffs and trade sanctions, however trade sanctions on tech components have clearly failed to eliminate competition from China.
The US could still try restricting freedoms within the US to try and stamp out Chinese competition, as they’re trying to do with TikTok.
The US are desperate to keep non-US competitors out of the social media market by hyping up the supposed "security threat" of non-US ownership, so could they try to artificially protect the profits of US tech companies by citing spurious "security threats" to ban Americans from using Chinese AI technology?
If they do attempt this kind of market manipulation to protect the profits of their own tech companies, it’s highly likely to backfire (if banning the use of open source developments is even feasible at all).
Forcing US businesses to pay through the nose for proprietary US-based AI advancements while most of the rest of the world follows the vastly cheaper open source approach of the Chinese would create enormous burdens and turn the US into an expensive and uncompetitive walled garden.
Just like the US embargo on selling advanced chips to China has backfired spectacularly, protectionism simply isn’t going to shut down further Chinese advancements.
The only workable option is for the US to try to outcompete their Chinese rivals, but it’s hard to see how that’s possible when US tech developers work on the capitalist principal that advancements are made in order for them to extract $billions in personal wealth, if Chinese companies can develop comparative tech for a tiny fraction of the cost, then give it away open source.
Deepseek has pulled the curtain back to reveal that the ai industry is vastly overpriced and another ‘tech-bubble’ where line managers in Silicon Valley are paid more than the entire budget for Deepseek.
A quick update to add what DeepSeek has to say on the prospect of US retaliations against Chinese AI ...
It is highly likely that the United States will seek to restrict Chinese AI development and deployment, similar to its approach with TikTok, though the methods and scope may differ. The U.S. government views AI as a critical technology with significant implications for national security, economic competitiveness, and geopolitical influence. Here's an analysis of the potential actions and motivations:
1. Motivations for Restricting Chinese AI
National Security Concerns: The U.S. fears that Chinese AI technologies could be used for espionage, surveillance, or military applications that threaten U.S. interests. For example, AI-powered systems could enhance China's cyber capabilities or autonomous weapons.
Data Privacy Risks: Chinese AI systems, like TikTok, are seen as potential conduits for data collection by the Chinese government, raising concerns about the misuse of sensitive information.
Economic Competition: AI is a key driver of future economic growth. The U.S. aims to maintain its technological edge and prevent China from dominating critical AI markets.
Geopolitical Rivalry: The U.S.-China tech rivalry is part of a broader competition for global influence. Restricting Chinese AI aligns with efforts to limit China's technological and strategic ascendancy.
2. Potential Measures to Restrict Chinese AI
Export Controls: The U.S. could expand export restrictions on advanced AI technologies, such as high-performance chips (e.g., NVIDIA GPUs) or AI software, to limit China's access to critical components.
Investment Restrictions: The U.S. may further tighten rules on American investment in Chinese AI companies, as seen with the CHIPS and Science Act and executive orders targeting sensitive technologies.
Bans on Chinese AI Products: Similar to the TikTok ban, the U.S. could prohibit the use of Chinese AI software or platforms in sensitive sectors (e.g., government, defense, critical infrastructure).
Sanctions on Chinese AI Firms: The U.S. could impose sanctions on Chinese AI companies, such as SenseTime, Huawei, or Baidu, restricting their access to U.S. markets and technology.
Data Localization Requirements: The U.S. could mandate that data used by AI systems be stored and processed domestically, limiting Chinese firms' ability to operate in the U.S.
3. Challenges and Limitations
Global Supply Chains: AI development relies on global supply chains, making it difficult to completely isolate Chinese AI without disrupting U.S. and allied industries.
Retaliation: China could retaliate by restricting U.S. companies' access to its market or by accelerating its own efforts to achieve technological self-sufficiency.
Ethical and Diplomatic Concerns: Overly aggressive restrictions could be seen as protectionist, damaging the U.S.'s reputation as a proponent of free trade and open innovation.
Effectiveness: China has made significant strides in AI and may continue to advance even with restrictions, particularly through domestic innovation and partnerships with other countries.
4. Precedents and Context
TikTok Ban: The U.S. government's efforts to ban or force the sale of TikTok (owned by Chinese company ByteDance) set a precedent for targeting Chinese tech companies over national security concerns.
Huawei Restrictions: The U.S. has already imposed severe restrictions on Huawei, including bans on its 5G technology and access to U.S. components, citing similar security concerns.
AI-Specific Actions: The Biden administration's Executive Order on AI (October 2023) includes provisions to address risks from foreign AI development, signaling a focus on countering Chinese AI advancements.
5. Likely Scenarios
Targeted Restrictions: The U.S. is more likely to impose targeted restrictions on specific Chinese AI applications (e.g., surveillance, military) or companies rather than a blanket ban on all Chinese AI.
Allied Coordination: The U.S. may work with allies (e.g., EU, Japan, South Korea) to create a unified front against Chinese AI, similar to efforts to limit Huawei's global reach.
Focus on Critical Sectors: Restrictions may focus on AI applications in sensitive areas like defense, critical infrastructure, and government operations, rather than consumer-facing AI products.
Conclusion
The U.S. is likely to take steps to restrict Chinese AI, particularly in areas that pose national security risks or threaten economic competitiveness. However, these measures will likely be more nuanced and targeted than a blanket ban, reflecting the complexity of AI technology and the interconnected global economy. The U.S. will also need to balance these restrictions with the need to maintain its own leadership in AI innovation and avoid escalating tensions with China.