Portuguese 
English 
Spanish 
3 min of reading 
0 comments 
AI Revolution! Chinese Company Creates Technology That Surpasses ChatGPT With Surprisingly Low Investment
The recent rise of China in the artificial intelligence (AI) sector is raising concerns in Silicon Valley. DeepSeek, an previously unknown Chinese AI lab, surprised the world by launching a large language model that is free and open-source open.
Despite being built on a budget of under US$ 6 million and using lower-capacity Nvidia H800 chips, the model has managed to surpass heavyweight competitors from the United States, such as Meta Llama 3.1, OpenAI’s GPT-4, and Anthropic’s Claude Sonnet 3.5. This turnaround is putting U.S. global leadership in the sector at risk.
The Technological Advancement of DeepSeek
In benchmark tests conducted by third parties, DeepSeek’s model excelled in various areas such as solving complex problems, mathematics, and coding.
-
Mercor paid $1.5 million per day for doctors, lawyers, and former Goldman Sachs bankers to teach artificial intelligence to do their jobs, and in 17 months, it went from zero to $500 million in annual revenue while its own contractors accelerated the replacement of their own work.
-
Casio Unveils Moflin, Robotic Pet With Artificial Intelligence Designed To Provide Emotional Comfort And Simulate A Permanent Affectionate Bond
-
Google Maps Revolutionizes GPS Navigation With Gemini Artificial Intelligence, Ask Maps Feature, and Advanced 3D Visualization for Route and Travel Planning
-
China Warns the U.S. of a “Terminator” Style Apocalypse: Military AI Rivalry, Sanctions Against American Startup, and Fear of Algorithms Deciding Who Lives or Dies Place the World on the Brink of a Nightmare That Seemed Like Fiction
Additionally, the company recently announced r1, a reasoning model that also surpassed OpenAI’s o1 model in several of these tests.
“Seeing DeepSeek’s new model is super impressive in terms of how they actually made an efficient open-source model in terms of inference time computing“, stated Satya Nadella, CEO of Microsoft, during the World Economic Forum in Davos.
Chinese Challenges and Solutions in Light of U.S. Restrictions
DeepSeek faced significant challenges due to the stringent semiconductor restrictions imposed by the U.S., which prevented China from accessing more powerful chips, such as Nvidia’s H100s.
However, the company’s advancements suggest that it has found creative ways to circumvent these barriers, possibly through techniques like model distillation.
This process allows a smaller model to learn from a larger model, reducing costs and optimizing computational efficiency.
“They can take a really good and large model and use a process called distillation to train a smaller model, making it smarter in a very cost-effective way“, explained Chetan Puttagunta, a partner at Benchmark.
The Role of the Hedge Fund and the Mystery Surrounding DeepSeek
Little is known about the internal structure of DeepSeek and its founder, Liang WenFeng. The company emerged from the hedge fund High-Flyer Quant, which manages approximately US$ 8 billion in assets.
The lack of transparency from the company has sparked speculation about its true potential and future plans.
Other Chinese Initiatives in AI
DeepSeek is not the only Chinese company making strides in the global AI landscape.
The startup 01.ai, founded by renowned researcher Kai-Fu Lee, managed to train an AI model with just US$ 3 million.
Moreover, ByteDance, the parent company of TikTok, recently announced an upgrade to its AI model, which also surpassed OpenAI’s o1 in important benchmarks.
“Necessity is the mother of invention“, stated Aravind Srinivas, CEO of Perplexity. “As they had to find alternative solutions, they ended up building something much more efficient.”
Impacts for the Future of Global AI
The rapid advancement of China in AI raises questions about the sustainability of massive investments by American giants in infrastructure and AI models.
If Chinese companies can develop competitive models with fewer resources and less advanced hardware, Western firms will need to reconsider their strategies to maintain leadership.
For the United States, this scenario serves as a wake-up call regarding the need to review its technology export restriction policies.
Furthermore, the efficiency of Chinese solutions could lead to a redefinition of research and development priorities in the West.
The emergence of DeepSeek as a threat to U.S. dominance in AI serves as a reminder of the dynamism of this sector.
China’s ability to develop high-performance models on reduced budgets challenges existing assumptions about the hardware and investment requirements needed for significant advancements.
As a result, companies and governments need to closely monitor China’s advancements and seek innovative solutions to maintain competitiveness in the global AI market. The next decade promises to be crucial in determining who will lead this technological revolution.
Seja o primeiro a reagir!