Hangzhou-based startup DeepSeek, backed by Chinese authorities, is advancing an early release of its R2 AI model following the disruptive impact of its R1 version. The new model aims to enhance coding and multilingual reasoning, sparking a global market frenzy and raising concerns over potential regulatory crackdowns in the tech arena.
To further establish its superiority, DeepSeek is making moves. With a cheap AI reasoning model that beat several Western rivals, the Chinese firm last month set off a $1 trillion-plus sell-off in worldwide equity markets.
Three sources close to the matter have learned that the Hangzhou-based firm is now moving swiftly to introduce the R2 model, which will replace the R1 that was released in January.
Without going into detail, two of them revealed that Deepseek is now aiming to deploy R2 as soon as feasible rather than in early May as originally anticipated, Reuters reports.
The business has expressed its desire for the new model to improve coding quality and extend its reasoning capabilities outside English. The specifics of the shortened release window for R2 have not been made public before.
An attempt to reach DeepSeek for comment on this article was unsuccessful.
Built with less powerful Nvidia chips, R1 is competitive with those produced by U.S. computer titans at a cost of hundreds of billions of dollars. Rivals are still comprehending the significance of this development.
The Indian IT services business Zensar’s COO, Vijayasimha Alilughatta, speculated that the release of DeepSeek’s R2 model may be a watershed point in artificial intelligence. He speculated that if DeepSeek were to succeed in developing affordable AI models, it would encourage other companies around the world to step up their own efforts, eventually releasing the industry’s few monopolies.
The United States government has made establishing AI leadership a top priority therefore R2 is sure to cause them concern. Dozens of Chinese enterprises have already claimed to have begun using DeepSeek models in their goods, and its release might further incite both authorities and businesses in the country.
The creator of DeepSeek, Liang Wenfeng, became a fortune with his quantitative hedge fund, High-Flyer. Very little is known about DeepSeek. A former employer of Liang’s characterized him as “low-key and introverted,” and since July 2024, Liang has avoided interviews with the media.
Twelve former workers and quant fund experts familiar with DeepSeek and its parent business High-Flyer were interviewed by Reuters. Research papers from 2019 onwards, as well as company social media posts and stories from state media, were also part of the review.
They painted a picture of a business that was less like a traditional corporation and more like a research institute, free from the strict hierarchies of China’s tech industry, which allowed it to become the driving force behind what many investors consider to be the most recent innovation in artificial intelligence.
Prior to R1’s worldwide impact, indications suggested that DeepSeek had won over Beijing. Official Chinese media stated in January that Liang, in his role as the AI industry’s appointed representative, met with Premier Li Qiang in Beijing. This put him ahead of the heads of more well-known companies.
The subsequent ruckus about how cheap its models are has given Beijing confidence that it can develop better than the United States; as a result, Chinese businesses and government agencies are using DeepSeek models at a rate no other company has seen.
Tech behemoths Tencent (owner of China’s most popular social media app WeChat), Baidu, and Lenovo have all integrated DeepSeek’s models into their products, while ten state-owned energy corporations and thirteen city governments in China have announced that they have implemented DeepSeek into their systems.
Alfred Wu, a specialist on the Chinese government at Singapore’s Lee Kuan Yew School of Public Policy, stated that both Chinese leaders Xi Jinping and Li “have signalled they endorse DeepSeek.”
“Now everyone just endorses it.”
In response to privacy concerns, governments in South Korea and Italy have removed DeepSeek from their respective app stores, prompting it to receive Chinese support.
“If DeepSeek becomes the go-to AI model across Chinese state entities, Western regulators might see this as another reason to escalate restrictions on AI chips or software collaborations,” stated Stephen Wu, the originator of the hedge fund Carthage Capital and an AI expert.
Another obstacle that Liang has recognized is the need to place restrictions on sophisticated AI chips.
“Our problem has never been funding,” he stated to Waves in July. “It’s the embargo on high-end chips.”
Luis Gochoco is a seasoned managing editor and writer with over a decade of experience covering politics, technology, gaming, and entertainment news. With a keen eye for breaking stories and in-depth analysis, he has established himself as a trusted voice in digital journalism. Luis is one of the key forces behind the success of GameNGuide, contributing to 12 million views through engaging and high-traffic content. He also played a pivotal role in generating 8 million views on International Business Times, shaping the platform’s technology and gaming coverage.
Discover more from Today's Esquire
Subscribe to get the latest posts sent to your email.