A breakthrough AI model from a Chinese startup is causing ripples in the tech world, sparking both excitement and concern, Cape {town} Etc reports.
DeepSeek, founded just a year ago by Liang Wenfeng, has unveiled its R1 AI, which rivals top models from companies like OpenAI, Meta, and Google but at a fraction of the cost.
The startup, which has captured significant attention, has reportedly spent just $5.6 million to develop R1, a fraction of the billions that American companies invest in their AI systems. This is especially striking given that the U.S. has worked for years to limit China’s access to advanced AI chips, citing national security risks.
DeepSeek’s achievement on less powerful chips has raised eyebrows, especially when considering the immense costs typically associated with AI development.
Liang Wenfeng, a Chinese hedge fund manager, is positioning DeepSeek as a major player in the AI field. His company has released several AI models, with the R1 version gaining rapid attention since its unveiling late last year.
Unlike previous models from the company, R1 has opened up new possibilities, not only because of its cost-efficiency but also because it’s open-source, allowing other developers to build upon it.
DeepSeek’s success story challenges the assumption that America’s AI dominance is secure. Wall Street reacted swiftly, with significant losses in the stock prices of major tech players like Nvidia and Meta as investors weighed the implications of this new, lower-cost competitor.
The low-cost nature of R1 has caused some scepticism. While the model’s price tag is a game-changer, questions remain about the accuracy of DeepSeek’s cost claims and whether it can truly compete at the highest levels. Still, with the app surging in popularity and nearing 2 million downloads, the potential for a shift in the AI landscape is undeniable.
Also read:
Microsoft SA to launch Artificial Intelligence skilling initiative
Picture: Justin Sullivan / Gallo Images
