China’s Deepseek Could Be Bigger Than ChatGPT

In just over a year, Chinese AI startup DeepSeek has sent shockwaves through the tech world with its groundbreaking advancements.

Offering performance comparable to top-tier AI chatbots like OpenAI’s ChatGPT but at a fraction of the cost, DeepSeek has emerged as a disruptive force, challenging the notion that AI development requires endless power and resources.

What Is DeepSeek?

Founded in 2023 by Liang Wenfeng, a former chief of the AI-driven hedge fund High-Flyer, DeepSeek operates with an open-source approach, allowing developers to inspect and enhance its software. Its app, released in early January, quickly became a global sensation, topping iPhone download charts in the US, UK, China, and beyond.

 

DISCOVER: AI Meets the Magazines: OpenAI Strikes Game-Changing Deal with Condé Nast

MORE: Why do AI assistants almost always have female voices?

What sets DeepSeek apart is its ability to articulate reasoning before delivering responses, a feature that differentiates it from competitors like ChatGPT.

The company claims its latest model, DeepSeek R1, performs on par with OpenAI’s offerings and excels in benchmarks like AIME 2024 for math tasks and MMLU for general knowledge.

 

Efficiency Redefines the AI Race

DeepSeek’s innovation challenges the industry norm of massive spending on AI infrastructure. Its models reportedly require far less computational power than those developed by OpenAI or Meta, raising questions about the necessity of the billions of dollars invested annually by US tech giants.

One user took to Instagram, commenting “It’s a great thing for humanity. We’re spending so much electricity on training and data centers. If we can reduce that by 95% and still reap the benefits, then this is truly a gift from China as they’ve made it open source.”

 

DISCOVER: Should you be hiding your children’s faces online?

MORE: I Tried the World’s First AI-yurvedic Skin Scan: Here’s What Happened

 

Liang Wenfeng, DeepSeek’s founder, emphasises efficiency over sheer investment. “More investment does not necessarily lead to more innovation. Otherwise, large companies would take over all innovation,” he said in an interview with Chinese outlet 36kr.

The success of DeepSeek has amplified concerns in the US about the effectiveness of restrictions on advanced semiconductor exports to China. These curbs were designed to slow China’s progress in AI, but DeepSeek’s breakthroughs suggest that engineers are finding ways to innovate with limited resources.

A Global Ripple Effect

DeepSeek’s rapid rise has already impacted the global AI market. Shares in companies like Nvidia, which have thrived on AI’s infrastructure boom, have taken a hit as investors reconsider the long-term need for high-cost hardware. Meanwhile, Chinese firms connected to DeepSeek, such as Iflytek, have seen their stock surge.

Marc Andreessen, a prominent tech investor, dubbed DeepSeek’s success “AI’s Sputnik moment,” highlighting its potential to reshape the global AI landscape.

 

DISCOVER: Mother Sues After Teen’s Tragic Love for Character.Ai Leads to Suicide

MORE: Robots in the Classroom: The UK’s Innovative Solution for School Absences

Challenges and Controversies

Despite its success, DeepSeek faces obstacles. Its AI models self-censor on sensitive topics like Tiananmen Square and Chinese political figures, limiting their versatility. Additionally, the company’s infrastructure has struggled to handle the surge in popularity, with a significant outage reported on January 27.

 

MORE: AR Smart-glasses revolutionise conversation for the deaf community

DISCOVER: Is This Tech The Future Of Jewellery?: These Bizarre Accessories Give You An AR Beauty Experience

A Catalyst for Change

DeepSeek’s impact extends beyond competition—it raises the stakes for regulatory oversight in AI development. As developers worldwide adopt its efficient models, the call for guardrails around AI use grows louder.

For now, DeepSeek’s efficient and accessible approach is forcing the global AI industry to rethink its strategies, potentially ushering in a more resource-conscious era of innovation.