US remains global leader in AI, but China rapidly closing gap with cheaper models: JP Morgan
New Delhi, June 24
The United States continues to dominate the global artificial intelligence ecosystem, but China is emerging as a formidable challenger through cost-efficient AI models that are increasingly attracting enterprise adoption, according to a report by financial services major J.P. Morgan.
In its report, "Semiquincententacles: the US grip on global markets at 250", J.P. Morgan said the US retains a commanding lead in AI readiness, innovation, infrastructure and investment, although Chinese firms are making rapid gains in the commercial deployment of AI models.
"The US is the most vibrant and prepared country for AI, with China close behind on some measures," the report said, citing multiple global AI readiness indicators.
According to the report, the United States remains at the forefront of AI development owing to its leadership in frontier AI laboratories, semiconductor technology, hyperscale cloud infrastructure and productivity gains linked to artificial intelligence.
J.P. Morgan noted that US companies continue to dominate the AI hardware ecosystem. Nvidia alone accounts for the bulk of global AI accelerator revenues, while technology giants such as Google, Amazon, Microsoft and Meta are increasingly deploying custom-designed chips to reduce costs and expand computing capacity.
The report also highlighted strong AI-driven productivity gains in the United States, stating that "Whether the issue is labor productivity or total factor productivity, the US leads the G10."
However, the report cautioned that China's AI ecosystem is advancing rapidly, particularly in the area of open-weight and low-cost AI models.
According to J.P. Morgan, the most cost-efficient segment of the global AI model landscape is increasingly being led by Chinese companies.
Referring to comparative assessments of AI performance and operating costs, the report said the "efficient frontier" is "dominated by China (DeepSeek, MiniMax, Xiaomi, Alibaba)" with only a limited presence from US models.The report further noted growing market acceptance of Chinese AI offerings as businesses seek to reduce operating costs amid rising token pricing by leading US frontier AI companies."OpenRouter shows a surge in API calls to Chinese models," the report said, adding that by April 2026, leading Chinese open-weight models "scored within a few dozen Elo points of closed frontier models and cost 10x - 50x less per token."J.P. Morgan observed that some enterprises have already begun shifting workloads from expensive frontier AI models to lower-cost alternatives, a trend that could intensify if pricing pressures continue.At the same time, the report stressed that frontier AI models developed by US firms remain essential for advanced applications such as cybersecurity, scientific discovery, agentic systems and large-scale reasoning tasks, ensuring that American companies retain a technological edge despite growing competition.The report noted that while the United States remains the global leader in artificial intelligence, China's rapid progress in developing competitive and significantly cheaper AI models has emerged as one of the most important challenges to US dominance in the sector.
— ANI
Reader Comments
The cost difference is staggering - 10x to 50x cheaper for Chinese models! With Indian startups always looking to cut costs, this could be a game-changer. But we must balance it with security concerns. Can't have our banking sector depending on foreign models without proper checks.
Meanwhile, in India we're still arguing about whether AI will take away jobs. Wake up and smell the coffee! China and US are fighting for global dominance, and we're stuck in debate mode. Need aggressive policy push for AI adoption in agriculture, healthcare, and education sectors. Time is running out!
Good analysis from JP Morgan. But let's not forget - cheap models don't automatically mean good models. For mission-critical applications like cybersecurity and scientific research, US frontier models still lead. China's strength is in commodity AI, which is fine for chat bots but not for serious applications. We need both types in our ecosystem.
The real story here is pricing. US companies charging 50x more? That's going to force Indian enterprises to look east. Already seeing IT service firms exploring Chinese models for internal tools. But government should create incentives for Indian-made AI models too. No point being dependent on either superpower for such strategic technology.
China's rapid progress is impressive but concerning. As an Indian, I worry about data sovereignty. If Indian companies start using Chinese AI models heavily, it could create new vulnerabilities. We need a balanced approach - leverage global innovations while building our own capabilities. The semiconductor mission is a step in the right direction, but more needed on AI models front.