Foxconn partners with Intel Corporation (INTC.US) to aggressively enter the AI infrastructure market: Breaking free from "iPhone dependence" and seizing the trillion-dollar AI server market.

On Thursday, Foxconn announced a strategic partnership with Intel Corporation (INTC.US), where both parties will collaborate on the research and deployment of next-generation AI infrastructure and intelligent computing platforms. As the world's largest electronics manufacturing contractor, Foxconn stated in a press release that the partnership aims to create a comprehensive AI solution covering chips, racks, systems, and application layers. The two companies plan to collaborate on devices used in AI data centers, including server racks equipped with Intel Corporation's Xeon processors and AI accelerator chips. They will also focus on high-speed interconnection technology, thermal design, and energy efficiency solutions in AI systems. Foxconn and Intel Corporation also plan to develop AI systems for applications beyond traditional data centers, including factories, smart cities, and Siasun Robot & Automation. Neither Foxconn nor Intel Corporation disclosed the specific financial amount of this collaboration, nor did they announce any customer names or product release schedules. Foxconn's Chairman and CEO, Liu Young-way, stated in the press release, "Our partnership with Intel Corporation combines the strengths of both companies in computing platforms, system integration, and global supply chain capabilities." The two companies also stated that they will explore cooperation in custom chip and system integration solutions. This collaboration reflects the strategic transformation and deep layout of the two tech giants in the artificial intelligence wave. For Foxconn, the world's largest electronics manufacturing contractor, this partnership is an important step in shedding traditional labels. In the past, Foxconn was known for heavily relying on assembling consumer electronics such as smartphones. However, facing market saturation in consumer electronics and a restructuring of the global supply chain, Foxconn has already started its "3+3" strategic transformation with core businesses in electric vehicles, digital health, and Siasun Robot & Automation, combined with core technologies in artificial intelligence, semiconductors, and next-generation communications. The financial data of Foxconn reveals a profound transformation in its business landscape. In the third quarter of 2025, cloud computing and network business (including AI servers) accounted for 42% of the company's revenue, surpassing the consumer electronics sector for the first time and becoming the largest source of revenue. This structural shift means that this company, once known globally for assembling iPhones, is now being rewritten as the "core builder of AI infrastructure." At the earnings conference for the first quarter of 2026, Chairman Liu Young-way confirmed that the shipment volume of AI cabinets for the full year of 2026 will achieve multiple growths, with high double-digit growth in shipment volume in the first quarter, increasing quarter by quarter. The company's management explicitly stated that capital expenditures in the next three years will prioritize the expansion of AI server capacity, rather than traditional smartphone assembly lines. Foxconn's acting CEO, Jue Zheng-heng, said, "The share of cloud products (including general servers and AI servers) is close to 50% (48%), showing that the Foxconn Group's layout in the field of AI has brought a structural shift." Jue Zheng-heng also revealed that the revenue share of AI servers in Foxconn's cloud products has exceeded 50% in the first quarter of this year. According to estimates, the annual sales volume of AI servers is expected to reach a level of "NT$2 trillion." In the AI wave, Foxconn deeply understands that the future computing center is no longer just a simple stack of components but a highly integrated complex physical system. By collaborating with Intel Corporation to develop server racks equipped with Xeon processors and AI accelerator chips, Foxconn is transforming its robust industrial manufacturing and scalable delivery capabilities into essential "hardcore assets" for AI data centers. More importantly, Foxconn's layout in the AI hardware field demonstrates a shrewd strategy of "multi-sided betting." While securing a significant share in high-end GPU server assembly through cooperation with NVIDIA Corporation, Foxconn's partnership with Intel Corporation further enriches its computing product line. This strategy of not being tied to a single supplier not only gives Foxconn stronger bargaining power and risk resistance in the global AI supply chain but also enables it to meet diverse computing needs in different customers and scenarios, achieving a strategic transformation from a "consumer electronics assembly factory" to a "global core builder of AI computing infrastructure." Intel Corporation's AI Transformation: From "Chips" to "Rack-level" In-and-Out Encirclement Intel Corporation has been passive in the AI chip competition. Its Gaudi series accelerators failed to shake NVIDIA Corporation's dominant position in the training market, and the Falcon Shores GPU product line was canceled due to a lack of competitiveness. However, Intel Corporation CEO Chen Liwu made a crucial judgment at the 2026 Taipei Computer Show (Computex): that the focus of the AI industry is shifting from "training" to "inference," which has significantly different requirements for computing architecture. In Agentic AI applications, the token consumption of a single intelligent agent when completing complex tasks can be up to 1000 times the consumption of a single round of inference. This means that the demand for computing power on the inference side will see explosive growth, with a much higher sensitivity to power consumption and latency than the training side. Intel Corporation's choice to cooperate with Foxconn at this time is precisely aimed at seizing this structural window. Chen Liwu's proposal of "stepping out of the single component view and turning to rack-scale systems" marks a significant shift in Intel Corporation's product philosophy. Traditionally, chip manufacturers' competitive focus has been on the performance metrics of single GPUs or CPUs, while rack-level solutions optimize computing, storage, network, cooling, and power management as a holistic system. The bottleneck of modern AI data centers has long surpassed the performance limits of a single wafer. Power consumption, cooling (such as liquid cooling systems), and high-speed interconnect technology between chips are the key factors determining the success or failure of computing centers. Intel Corporation is well aware that only by providing comprehensive solutions from "chip to system level" can it appeal to enterprise customers eager for cost control and efficiency. In this collaboration, Intel Corporation deeply integrates its Xeon processors and AI accelerator chips with Foxconn's advanced rack design and cooling solutions. Intel Corporation sees the golden turning point of the AI industry shifting from "large model training" to "large-scale inference and intelligent agent applications." In inference and edge AI scenarios, enterprises have a more urgent need for systems that are cost-effective, low-latency, and can handle both traditional general computing and AI computing. Intel Corporation's rack-level new thinking of "hardware and software integration, whole-machine delivery" can perfectly unleash the general scheduling advantages of its processors, establishing an exclusive computing moat for Intel Corporation in the broader enterprise AI inference market. The Collaboration of "Transformers" and "Followers" The ambitions of the two giants clearly extend beyond the four walls of traditional data centers. The collaboration is explicitly stated to extend to applications in factories, smart cities, and Siasun Robot & Automation. In the context of the global AI infrastructure investment frenzy, this collaboration is of strategic significance. In 2026, the total capital expenditures of Amazon.com, Inc., Alphabet Inc. Class C, Meta, and Microsoft Corporation, the four major tech giants, amounted to $650 billion, nearly doubling from 2024. Foxconn, as the largest AI server rack assembler for NVIDIA Corporation, is expected to have a market share of over 40% in the AI server market in 2026. Their US AI server factory plans to reach a capacity of 2,000 racks per week in mid-2026. The cooperation with Intel Corporation can be seen as Foxconn's layout in a diversified supply chain - while deeply binding with NVIDIA Corporation, expanding cooperation with Intel Corporation helps reduce the risk of single dependence and seize new opportunities in the inference market. As Intel Corporation CEO Chen Liwu said, "We are not chasing yesterday's AI, but defining tomorrow's AI infrastructure." And Foxconn is trying to be the builder behind that "definer." The collaboration between the two companies may reshape the competitive landscape of global AI data center equipment.