A Chinese AI, DeepSeek, has shaken the world. American tech companies and the US stock market have been shocked to an extent they couldn't have imagined.
 |
| China's New AI Model Deepseek |
"The release of DeepSeek AI from a Chinese company should be a wake-up call."
On January 20, 2025, a Chinese research lab launched its AI chatbot named DeepSeek R1. Along with it, they published a research paper stating that this chatbot surpasses the most advanced chatbots worldwide in benchmarks like math and reasoning. This includes OpenAI's ChatGPT o1 model, Meta's Llama, and Google’s Gemini Advanced. DeepSeek has outperformed all of them.
DeepSeek is the best in terms of performance and efficiency, and it was developed at a fraction of the time and cost. The most surprising part? It is completely free for all users. On the other hand, OpenAI charges $200 per month for access to its ChatGPT o1 Pro model.
Moreover, the cost of training DeepSeek was reported to be only $5.6 million, whereas American companies like OpenAI, Meta, and Google spend billions of dollars on their AI models.
Coincidentally, a while back, when OpenAI’s founder Sam Altman was asked whether foundational AI models like ChatGPT could be made in India with just $10 million, he arrogantly replied that it would be "hopeless."
"It's totally hopeless to compete with us in training foundational models. You should try, though; it’s your job to try anyway."
Today, Sam may be feeling hopeless himself because, just a week after DeepSeek’s launch, it became the most downloaded app in the US on the App Store and Google Play Store, surpassing ChatGPT. The next day, it became the number one app in India and other countries. By January 27, it had disrupted the American financial markets.
"The launch of DeepSeek, a Chinese-built chatbot, immediately rattled investors and wiped out a staggering $1 trillion off the US tech index."
Before DeepSeek’s launch, NVIDIA was the most valuable company in the world, valued at $3.5 trillion. However, in just one day, its valuation dropped to $2.9 trillion. NVIDIA, which specializes in manufacturing computer chips for training and operating AI, saw its shares fall by 17% in a single day, losing $589 billion in valuation—the biggest loss in corporate history. The Nasdaq tech index dropped by 3.1%.
The Story Behind DeepSeek
The credit for developing DeepSeek goes to a 40-year-old Chinese entrepreneur, Liang Wenfeng. He rarely makes public appearances and prefers to keep his identity hidden. In 2015, he founded a hedge fund called High Flyer, which used AI and mathematics for investments. In 2019, he established High Flyer AI to research artificial intelligence algorithms. By May 2023, using his hedge fund earnings, he began a side project to create an AI model purely out of scientific curiosity, with no profit motives.
Instead of hiring engineers, Liang recruited PhD students from China’s top universities. He trained his AI model using the most difficult questions worldwide. Within just two years and a few million dollars, he launched the DeepSeek R1 model.
Only 200 people were involved in creating DeepSeek, 95% of whom were under 30 years old. In contrast, OpenAI has over 3,500 employees. DeepSeek is also China’s only AI firm that is not funded by major tech giants like Baidu, Alibaba, and ByteDance.
DeepSeek’s Advanced Reasoning Model
DeepSeek operates as a "Chain of Thought" model, similar to OpenAI's ChatGPT o1. This model mimics human reasoning by analyzing multiple perspectives before answering a question.
For example, if you ask ChatGPT-4o, "Which is bigger: 9.11 or 9.9?" it might instantly respond incorrectly. However, ChatGPT o1 and DeepSeek first think critically and counter-question themselves before answering correctly. DeepSeek, like ChatGPT o1, enhances logical accuracy significantly.
DeepSeek also displays its thought process step by step before answering. If asked an original insight about humans, it will think in multiple directions before concluding, unlike previous models, which generate instant responses.
The Downsides of DeepSeek
The biggest downside of DeepSeek is its censorship. If you ask politically sensitive questions about China, such as:
- "What happened in Tiananmen Square in 1989?"
- "What are the biggest criticisms of Xi Jinping?"
- "Is Taiwan an independent country?"
DeepSeek will simply respond, "Sorry, I am not sure how to approach this type of question yet. Let's talk about math, coding, and logic problems instead."
However, it freely criticizes other world leaders like Joe Biden, Donald Trump, and Vladimir Putin. Chinese AI models undergo rigorous government testing before being approved, ensuring they provide "safe" responses to political questions.
DeepSeek’s Open-Source Advantage
Unlike OpenAI, which initially promised open-source AI but later restricted access, DeepSeek is open-source. Its code is publicly available, allowing developers to download, modify, and run the AI locally. Companies like Perplexity AI and Microsoft have already integrated DeepSeek R1 into their platforms after removing censorship.
AI Wars and NVIDIA’s Fall
In 2022, the US government imposed export controls to restrict access to high-end NVIDIA chips, preventing Chinese companies from training AI models effectively. However, DeepSeek adapted by developing more efficient software that required fewer resources, proving that AI advancement doesn’t necessarily require massive financial investment.
This innovation is why NVIDIA suffered the most in the stock market. Until recently, companies like Meta, OpenAI, and Google claimed that AI required more chips, energy, and money to scale. DeepSeek proved otherwise, achieving groundbreaking AI efficiency with minimal resources.
A Lesson for India
DeepSeek’s success serves as a motivation for India. If China can develop world-class AI with limited resources, India can too. The key is focusing on AI innovation and training a new generation of AI experts.
The Future of AI
Those who ignore AI advancements today risk falling behind. AI will impact every sector globally, making it essential for individuals and businesses to upskill. The rapid evolution of AI is undeniable, and adapting to it is no longer optional—it’s necessary for staying competitive in the future.