JPMorgan: China's AI "Battle of 100 Models" Shifts to Enterprise Value Competition
Alina Collins
JPMorgan's head of China equity research Alex Yao says the focus of China's AI race is shifting from raw model performance to measurable business value, with the industry consolidating fast around a few globally competitive players.
Is the "battle of 100 models" ending?
Alex Yao told the South China Morning Post that China's AI sector is moving rapidly from fragmentation toward consolidation.
This means → most small and mid-sized model makers will be squeezed out, leaving a handful of leaders.
The winners will be firms that convert consumer-facing features into reliable enterprise infrastructure. In plain terms = a chatbot alone is not enough — the model has to help businesses get work done and generate revenue.
Does the gap with the US matter?
Yao acknowledged that Chinese AI models still trail top US models on some benchmarks.
But he argued this is not decisive, for two reasons: ① Chinese users cannot access US models directly; ② real-world application value now matters more than benchmark scores.
"You don't need a model with Einstein-level intelligence," he said. "Once a model reaches the level of a strong master's graduate, it can start doing real work."
Will Chinese consumers actually pay for software?
Yao believes concerns about Chinese users' reluctance to pay for software may be overstated — adoption hinges on whether a product delivers "clear and verifiable value."
In plain terms = users are not opposed to paying; many past software products simply failed to feel worth the price.
This reflects a deeper point: the real barrier to AI monetization is not technology — it is perceived value.
Who has already started charging?
ByteDance launched a subscription service for Doubao in early May, priced at 68 to 500 yuan per month.
This means → leading players are already testing real willingness to pay; the path to monetization is no longer hypothetical.
This signals that China's AI competition is entering its next phase: from "whose model scores higher" to "who can generate revenue first."
Content is for reference only, not financial advice.