China could lead the world in AI software by 2030, and still not fix its economy
A new Bloomberg Intelligence forecast sees Chinese firms closing the hardware gap quickly, but struggling to turn AI advances into profit.
Bloomberg Daybreak: US Edition · Uncertainty in Israeli Election; Bolsonaro Flees BrazilBloomberg reporter John Tucker, describing a new Bloomberg Intelligence report, said China is "well-placed to lead the world by 2030 in the large language models."
Put simply, large language models are giant software systems that learn from vast amounts of text and can write, answer questions, and hold a conversation. They are the engines inside today's chatbots and many other AI tools, and the report puts them at the core of AI.
On the physical side of AI, the report draws a distinction between now and soon. For the hardware needed to run AI, "from chips to optical networking components" that move data quickly between computers, Tucker said "Chinese firms are mostly behind now but will be competitive in less than four years."
That forecast covers both halves of the AI race, the software that provides the intelligence and the hardware that provides the computing power.
Then came the limit on what that advance would mean for the wider economy. Tucker, summarizing the report, said "none of this is going to rescue China's economy from a slowdown."
Tucker added that Chinese AI firms, "like peers elsewhere are struggling to figure out how to make money out of their innovations."
Questions people ask
What exactly is a large language model?
A large language model is software trained on huge collections of books, websites, and other text so it can predict words and generate answers. It powers chatbots, search summaries, translation, and drafting tools. Bigger and better trained models tend to be more capable, but also far more expensive to build and run.
What kind of hardware matters in the AI race?
They mean the specialized chips that train and run AI, the high capacity memory those chips need, and the fast networking that links thousands of chips together in data centers. Shortages or export limits in any of those parts can slow AI progress, which is why chip plants and component suppliers have become central to the competition.
Why do analysts think China could take the lead?
China has a large pool of engineers, enormous amounts of domestic online data, strong government support, and fast building of data centers and power capacity. U.S. export controls have limited China's access to the most advanced foreign chips, which has pushed Chinese companies to develop domestic alternatives.