AI operations have fundamentally reshaped the revenue and profit structure of Foxconn Industrial Internet Co., Ltd.. In 2025, the cloud computing segment contributed 66.75% of the company's total revenue, with its gross profit accounting for 54.83% of the company's overall gross profit, and the growth rate of profit contribution from cloud computing has started to outpace revenue expansion. Meanwhile, during this rapid AI growth phase, the procurement share from the top five suppliers climbed to 58.36%, direct material costs for 3C electronics reached 92.50% of total costs, and both inventory and short-term borrowings have risen significantly. The company's business scope is extending toward system-level segments such as AI racks, liquid cooling, power management, and high-speed networking, shifting its fundamentals from a focus solely on AI server revenue growth to a focus on profit contribution and working capital efficiency.

Supplier concentration has more than doubled, signaling a shift in procurement structure

In 2023, purchases from the top five suppliers accounted for 27.23% of total annual procurement; this figure jumped to 56.27% in 2024, and further climbed to 58.36% in 2025. Over the same period, sales to the top five customers dipped slightly from 65.33% to 62.01%. In other words, during the rapid AI business expansion, customer concentration downstream remained largely stable, but upstream procurement has become increasingly concentrated among a handful of suppliers. This data does not directly prove that any single chipmaker has become a primary supplier, nor does the annual report disclose supplier names, so unverified attributions are not appropriate. However, timing-wise, the rise in supplier concentration coincides with the scaling of AI server operations. Since the annual report does not break down procurement by business segment, this metric is best viewed as supporting evidence of changes in the supply chain structure. From 2023 to 2025, cloud computing revenue grew from RMB 194.308 billion to RMB 602.679 billion, with its share of total revenue rising from 40.79% to 66.75%; AI servers as a proportion of cloud revenue increased from roughly 30.00% in 2023, surpassing 40.00% in 2024, and in 2025, AI server revenue from cloud service providers grew more than 3.00 times year-over-year. At the same time, direct materials' share of 3C electronics costs has continued to rise, from 89.09% in 2023 to 91.38% in 2024, reaching 92.50% in 2025. It should be noted that the cost figures above cover the company's overall 3C electronics products, not the cost structure of AI servers alone, and therefore can only serve as auxiliary evidence of changes in product mix. This indicates that the rapid growth in AI revenue carries a distinct hardware manufacturing characteristic. High-value components can drive up overall system sales, but higher sales figures do not necessarily mean manufacturers capture a proportional share of added value.

Slight improvement in gross margin

The financial notes offer another lens for assessing revenue quality. Foxconn Industrial Internet Co., Ltd. disclosed that in certain business operations, customers also serve as suppliers of some raw materials, a model known as Buy and Sell. The company, based on factors such as contractual responsibilities, inventory risk bearing, and pricing power, determines that it acts as the principal in these transactions and therefore recognizes revenue at the full transaction value. Given the judgments involved in determining the principal, inventory risk bearer, and independent pricing capability, this revenue recognition has been flagged by auditors as a key audit matter for several consecutive years. This accounting treatment is not abnormal in itself, but it implies that the company's revenue scale must be evaluated alongside its cost structure. For manufacturing businesses with a high proportion of material costs, revenue growth could stem from higher shipment volumes or from increased product value, with each having different implications for actual profit contribution. The company has not disclosed the specific share of the Buy and Sell model within its AI server business, so it is not possible to gauge the extent to which AI server revenue is affected by this model. A more prudent approach is to examine operating revenue alongside gross profit, gross margin, and operating cash flow to determine how much profit and cash benefit has ultimately been retained from business scale expansion. This was especially evident in 2024, when AI server revenue grew over 150.00% and cloud computing revenue rose 64.37%, yet the cloud computing gross margin actually declined from 5.08% to 4.99%. The revenue surge first translated into manufacturing scale, not margin expansion.

From servers to full racks: broadening the value frontier

The changes in 2025 carry more fundamental significance. The cloud computing gross margin recovered to 5.73%, and while it remains a typical low-margin hardware business, the gross profit from cloud computing jumped from RMB 15.953 billion to RMB 34.533 billion, already contributing 54.83% of the company's total annual gross profit. The company's business scope has further expanded from server systems and components to rack-level systems, high-speed networking, liquid cooling, and power management. This suggests that Foxconn Industrial Internet Co., Ltd. is attempting to counterbalance the natural constraint of high upstream core component concentration. Its approach is not to move away from core chip suppliers but to add segments where its own involvement and added value are greater, including rack design, liquid cooling, power management, high-speed switching, and system integration. In the first half of 2026, the company explicitly disclosed coverage of full-rack design, liquid cooling, and power management; during the same period, GPU AI rack shipments grew 3.20 times year-over-year, while ASIC AI rack shipments grew 3.00 times. Therefore, rising supplier concentration alone is not sufficient evidence of deteriorating fundamentals. The real question is whether the company can capture increasing added value from manufacturing, design, and system integration within a highly concentrated upstream system. The gross margin recovery in cloud computing in 2025 provides initial evidence, but it is not yet enough to prove that the business model has completed its upgrade.

Profit improving while financial flexibility narrows

Another constraint comes from capital occupation. In 2025, the company's inventory rose from RMB 85.266 billion to RMB 150.913 billion, a year-over-year increase of 76.99%; short-term borrowings grew from RMB 35.992 billion to RMB 104.229 billion, up 189.59%; meanwhile, net operating cash flow dropped from RMB 23.820 billion to RMB 5.238 billion. The company attributed the cash flow decline to strong AI server demand and increased stockpiling. In the first half of 2026, inventory further rose to RMB 192.109 billion, and accounts payable reached RMB 178.459 billion, up 27.30% and 29.36% respectively from the end of 2025; short-term borrowings remained stable at around RMB 102.325 billion. Net operating cash flow recovered to RMB 7.391 billion year-over-year, but relative to revenue of RMB 557.861 billion, it remains modest.

The revenue side of this AI expansion cycle has been substantial. With rising upstream procurement concentration and a growing share of direct material costs, high-value chips and components increasingly amplify revenue scale. Building on this, the company is extending into full racks, liquid cooling, power management, and high-speed networking, aiming to increase its value share within the entire computing infrastructure by broadening system integration. The notable growth in cloud computing gross profit in 2025 shows that this shift is beginning to reflect in profits, but at the same time, inventory, accounts payable, and short-term financing have also expanded in tandem, placing higher demands on working capital and supply chain management capabilities. In 2025, the recovery in cloud computing gross margin and the fact that profit contribution crossed the halfway mark indicate that the AI business is no longer just expanding revenue scale; simultaneously, the rapid growth in inventory, accounts payable, and short-term borrowings also demonstrates that high growth is significantly raising the company's requirements for supply chain organization and working capital. Moving forward, the key metric for evaluating the AI business should shift to how much gross profit new revenue ultimately retains and how much capital it ties up. Cloud computing gross margin, gross profit contribution, the share of system-level products, and inventory and financing scales will better reflect the quality of this expansion than AI server growth rates alone.