DeepSeek R1: This Is What’s Driving The Hype Around China’s New AI System

What is DeepSeek’s new model, R1?

DeepSeek develops large language models (LLMs), the same type of AI systems used by OpenAI’s ChatGPT, Meta’s LLaMA, Google’s Gemini, and so on.

These chatbot-like systems are trained on vast amounts of text data – such as books, news articles, and web pages – from which they learn how to construct sentences. LLMs can be used for all kinds of purposes, such as writing computer code, problem-solving, content generation, and even writing news articles (with mixed success). 

On January 20, 2025, DeepSeek released its latest AI model, called R1. In a nutshell: it’s extremely smart, very efficient, and shockingly low-cost.

In terms of technical ability, DeepSeek claims that R1 matches or beats its competitors, including OpenAI’s o1, by several different metrics. Many investors and tech experts agree. 

Crucially, it’s able to perform these tasks with greater efficiency, requiring less computing power and less powerful chips to achieve the same (if not better) results as rival models.

Another key point: DeepSeek-R1 is open-access, meaning it is freely available for public use. Users have access to its code, allowing them to integrate it into their own applications, or experiment with it for research and development.

“DeepSeek has demonstrated that large language models can learn to solve complex tasks with fewer computing resources and less data for training. As a result, future chatbots may respond faster and be trained with less but higher quality data, consume less energy, and may be less costly to develop,” Wolfgang Meyer, an Associate Professor in STEM at the University of South Australia, said in a statement.

“While we don’t yet fully understand the training process and capabilities of DeepSeek, its results may shape how future chatbots are developed,” explained Meyer.

However, critics have already highlighted significant downsides, including its compliance with Chinese government censorship. For example, DeepSeek refuses to answer questions about the 1989 Tiananmen Square protests, an event that the Chinese Communist Party has consistently sought to downplay.

Some commentators have also noted that DeepSeek is reportedly sending tons of user data from the US to China, raising big security and privacy concerns.

Who is behind DeepSeek?

DeepSeek is an AI company based in Hangzhou, a city that’s home to many Chinese tech companies, including Alibaba, kind of like China’s equivalent to Amazon.

The company was founded in 2023 by Liang Wenfeng, the 40-year-old entrepreneur behind the hedge fund High-Flyer. Wenfeng bought thousands of Nvidia computer chips before the Biden administration began limiting exports of chips to China and used them to build an AI side project, according to the Financial Times.

“When we first met him, he was this very nerdy guy with a terrible hairstyle talking about building a 10,000-chip cluster to train his own models. We didn’t take him seriously,” one of Liang’s business partners told the Financial Times. 

“He couldn’t articulate his vision other than saying: ‘I want to build this, and it will be a game change.’ We thought this was only possible from giants like ByteDance and Alibaba,” the associate added. 

DeepSeek has reportedly said it spent just $5.6 million on computing power for its base model. While this wouldn’t cover the total cost of the project, it’s certainly a lot cheaper than the hundreds of millions or even billions of dollars that American firms have allocated to AI development.

This lean, cost-effective approach is a key factor in why DeepSeek’s new model is causing a stir in the tech industry.

“DeepSeek’s open-source model was developed at a much lower cost and requires less powerful hardware to run than current market leaders such as ChatGPT,” Dr William Darler from the University of Leicester School of Business said in a statement sent to IFLScience. 

“Early consensus from experts is that DeepSeek’s AI models look robust but it currently lacks the volume of training data to compete with the best closed-source models. This is likely to prompt generative AI organisations to rethink their business models which may be good news for organisations and consumers for whom the current costs may be prohibitive,” added Darler.

What’s the big deal about DeepSeek?

When markets closed on Monday, January 27, almost $600 billion was slashed from the value of American AI microchip maker Nvidia, marking the biggest one-day market value drop of any company in history. Clearly, investors in the US and Europe were spooked. 

On the same day, DeepSeek overtook OpenAI as the most downloaded free app in the US on Apple’s App Store.

DeepSeek’s meteoric rise means American AI companies now face significant global competition. Suddenly, they aren’t the only show in town. Some argue that the rise of Chinese AI companies will foster healthy competition with the US, driving advancements and accelerating progress in the field.

Silicon Valley venture capitalist Marc Andreessen posted on X: “DeepSeek-R1 is AI’s Sputnik moment,” referencing the Soviet satellite that sparked the space race of the Cold War and paved the way towards the Apollo Moon landings. 

On the other hand, the growing threat to American AI supremacy has already disrupted the US tech industry and is likely to stoke up geopolitical tensions even further. 

“This is the clearest signal yet of where the future is heading. AI will be the defining technology of our time, and the race to dominate it will shape global markets for decades to come,” Nigel Green, CEO of global financial advisory giant deVere Group, said in a statement seen by IFLScience.

“China’s technological advances, particularly in AI, are eroding the US’s ability to use tariffs as a tool to maintain global supremacy. The balance of power is shifting, and Washington must recognize that it can’t always dictate terms to Beijing as it once did. This new reality will have far-reaching consequences for investors and policymakers,” he added.