While US banks debate building versus buying AI, China’s state-owned giants are all-in on building.
A new report from Morgan Stanley finds that China's largest banks have moved past experimentation and are deploying artificial intelligence on a massive scale, integrating the technology into their core infrastructure. Four major state-owned banks each saw IT investment exceed RMB25 billion ($3.4 billion) last year as they race to enhance efficiency, strengthen risk monitoring, and protect profitability from a low interest rate environment.
"AI has entered a phase of scaled implementation and become part of core banking infrastructure," the Morgan Stanley report wrote. The broker believes the technology enables Chinese banks to expand services to more corporate and retail customers without increasing headcount, while offsetting pressure from declining net interest margins.
The spending, which accounts for 3 to 4 percent of each bank's total revenue, is being used to build enterprise-level AI platforms, internal knowledge bases, and digital workflow tools. While Morgan Stanley did not name all four, it identified Industrial and Commercial Bank of China (ICBC), China Construction Bank (CCB), and China Merchants Bank (CMB) as industry bellwethers combining strong technological capabilities with solid execution.
The move to bring AI development in-house is a strategic response to preserve stable returns on equity (ROE) and more precisely match financing with needs in the real economy. McKinsey & Co. has estimated that banks that are pioneers in AI could open a four-percentage-point gap in return on tangible equity over slower-moving peers, a prize Chinese banks are now aggressively pursuing.
A Different Answer to the Build-vs-Buy Question
The unified direction among China's top lenders provides a stark contrast to the divided strategy seen in the United States. There, JPMorgan Chase is spending $18 billion annually on technology to build most of its AI tools in-house, while Capital One has pursued major acquisitions, buying AI-native Brex for $5.15 billion and Discover Financial for $35.3 billion to acquire technology and talent.
According to a 2026 report from venture capital firm Team8, 81% of North American banks have revised their build-vs-buy thinking because of AI. The dividing line is data gravity. Banks are choosing to build applications that touch their proprietary data—analytics, workflow orchestration, and personalization—while buying services like fraud detection where a vendor's scale across the entire system provides a structural advantage.
China's state-owned banks, which sit on some of the world's largest proprietary datasets, appear to have taken this logic to its conclusion. By building their own platforms, they prevent handing over their most valuable asset—client data—to third-party vendors, which AI makes exponentially more valuable.
From Tech Spend to Core Infrastructure
The scale of the investment signals that, like JPMorgan in the US, Chinese banks have reclassified AI from a discretionary innovation item to core infrastructure, on par with payment systems and cybersecurity. The goal is to create a lasting competitive moat by embedding AI into every facet of the bank, from front-office client interaction to middle-office risk and back-office processing.
This allows them to automate manual workloads and, more importantly, create a more intelligent and responsive risk management framework. For an economy grappling with a property downturn and shifting trade patterns, the ability for banks to more accurately price risk and allocate capital is a significant national advantage. Morgan Stanley notes this ultimately improves the overall competitiveness of Chinese enterprises.
While the spending figures are large, the report suggests the key differentiator will be execution. The banks that succeed will be those that can translate the massive investment into broad internal applications and a fundamental transformation of workflows. For now, Morgan Stanley is betting that ICBC, CCB, and CM Bank are the names to watch.
This article is for informational purposes only and does not constitute investment advice.