Since yesterday, there has been a huge buzz everywhere—everyone is talking about DeepSeek. Today, we take a closer look at an exciting new project that is changing technology and data analysis in China—DeepSeek. This powerful tool is making a big impact on industries like finance and healthcare, using artificial intelligence and big data to create new possibilities. In this blog, we will explain What is DeepSeek and Why is it Impacting the AI sector?
DeepSeek is a chatbot application and software introduced by China that operates on AI principles, just like OpenAI created ChatGPT, or Google developed Gemini, and Meta launched its AI models. Right now, all major tech giants are working on their own AI agents, AI chatbots, AI software, and AI platforms. China has introduced an AI platform called DeepSeek, which has caused a massive stir worldwide. Yesterday, Wall Street dropped by $1 trillion.
This is surprising because Wall Street was extremely bullish, and tech giants like Google, Microsoft, Meta, and media companies were dominating. However, due to DeepSeek, the market for these tech giants crashed, leading to significant losses.
Liang Wenfeng, a Chinese developer, created this AI platform. It was launched as a project under Hiflyer, initially as a site called Hussle. The main goal was to provide affordable and high-quality AI to people.
There are two key factors usually considered essential in AI development:
The secret behind DeepSeek’s popularity is that it has broken both of these rules. Unlike previous AI platforms that required billions of dollars for development and training, DeepSeek was built with only $5.5 million—without relying on high-end, expensive technology. Before this, whether it was Sam Altman’s ChatGPT, Google’s Gemini, or Meta, which is spending $60 million or more, all required massive investments.
Just a few days ago, former U.S. President Donald Trump announced the StarGate project, where OpenAI founder Sam Altman was also present. Trump stated that the U.S. would build data centers across the country and take AI to new heights. Additionally, OpenAI introduced an agent called “Operator”. So, what is this agent and how does it work? Well, even if you’re not directly working on a computer, you can instruct the agent to go online and search for information. While you’re sitting there, with your mouse moving, the agent will search for what you need on the internet.
For example, you could say, “I need to visit Islamabad this weekend, and you should check for hotel rooms available in this price range, preferably near the Secretariat or Blue Area.” The agent would go ahead and search for the information, then bring it back to you. If you give it permission, saying “Book a room within this range,” it would also handle the booking for you. This is the kind of task that OpenAI’s ‘Operator’ agent can perform.
The main reason being cited is that DeepSeek became the top-rated application on the Apple Store in three countries: the USA, UK, and China. Even though Apple does not have its own AI platform, the Apple Store influences market trends. As a result, Wall Street experienced a major market crash when investors realized that China had launched an AI platform that is affordable, offers the same capabilities as Claude 3.5, Gemini, and ChatGPT 4.0, but was built at a much lower cost. Most importantly, DeepSeek is open-source, making it a game-changer in the AI industry.
Let me first share how much the stock prices of major tech giants have dropped. For example, Nvidia‘s stock fell by 12.5%, resulting in a $465 billion loss in just one day. In addition, Broadcom experienced a 14% decline. Stocks for companies like AMD and Microsoft also saw declines. AMD‘s stock dropped by 4.7%, while Microsoft‘s stock fell by 4.4%.
This happened after the U.S. imposed restrictions on China, preventing the export of advanced Graphics Processing Units (GPUs), which are crucial for high-end computations, complex designs, and AI applications. Nvidia, known for producing these GPUs, saw significant impacts due to this. GPUs are essential for AI as they enable fast processing of complex algorithms, making them a key component in the development of AI platforms like DeepSeek.
Nvidia’s state-of-the-art and expensive chips were previously used for AI development, but the U.S. imposed restrictions, preventing these chips from reaching China. As a result, China had no choice but to use less advanced GPUs, which are more affordable. This shift created a significant challenge for the creators of DeepSeek.
The challenge was that these chips, which are not very advanced, cannot run highly complex algorithms. The question was how to achieve the same results as ChatGPT with them. This is where the skills of their software programmers came into play. Their software engineers actually altered the architecture of the LLM (Large Language Model) and used smaller expert models. In other words, they employed many smaller expert models and a central brain. Whenever a task is given, the central brain checks these smaller expert models to determine which one is most suitable for the task.
This technique allowed them to determine which of these smaller expert models was best suited to answer your question. Due to this architectural change, they were able to perform tasks similar to what ChatGPT does, but with cheaper and less advanced GPUs. In reality, the U.S. wanted to maintain its monopoly on AI and keep it intact. This is why it wasn’t supplying China with the more advanced Nvidia chips. Let me show you some interesting statistics.
Between 2003 and 2007, the U.S. led 60 out of the world’s top 64 technologies, while China was only ahead in three. However, by 2025, the U.S. leads in only seven of the top 64 technologies, while China is ahead in 57. This has raised concerns in the U.S., fearing that China might surpass it in the technology race, particularly in AI. The U.S. wants to maintain its dominance in AI because it recognizes that the future of the world will be shaped by AI. Additionally, AI has one of its most remarkable qualities.
This is an open-source model. An open-source model means that its code is available to you on the web. If you wish to make any changes to it, you can do so, as it is up to you. You can use it in your own way. Due to its open-source nature, it has gained more popularity, and people believe they can use it for various use cases, which in technology terms are known as ‘use cases.’ Now, let me quickly show you how you can install it.
This model, called DeepSeek R1, was launched on the 20th of this month and has created a global sensation. If you want to install it, simply go to the Play Store or Apple Store and download the DeepSeek application. However, if you wish to use it as an open-source model, you need to first download Llama, which is a website and application. After installing Llama, you can select DeepSeek within the app and start using it.
You can then download the specific version of DeepSeek you want. After that, install an application called ChatBox, which provides a user-friendly interface. This is a quick guide for using its open-source model: first, go to Llama and download it. Then, select and install your preferred version of DeepSeek. Finally, use ChatBox as the UI interface to access and interact with the open-source model efficiently.
DeepSeek describes itself as an advanced AI model first launched on March 15, 2023, as version one. Its initial release aimed to assist businesses and developers, quickly gaining popularity due to its innovative use of natural language processing and machine learning. Six months later, on September 20, 2023, DeepSeek released its second version, further enhancing its capabilities.
DeepSeek’s second version, launched on September 20, 2023, introduced multilingual support for over 50 languages and real-time learning capabilities. This update positioned DeepSeek as a serious competitor to platforms like ChatGPT and Gemini. On January 10, 2024, DeepSeek released another version with further improvements, including modular architecture. Modular architecture allows different modules to perform specific tasks, making the system more efficient and adaptable.
The reason behind this is that when writing a long software program, if you need to change a particular element, you can simply extract that specific module, modify it, and reintegrate it without disrupting the entire system. Instead of altering the entire structure, modular technology allows for targeted changes, making it highly efficient. This approach is considered superior because it enables improvements, technological advancements, and seamless scalability.
This means it can also be expanded while keeping hardware costs low. DeepSeek has been downloaded over 10 million times—according to DeepSeek itself, though I haven’t verified this independently. DeepSeek claims its standout feature is its exceptional capabilities, whether in natural language processing (NLP), chatbot creation, data analysis, or automating customer support. DeepSeek is highly user-friendly and provides insightful, responsive AI-driven solutions on a platform-based AI design.
This is what makes DeepSeek a game-changer compared to its predecessors, which offer solutions of similar size. DeepSeek empowers users to choose features according to their needs, which not only reduces costs but also ensures that the platform is specifically tailored for its intended use cases. Another notable feature is its extensive language support. DeepSeek can understand and process content in over 50 languages. In fact, everything I’m sharing with you now is also something that DeepSeek has conveyed.
DeepSeek can quickly adapt to new data and customer inputs, ensuring it remains up-to-date and relevant. Thanks to its superior algorithms and open-source nature, DeepSeek’s operational costs are significantly lower. On average, businesses spend between $00 and $00 annually on DeepSeek, depending on their usage, while OpenAI typically costs more.
While ChatGPT costs between $100,000 to $500,000 annually, primarily due to reliance on large datasets and expensive hardware, Google’s cost can range from $200,000 to $1 million, as it is designed for enterprise-level applications. Meta also has an annual cost ranging from $75,000 to $300,000. This cost advantage makes DeepSeek a preferred choice for startups and small businesses, enabling them to use advanced technology without the burden of large AI budgets.
Another key factor in DeepSeek’s success is its open-source nature. Unlike proprietary platforms, DeepSeek’s code is freely available to the public. This encourages innovation, as developers worldwide can contribute to the platform, add new features, and enhance existing versions. When I asked DeepSeek whether it uses Nvidia GPUs, it responded that it utilizes Nvidia GPUs for complex computations, providing faster and more efficient performance.
It works efficiently due to its superior algorithms, ensuring it can run on low-cost hardware, making it accessible to a wider audience. DeepSeek has shaken the AI industry, causing a stir. OpenAI’s valuation dropped by 15%, as investors shifted toward a more affordable and open-source model.
They are concerned about competition with the cloud, as everything is essentially running on speculation. The chips from Nvidia are very expensive and in high demand, which led to Nvidia crashing. OpenAI is also facing a major challenge, as it now has a competitor available at a much lower price. To sum up, DeepSeek is being called a game changer because it is simple, cost-effective, flexible, and connects to AI platforms in innovative ways.
It is available in a way that no other platform offers. Its open-source nature makes it highly friendly and fosters its growth, while its low cost makes it an ideal choice for businesses of all sizes. Whether you are a startup looking to expand or a large company seeking a cost-effective solution, DeepSeek has a lot to offer. With its rapid development and global adoption, it is clear that DeepSeek is here to stay. This was the story of DeepSeek—how it has stirred the industry and how…