Foxconn is targeting a larger share of the global artificial intelligence server market, aiming to increase its position from about 40% currently to as much as 50% within the next few years as technology companies accelerate investment in computing infrastructure.
Chairman Young Liu said the Taiwanese electronics manufacturer could capture approximately 45% of the market next year before reaching the longer-term target. The ambition would strengthen Foxconn’s position in one of the fastest-growing segments of the technology supply chain, where demand for servers, data center equipment and advanced computing systems has expanded alongside the development of generative AI.
“I think we can grow up to 50 percent,” Liu told CNBC’s Christine Tan in an interview, referring to the company’s share of the global AI server market.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Formally known as Hon Hai Precision Industry, Foxconn is best known for assembling Apple’s iPhones. However, the company has increasingly shifted its growth strategy towards AI infrastructure, seeking to capture more value from the hardware and systems required to train and operate increasingly powerful AI models.
Liu expects AI infrastructure to account for 60% to 65% of Foxconn’s revenue within the next three to five years, excluding edge AI devices that run machine-learning applications directly on local hardware.
The revenue forecast signals a significant change in the company’s business mix. Foxconn’s traditional electronics manufacturing operations have made it a major supplier to global consumer technology brands, but the expansion of AI infrastructure offers an opportunity to build a larger presence in higher-value enterprise and data center markets.
Liu said Foxconn’s production capacity was already fully booked for at least the next 18 months, indicating that customers are committing to substantial volumes of AI infrastructure. Demand has so far been driven mainly by AI model developers and cloud service providers, while governments and traditional enterprises have yet to adopt the technology at the same scale.
AI demand remains “quite strong,” he said.
Foxconn is also planning to expand its manufacturing presence in the United States, where technology companies and policymakers are seeking more domestic capacity for critical computing infrastructure.
Liu said the company intends to move further up the supply chain to serve the US market more effectively. Rather than concentrating only on assembling individual components, Foxconn wants to expand its ability to manufacture more complex systems, including AI server racks and data center equipment.
“What we are going to do is we’re going to move up the supply chain,” he said.
The approach would give Foxconn a broader role in the infrastructure supporting AI computing, potentially extending its activities from server manufacturing into integrated data center systems and computing services.
However, manufacturing in the United States carries higher costs than production in some Asian markets. Liu said Foxconn would initially prioritize products with stronger margins, particularly AI server racks and data center equipment, where the value of each manufactured system can justify higher operating expenses.
He also suggested that greater production scale could eventually reduce those costs. Expanding manufacturing capacity and increasing output could improve efficiency, although the extent of any savings will be determined by labor costs, component sourcing, logistics, and the availability of skilled workers.
The US expansion also comes amid efforts by major technology companies to secure reliable supplies of AI infrastructure. Cloud providers and model developers are investing heavily in computing capacity, creating opportunities for manufacturers capable of delivering large volumes of servers and integrated systems.
Foxconn’s existing position in electronics manufacturing gives it an established production base and relationships with major technology customers. Its challenge is to translate those advantages into a stronger position in the more specialized AI infrastructure market, where performance requirements, supply-chain coordination, and delivery schedules are becoming increasingly demanding.
Moving up the supply chain could also improve Foxconn’s exposure to the economics of AI infrastructure. Manufacturing complete server racks and data center equipment may allow the company to capture more revenue per customer than assembling individual components. However, it could also require greater investment and expose the company to more complex operational responsibilities.
Foxconn Warns Against Excessive AI Investment
Despite its ambitious growth targets, Liu warned that the expansion must be managed carefully, particularly as spending across the AI industry reaches unprecedented levels.
“Overinvestment will kill the company,” he said, stressing that Foxconn would maintain tight control over capital expenditure as it expands its AI business.
The company currently plans to spend between $3 billion and $5 billion annually on AI-related investments, although Liu said expenditure could exceed that range depending on how far Foxconn pursues vertical integration.
Vertical integration would involve taking control of more stages of the production and delivery process, potentially allowing Foxconn to manufacture a wider range of components and systems rather than relying as heavily on external suppliers. While this could improve efficiency and strengthen its position with customers, it would also increase the amount of capital tied up in the business.
Liu’s warning reveals a risk facing companies across the AI supply chain. Demand for computing infrastructure is currently strong, but manufacturers are making investment decisions based on expectations that AI spending by cloud providers, model developers and enterprises will continue expanding.
If those customers reduce their spending, companies that have committed substantial capital to new factories, equipment and production capacity could face lower utilization and weaker returns.
Asked how Foxconn would respond if a major customer cut back on AI spending, Liu said weaker demand from one customer could be offset by growth elsewhere. That approach offers some protection against customer-specific spending changes, particularly if demand remains broad across cloud providers and AI developers.
However, he acknowledged that an industry-wide slowdown would create a much bigger challenge for Foxconn and the wider AI supply chain.
Foxconn is positioning itself to benefit from the next phase of AI infrastructure expansion, with plans to increase server market share, deepen US manufacturing and take on more complex production responsibilities. But Liu’s warning on overinvestment shows that the company is also aware that the current spending boom carries substantial risks.