0.3 percentage points. That’s how much one investment bank estimates China’s AI economy will add to national GDP in 2026, a modest number that sits oddly next to how much the AI boom is reshaping specific Chinese cities.
Beijing, Shanghai, Hangzhou, and Shenzhen are delivering their strongest contribution to national growth in at least two decades. Most of the rest of the country is sliding toward the bottom edge of Beijing’s full-year growth target. It’s a split economists have started calling “K-shaped divergence,” a framing YourNewsClub notes undersells how narrow the winning side of that K actually is: four cities, not a region, not a sector spread nationally.
In Hefei, the memory-chip capital that lured manufacturer CXMT with heavy state backing, factories reportedly can’t keep up with demand. In Changchun, a rust-belt auto hub, officials admitted to “unprecedented difficulties,” a phrase quietly scrubbed from official channels after it drew attention online, a contrast YourNewsClub spots as sharper than the “K-shaped” framing alone suggests, since one side of that K is actively degrading in real time, not just growing more slowly.
Freddy Camacho, who studies the political economy of computation, materials, and energy as dominance assets, thinks the real test is whether Hefei is a template or an accident. “Patient state capital chasing one industry for years, that’s the model everyone’s now trying to copy,” he said. Whether it’s actually repeatable, or whether it depended on advantages that only existed once, is what YourNewsClub clocks as the more important question than this year’s growth numbers themselves.
The optimism inside the boomtowns is starting to show up in public markets, too. Beijing-based GigaAI, a world-model startup headquartered in one of the cities driving this divergence, is nearing a Hong Kong listing at a $3 billion valuation, aiming to become the first world-model company globally to go public. Moonshot AI, maker of the Kimi K3 model that briefly outperformed some Western frontier systems on independent benchmarks, is reportedly targeting its own IPO within six months.
That IPO wave carries its own warning sign. Chinese AI stocks have rallied roughly 80% this year on the back of exactly this boomtown story, and some analysts are now questioning whether elevated valuations can be sustained without clearer evidence that AI earnings growth translates into the kind of broad-based economic activity a K-shaped chart, by definition, says it hasn’t produced yet.
Even inside the boomtowns, the money isn’t reaching households evenly. Retail sales are reportedly shrinking in several of the fastest-growing hubs, undercut by the same automation driving the manufacturing surge.
Jessica Larn, who studies macro-level technology policy and infrastructure impact of AI, sees a policy question underneath the economics. “The real issue is whether Beijing has the fiscal tools to redistribute AI gains before the gap widens further,” she said, a redistribution problem separate entirely from whether the underlying technology itself keeps advancing. Beijing can win the AI race and still lose the country’s regional cohesion in the process.
One caveat matters here: a single half-year isn’t a trend. Whether the boomtown map widens or narrows further is, as YourNewsClub benchmarks it, the actual thing worth watching in the second half, not this quarter’s number.