In what insiders dub as AI’s “Sputnik Moment,” the hitherto unknown Chinese app sent shockwaves across the industry with its cost-effective, open-source approach.
Towards the end of January, a major shockwave shook the tech industry: a previously obscure Chinese AI startup proved to be a fierce competitor against the likes of OpenAI’s ChatGPT and Anthropic’s Claude.
On January 27, stock markets across America and Europe reeled from a whiplash caused by DeepSeek, a hedge fund-turned-AI startup that has blown analysts away with its latest large language model, R1, released on January 20. In what can be best described as a “Sputnik Moment,” its approach to developing AI models that are cost-effective, energy-efficient, and accessible resulted in the share price of Nvidia, America’s leading AI chip maker, to tumble by 17%. This erased $600 billion of market value, per The Economist, the biggest one-day loss in the history of America’s stock market.
But what is DeepSeek? Why did it send the tech industry buzzing and stocks plummeting? The POST dives deep into AI’s current darling.
Related story: Generation Beta begins in 2025. How will these babies born from 2025 to 2039 mark their era?
What is DeepSeek?
DeepSeek is a Chinese AI company that specializes in creating open-source AI models. Founded in 2023 by Liang Wenfeng, who is also the CEO, the company is based in Hangzhou, Zhejiang, and is funded by the Chinese hedge fund High-Flyer.
It is known for attracting young and talented AI researchers from top universities, promising high salaries and an opportunity to work on groundbreaking projects, The New York Times reports. The past few years have seen DeepSeek releasing large language models (LLMs), the kind of technology that underpins chatbots like ChatGPT and Gemini. On Jan. 10, it released its first free chatbot app, which was based on a new model called DeepSeek-V3.
Related story: AI relationships: Yay or nay?
Why did it upend the tech sector?
We can distill the reasons why the app with the cute whale logo sent major shockwaves across the ultra-competitive tech landscape into two main reasons, with the first being its cost-effectiveness.
DeepSeek’s AI model, DeepSeek-R1, performs comparably to models like OpenAI’s GPT-4o but at a fraction of the cost. DeepSeek claims it developed its latest model for only $5.6 million—just 10% of the cost of Meta’s Llama—a sharp contrast with the $100 million spent on GPT-4 in 2023. It also uses significantly less computing power than similar AI models. With this, investors realized that DeepSeek’s LLMs can be as effective as those of US tech giants like Google and OpenAI but at a much lower cost, leading to Nvidia’s stocks going on a free fall on that fateful January day.
And if DeepSeek’s approach can truly lead to LLMs that are less energy-intensive to set up, that’s also good for the environment. The high electricity consumption related to AI is driving up carbon emissions, and the firm’s approach could offer a path towards more sustainable AI scaling.
Related story: Germany has a new tourism ambassador—and she was created by AI
Another reason why DeepSeek has caught the world’s attention is that it’s open source, meaning anyone can copy, download, and build on it. Its code and technical specificities are freely shared, enabling global developers and organizations to access, modify, and implement.
This contrasts with companies like OpenAI, whose models, data sets, and algorithms are kept secret, according to the World Economic Forum (WEF). Although Meta and Google’s models are open for anyone to view, the WEF further explains, they are not “truly open source” because the way users apply the models is restricted by licenses and the training data sets aren’t made public.
DeepSeek being open source also adds to its accessibility, especially with its new chatbot. By end-January, it had overtaken ChatGPT as the most downloaded free app on the Apple App Store in the US. Many believe its increasing popularity could be a significant step in democratizing AI, allowing smaller companies, start-ups, and individual developers to build on DeepSeek-R1.
What are its cons?
These said, not everything about DeepSeek is unicorns and rainbows. It has faced criticism for alleged censorship, both in its answers and training data. A report on The Guardian, for instance, praised the app, saying it “worked well, until we asked it about Tiananmen Square and Taiwan.”
Related story: 14% of Filipino workforce at risk of losing their jobs to AI. Here’s how to protect your career
More recently, it has been prohibited by some governments, citing privacy concerns. Per Al Jazeera, DeepSeek is now banned on government devices in South Korea, Australia, and Taiwan. All three countries cite “security concerns” about the Chinese technology and a lack of clarity about how users’ personal information is handled by the operator. More countries, including the United States, might follow suit.
Some people also argue open-source AI poses considerable risks, as it could be used for malicious purposes such as developing bioweapons and facilitating the spread of mis- and disinformation.
Related story: Sick of Google’s AI summaries? Try swearing at it instead
In sum
Either way, DeepSeek has compelled the AI industry to reevaluate competitiveness. Its ability to deliver comparable AI performance at a much lower cost and with an open-source approach has disrupted the formerly ironclad industry consensus that advanced AI systems require massive investment.
But as in all technological breakthroughs, DeepSeek’s quick rise must be met with caution. While its open-source framework promotes inclusivity and transparency, it also raises important questions about data privacy, geopolitical dynamics, and security. As artificial intelligence continues to evolve, navigating these challenges is crucial to unlocking its full potential while ensuring its responsible adoption.
Related story: Meta is done with fact-checking: What this means for Filipino users
