DeepSeek may launch new AI model earlier than planned — Reports

Chinese AI company DeepSeek is accelerating the release of its next-generation AI model, R2, in a move to solidify its position as a leading player in artificial intelligence.

The Hangzhou-based company had initially planned to release R2 in early May, but according to three individuals familiar with the company, it now aims for an earlier launch. Two of these sources declined to specify a revised date.

DeepSeek’s push comes after its R1 model, introduced in January, triggered a global stock sell-off exceeding $1 trillion by outperforming many Western AI models despite being built with less-powerful Nvidia chips. The company believes R2 will deliver enhanced coding capabilities and expand reasoning across languages beyond English. Details regarding the accelerated release timeline have not been previously reported.

DeepSeek did not respond to a request for comment.

Competitive Impact and Industry Reactions

The AI industry is still assessing the implications of DeepSeek’s R1, which competes with models developed by US tech giants at costs reaching hundreds of billions of dollars.

“The launch of DeepSeek’s R2 model could be a pivotal moment in the AI industry,” said Vijayasimha Alilughatta, chief operating officer of Indian tech services provider Zensar. 

He added that DeepSeek’s success in creating cost-effective AI models “would likely spur companies worldwide to accelerate their own efforts … breaking the stranglehold of the few dominant players in the field.”

Beyond industry competition, the US government is likely to be concerned about R2’s release, as AI leadership remains a national priority. Meanwhile, dozens of Chinese companies have already begun integrating DeepSeek models into their products, further strengthening China’s position in the AI race.

The Low-Key Billionaire Behind DeepSeek

DeepSeek remains a largely enigmatic company, led by its reclusive founder, Liang Wenfeng. A former quantitative hedge fund manager, Liang built his fortune through High-Flyer, one of China’s most successful quant funds. Since DeepSeek’s rise, he has avoided media attention, with his last public comments dating back to July 2024.

According to former colleagues, Liang has always preferred operating behind the scenes. “Liang gave us control and treated us as experts. He constantly asked questions and learned alongside us,” said Benjamin Liu, a 26-year-old researcher who left DeepSeek in September. “DeepSeek allowed me to take ownership of critical parts of the pipeline, which was very exciting.”

Reuters interviewed a dozen former employees and professionals in the quant fund industry, as well as reviewed state media articles, social media posts, and research papers dating back to 2019. The findings depict a company that functions more like a research lab than a traditional profit-driven enterprise, breaking away from the hierarchical structure typical of China’s tech sector.

The Power Behind DeepSeek’s AI Models

DeepSeek’s breakthroughs have been fueled by High-Flyer’s substantial investment in research and computing power over the past decade.

As early as 2020, High-Flyer executives declared they were going “all in” on AI, reinvesting 70% of revenue into AI research. The firm spent 1.2 billion yuan on two AI supercomputing clusters in 2020 and 2021, the second of which—Fire-Flyer II—was powered by approximately 10,000 Nvidia A100 chips.

Regulators initially questioned the fund’s growing accumulation of chips. “Regulators wanted to know why they need so many chips? How they were going to use it? What kind of impact would that have on the market?” said a person with direct knowledge of officials’ inquiries. However, authorities ultimately decided not to intervene—a decision that proved critical when the US imposed an export ban on A100 chips in 2022, by which time DeepSeek had already secured its computing infrastructure.

Some Western AI experts, including Scale AI CEO Alexandr Wang, have speculated that DeepSeek possesses as many as 50,000 high-end Nvidia chips. However, Wang has not provided evidence, and DeepSeek has not addressed these claims. Former DeepSeek employees attribute the company’s success not to chip availability but to cost-effective AI architecture.

ALSO READ FROM NIGERIAN TRIBUNE