BEIJING: Chinese startup DeepSeek has taken the global tech landscape by storm, challenging the dominance of established Western players like OpenAI and Nvidia.
DeepSeek’s release of a major upgrade to its V3 large language model has heightened the competition, offering more affordable AI solutions that are forcing global companies to rethink their strategies.
The latest version of DeepSeek’s AI model, named DeepSeek-V3-0324, was made available through the AI development platform Hugging Face.
The new model features significant improvements in reasoning and coding abilities over its predecessor, according to benchmark tests published on Hugging Face, which demonstrate its enhanced performance across multiple technical metrics.
This release marks the startup’s ongoing push to establish a significant presence in the global AI market.
DeepSeek first gained attention in January when it unveiled a model that purportedly cost just several million dollars to develop—far less than the multi-billion-dollar investments made by its Western counterparts.
Since then, Chinese tech giants have followed suit, flooding the market with a rapid succession of low-cost AI models that undercut the premium offerings of companies like OpenAI and Google’s parent Alphabet Inc.
Over the past two weeks alone, Chinese companies have rolled out several product updates or new releases.
Notable among these were Baidu Inc.’s unveiling of its Ernie X1 model, which competes directly with DeepSeek’s R1 model, and Alibaba Group’s launch of its AI agents and reasoning model upgrades.
Tencent Holdings and Ant Group have also made headlines with their new AI initiatives. Meituan, the world’s largest meal delivery service, has also announced a substantial investment in AI technologies.
These developments are not just a sign of Chinese companies jumping on the AI bandwagon; they represent a broader effort to set new global standards for AI models.
Almost all of these models are open-sourced, making them accessible for developers worldwide.
According to Amr Awadallah, CEO of Vectara Inc., a company that helps businesses build AI solutions, the rapid rise of Chinese AI startups could lead to significant margin compression across the AI ecosystem.
“We’re about to see a trend toward significant margin compression for companies throughout the ecosystem. Not just AI model builders, but the large AI enablers that are also driving the growth of the industry,” Awadallah said.
In response to DeepSeek’s success, OpenAI has signalled that it may explore giving away some of its technology in a bid to remain competitive. However, it is also considering charging more for its most advanced products.
“The hard work goes on. The EU’s priority is a fair, balanced deal instead of unjustified tariffs,” Sefcovic wrote on X after talks with US officials including US Commerce Secretary Howard Lutnick Tuesday.
The global AI industry is now facing an open-source revolution, with Chinese developers leading the charge.
These models are already being replicated and used across markets in the US, India, and beyond, despite attempts by some governments to restrict access to DeepSeek’s technology on certain devices.
According to Kevin Xu, a tech investor based in the US, the aggressive pricing strategies of Chinese cloud providers hosting AI development are likely to have a significant impact beyond China’s borders.
“It’s just sort of a natural evolution of that price war within the Chinese ecosystem spreading out into other markets,” Xu remarked.
Chinese tech companies, including Alibaba, are also focusing on integrating AI into broader products and services, with global names like Apple and BMW already partnering with the Chinese tech giant to incorporate its AI offerings into their devices.
This growing market share is beginning to challenge the infrastructure investments that Western companies such as OpenAI and Microsoft have made in AI development.
James Wilton, founder of tech consultancy Monevate, remarked on the shift: “If you assume the Chinese LLM players’ intent is to disrupt the market and steal share, it’s worked. They might be willing to eat up the costs right now, but it won’t stay free forever.”
This surge of open-source, high-performance AI models is expected to expand into fields such as computer vision, robotics, and image generation in the coming months.
Experts predict that as Chinese companies continue to optimise and scale their AI technologies, the demand for AI-powered devices will also rise—adding yet another layer of competition for US firms.
In an interview at a conference in China, Apple CEO Tim Cook described DeepSeek’s AI models as “excellent” and praised the efficiency of the model.
However, the rapid pace of Chinese AI development is raising questions about whether the US will need to reassess its massive investments in AI infrastructure.
According to Bloomberg Intelligence, tech giants like Amazon, Meta, and Microsoft are projected to increase their spending on AI systems by 44% in 2025 compared to the previous year.
