IDC: Overseas AI budgets are rapidly shifting from hardware to software, and the three emerging markets have become a new blue ocean for Chinese manufacturers
The Zhitong Finance App learned that IDC Consulting published an article stating that when Chinese companies go overseas, they need to switch from “selling hardware” to “ability to sell” — relying on mature domestic scenario experience to transform solutions into software and platform services, focusing on the three major industries of card finance, retail, and software information services, and pioneering implementation in the three high-certainty scenarios of computing power operation and maintenance, intelligent customer service, and risk control.
Global AI budget flows are diverging: the three major emerging markets have entered the budget release window
Global investment in AI is experiencing explosive growth. According to IDC data, total global AI IT investment is expected to increase from US$695.4 billion in 2025 to US$3.12 trillion in 2030. The share of generative AI investment will increase year by year, and will be close to 60% of total AI investment by 2030. In terms of regional distribution, the US ranks first in the world with a share of over 60%, Western Europe second, and China third with a 9.0% share.
However, there is a significant “time difference” in the pace of growth between regions. Latin America is the region with the fastest growth in AI investment in the world, with generative AI growing at a rate of 60.8%; the Middle East and Africa are growing steadily, driven by a national-level digital vision; demand for Asia-Pacific (excluding China and Japan) enterprises to apply generative AI to improve operations is being concentrated unleashed. The current generative AI growth rate in all three regions is over 50%, and they are in the window of technology implementation and budget release.
In contrast, although the mature markets in Europe and the US are definitely larger, the competitive pattern is relatively solidified, and Chinese manufacturers have high entry costs and little room for differentiation. However, the three emerging markets mentioned above happen to be at the stage where “computing power is in place, but I don't know how to use it” — this is the gap that Chinese manufacturers are best at filling.
IDC suggests selecting the first destination for overseas markets in Latin America, Middle East Africa, and Asia Pacific, and using budget incremental windows corresponding to generative AI growth rates above 50% in these regions to prioritize seizing the cognitive share and customer resources of emerging markets and avoid a head-on war of attrition in the European and American markets.
Overseas AI budgets are shifting from hardware to software: AI platforms and agent tool chains are the main forces in the next wave of growth
Deconstructing the three major technology segments of AI hardware, software, and services shows that there is a clear difference in product form preferences between overseas markets and domestic markets. Hardware accounts for 74.3% of China's AI spending in 2026, showing strong infrastructure-driven characteristics. But overseas, the situation is quite different — Latin America accounts for only 18.0% of AI hardware, while software accounts for 60.9%; the Middle East and Africa together account for 67.9% of software and services. This means that moving the domestic model of “selling servers and selling computing power cards” directly overseas will cause serious dissatisfaction.
What is more noteworthy is the internal growth structure of the software. Although AI embedded applications are still the largest category in stock, AI platforms are more explosive (global CAGR 59.5%) and will soon rise to the top of the software segment in emerging markets such as Latin America and the Middle East and Africa. Among them, the basic model is the core foundation of an enterprise's ability to build AI, and the growth rate of intelligent construction, deployment, and orchestration tools leads all sub-markets (global CAGR 74.5%), indicating that while selecting large models, overseas companies are simultaneously exploring “how to use, manage, and connect models”.
IDC proposes to standardize domestic proven software platforms, agent orchestration tools and industry kits to replace single hardware sales and export an “out-of-the-box” overall solution overseas. The focus is on deploying middleware capabilities such as agent development platforms and multi-model routing and orchestration. Currently, the supply of such products is insufficient in overseas markets, and it is the best window for Chinese manufacturers to differentiate themselves. In regions that are highly dependent on services, such as the Middle East and Africa, it is possible to establish cooperation with local system integrators to implement projects using their channels and delivery capabilities.
Anchoring the three major industries and two main lines: finance/retail/information services are the largest reservoirs for overseas AI budgets
After clarifying the region and technical form, overseas companies need to target the industries and scenarios with the most concentrated budgets in order to achieve accurate implementation.
In terms of industry distribution, software information service companies are the world's largest source of AI budgets, accounting for 32.2%; the banking industry is at the top of the AI spending rankings in Latin America, Western Europe, Middle East Africa and Asia Pacific; and retail accounts for 11.6% of the world. The three major industries make up the largest reservoirs of overseas AI budgets.
Looking at specific scenarios, AI computing power infrastructure accounts for the largest share of expenses in markets such as China and the US all year round, indicating that the management and scheduling of computing power resources has always been the main source of capital investment; at the same time, intelligent customer service and field support have been widely implemented around the world, and threat detection and fraud analysis are particularly significant among major financial and telecommunications customers. The two directions of computing power base and business efficiency improvement form the two parallel main lines of current AI spending.
IDC proposes to provide AI computing power infrastructure operation, maintenance and management services for compute-intensive customers, and cooperate with local cloud vendors or data center operators to output computing power scheduling and optimization capabilities. After localizing and adapting mature domestic applications such as intelligent customer service, field support, and sales assistance, it was quickly launched into high-demand markets such as Latin America and Southeast Asia. For major customers in the overseas financial and telecommunications industry, the focus is on risk control scenarios such as threat detection and anti-fraud — these types of demand are rigid and customer unit prices are the best breakthroughs for Chinese manufacturers to open up leading customers.
Summary and outlook: three time differences, one way out to sea
IDC data clearly points to a conclusion: the biggest opportunity for Chinese companies to go overseas is not to “sell computing power,” but to “sell ability.”
The hardware-dominant logic of the Chinese market (74% of the expenditure is on hardware). In Latin America, Middle East Africa, and Asia Pacific, it is necessary to switch to a style of play that prioritizes software and services. These three regions are going through the stage China went through three or four years ago — computing power is in place, but I don't know how to use it or how to use it effectively. However, Chinese manufacturers happen to have complete engineering experience from computing power management to scene implementation. This is currently the scarcest and most willing to pay in overseas markets.
Three routes to the sea are worth prioritizing: 1. Regionally, it focuses on the three high-growth markets of Latin America, Middle East Africa and the Asia-Pacific region, and avoids head-on competition in mature markets in Europe and the US; 2. In terms of products, standardized output of domestic proven software platforms, agent orchestration tools and industry kits will replace single hardware sales; 3. In the scenario, key card computing power infrastructure operation and maintenance management, as well as high-budget links such as intelligent customer service and risk control in the financial and retail fields.