Taiwanese contract manufacturing giant Foxconn (Hon Hai Precision Industry Co., Ltd.), the world’s largest electronics manufacturing services (EMS) provider and primary server assembler for chip designer Nvidia, reported an approximate 47% year-on-year revenue surge for the third quarter ended September 30, 2026.

Revenue for the July–September quarter climbed to T$3.03 trillion (approximately $95.4 billion), comfortably beating consensus analyst forecasts of roughly T$2.85 trillion to T$2.92 trillion. The performance marks the company’s highest-ever revenue haul for a third quarter, fueled by enterprise adoption of AI computing racks and the seasonal assembly ramp of next-generation smartphones.

Key takeaways

  • Record third-quarter revenue: Foxconn generated T$3.03 trillion ($95.4 billion) in the third quarter of 2026, representing an approximate 47% surge year-on-year, easily surpassing market expectations.
  • AI servers outpace expectations: The Cloud and Networking Products division served as the primary growth engine, driven by volume deliveries of high-density AI server systems powered by Nvidia’s advanced computing architectures (including Blackwell and GB200 enterprise racks).
  • September monthly strength: Revenue in September alone crossed T$1.08 trillion, expanding roughly 34% year-on-year and setting an all-time monthly record for the manufacturer.
  • Consumer electronics seasonal ramp: Foxconn’s Smart Consumer Electronics vertical (anchored by its contract manufacturing of Apple’s flagship iPhone series) posted solid sequential expansion as assembly lines in Zhengzhou, China, and Tamil Nadu, India, ramped to peak seasonal output.
  • Positive Q4 momentum: Looking forward, management signaled that fourth-quarter performance is on track to match or exceed historical seasonal peaks, bolstered by long-term cloud hyperscaler orders from Microsoft, Google, and Amazon Web Services (AWS).

Revenue breakdown: The AI server transformation

For decades, Foxconn was known as the assembler of consumer gadgets, with Apple’s iPhone hardware accounting for more than half of its aggregate business.

However, over the past two years, Foxconn has transformed into a high-margin enterprise data center builder:

FOXCONN REVENUE ENGINE: SHIFTING FROM CONSUMER GADGETS TO AI INFRASTRUCTURE

HISTORICAL MODEL:
[ Consumer Electronics (iPhones, Consoles, Tablets) ] ──► ~60%–65% of Total Top Line
[ Enterprise Servers & Networking ] ──────────────────► ~20%–25%

CURRENT OPERATING MIX (2026):
[ Cloud & Networking / AI Server Systems ] ──────────► Surging Share & Primary Margin Driver
- Turnkey liquid-cooled server racks (Nvidia GB200 / NVL72)
- Hyperscaler cloud data center compute boards
- Proprietary thermal cooling loops & power delivery modules
                     │
                     ▼
[ Q3 2026 Top-Line Closes at Record T$3.03 Trillion (+47% YoY) ]
Operating SegmentQ3 2026 Trajectory (YoY)Primary Growth Drivers & Product Lines
Cloud and Networking ProductsSignificant / Exponential GrowthHyperscaler AI clusters, Nvidia GB200 NVL72 racks, AI switches
Smart Consumer ElectronicsStrong Sequential ExpansionSeasonal assembly ramp for next-gen flagship smartphones
Components and Other ProductsSteady Double-Digit GrowthConnectors, precision optics, liquid cooling manifolds, EV parts
Computing Products (PCs/Laptops)Moderate / Stable DemandCorporate PC replacements, enterprise workstations

Source: Compiled from Hon Hai Precision Industry monthly revenue filings and exchange disclosures.

The Cloud and Networking Products segment led the quarter. Management noted that demand for AI servers expanded by triple digits on an annual basis, with server revenue now outpacing traditional personal computer component manufacturing by a wide margin.

The Nvidia alliance: Foxconn as the engine of the GB200 rollout

Foxconn’s revenue beat is tied directly to its role in Nvidia’s hardware ecosystem.

While semiconductor foundries like TSMC fabricate the underlying silicon, turning thousands of AI chips into functional, liquid-cooled, multi-cabinet data center supercomputers requires mechanical, thermal, and electrical integration.

Foxconn has secured a major share of manufacturing orders for Nvidia’s GB200 NVL72 platform—a system integrating 72 Blackwell GPUs and 36 Grace CPUs into a single liquid-cooled enterprise rack:

THE AI SERVER VALUE CHAIN:

[ TSMC ] ──► Advanced Foundry (CoWoS Packaging of Blackwell GPUs & Grace CPUs)
                │
                ▼
[ NVIDIA ] ─► Designs Architecture, CUDA Software Stack & Compute Topologies
                │
                ▼
[ FOXCONN ] ► Turnkey Industrial System Integration:
              - Designs high-efficiency cold plates & coolant distribution units (CDUs)
              - Fabricates high-speed bus interconnect backplanes & power drawers
              - Assembles, tests, and ships complete plug-and-play multi-megawatt server racks
                │
                ▼
[ CLOUD HYPERSCALERS (Microsoft, AWS, Google, Meta, Sovereign AI Hubs) ]

By supplying not just the bare circuit boards, but the complete turnkey server cabinets—encompassing power supplies, high-speed copper interconnects, and advanced liquid cooling distribution units (CDUs)—Foxconn captures a higher dollar value per server rack compared to legacy server assembly.

Consumer electronics seasonal lift: The iPhone factor

While enterprise AI drove the year-on-year growth rate, Foxconn’s consumer electronics division provided the baseline volume for the T$3.03 trillion milestone.

The third quarter coincides with the annual production ramp for Apple’s newest smartphone generation. Foxconn’s massive manufacturing hubs operated at high utilization:

  • Zhengzhou, China (“iPhone City”): Scaled employment rosters with peak seasonal bonuses to assemble premium Pro and Pro Max smartphone variants.
  • Sriperumbudur (Tamil Nadu, India): Continued its expansion under India’s Production-Linked Incentive (PLI) scheme, manufacturing base and premium units simultaneously for domestic retail distribution and international export.

The simultaneous execution across both consumer electronics and enterprise cloud infrastructure allowed Foxconn to cushion minor supply-chain bottlenecks in individual components.

Strategic capital expenditures: Building mega-factories in Mexico and Taiwan

To sustain the delivery of AI supercomputing systems, Foxconn has accelerated capital outlays across its global manufacturing footprint:

  1. The Mexico Mega-Facility: In late 2025 and 2026, Foxconn invested hundreds of millions of dollars to build an AI server manufacturing facility in Guadalajara, Mexico. Situated near the US border, the plant is designed to assemble and ship complete GB200 server racks directly to North American data center hubs operated by Microsoft, Oracle, and Google, minimizing ocean transit times and shielding products from Asia-Pacific shipping friction.
  2. Kaohsiung Advanced Computing Center (Taiwan): In partnership with Nvidia, Foxconn is constructing an advanced computing center in southern Taiwan powered by Blackwell architectures to develop autonomous factory software, digital twin simulations, and smart electric vehicle platforms.
+-----------------------------------------------------------------------------------+
|               FOXCONN'S GLOBAL AI INFRASTRUCTURE MANUFACTURING TRIAD              |
|                                                                                   |
|  NORTH AMERICA (Mexico)       EAST ASIA (Taiwan & China)    SOUTH ASIA (India)    |
|  ---------------------------  ---------------------------   -------------------   |
|  - Guadalajara Mega-Factory   - Advanced R&D and Kaohsiung  - Expanding Bengaluru |
|  - GB200 NVL72 server racks     supercomputing facility.      & Tamil Nadu plants.|
|    built for US hyperscalers. - Precision component tooling - Core smartphone     |
|  - Direct land freight to US    and motherboard production.   assembly & robotics.|
+-----------------------------------------------------------------------------------+

What could happen next

  • Full Q3 2026 investor conference: Foxconn will publish its complete audited third-quarter financial statement on November 14, 2026, detailing net profit after tax, gross margins, and operating cash flows.
  • Hon Hai Tech Day (HHTD): Investors will look to Foxconn’s annual flagship technology showcase for demonstrations of next-generation liquid cooling manifolds, proprietary robotics, and updates on its electric vehicle (EV) commercial partnerships.
  • Blackwell supply ramp validation: Analysts will monitor fourth-quarter shipment volumes to evaluate whether component supplies of high-bandwidth memory (HBM3e) and advanced packaging remain unconstrained as cloud hyperscalers accelerate data center deployments.

Frequently asked questions

How much did Foxconn’s revenue grow in Q3 2026?

Foxconn reported an approximate 47% year-on-year surge in revenue for the third quarter (July–September 2026), reaching a record T$3.03 trillion (approximately $95.4 billion).

Why did Foxconn’s revenue jump so sharply?

The revenue jump was driven primarily by strong enterprise demand for artificial intelligence servers—specifically systems powered by Nvidia’s advanced architectures—alongside the seasonal production ramp of flagship consumer smartphones.

What is Foxconn’s relationship with Nvidia?

Foxconn is Nvidia’s primary manufacturing and system integration partner for advanced AI hardware. Foxconn designs and builds turnkey, liquid-cooled enterprise server cabinets (such as the GB200 NVL72), integrating Nvidia’s GPUs, CPUs, network switches, and custom cooling systems for global cloud data center clients.

Where does Foxconn manufacture its AI servers and smartphones?

Foxconn manufactures AI servers across facilities in Taiwan, China, and Mexico (specifically a mega-plant in Guadalajara serving North America). Its consumer electronics and smartphones are assembled in large-scale complexes across China (such as Zhengzhou) and expanding manufacturing hubs in India (such as Tamil Nadu and Karnataka)

Get the day’s top stories in your inbox

One concise email. No spam, unsubscribe anytime.