Inventec

英業達 / 英業達股份有限公司

A major Taiwanese original-design manufacturer that builds and integrates GPU servers, rack-scale systems, and notebooks for global technology brands.

TWSE: 2356
Profile as of September 13, 2026
#79 in the AsiaAI Tech Index
How the ranking works →

Inventec is scaling rapidly in the AI-server market by manufacturing high-volume GPU systems for major OEMs. However, its low-margin ODM model leaves it highly dependent on upstream accelerator suppliers like NVIDIA and downstream branded customers for pricing power.

Key figures

Revenue FY2025: NT$691.169 billion
Revenue growth FY2025: 6.94%
Operating margin FY2025: 1.8%
Net income FY2025: NT$8.696 billion
R&D spend FY2025: Not disclosed
Capital expenditure FY2025: NT$10.420 billion
Market capitalization Sep 11, 2026: NT$222.6 billion Estimate
AI-server shipments growth 2025: 40% year over year
AI share of server revenue 2025: 40% to 50%
Planned 2026 capital expenditure 2026 plan: US$1 billion

Overview

Inventec was founded in 1975 and has grown into a major Taiwanese original-design manufacturer (ODM). It designs and manufactures notebooks, servers, smart devices, and related electronics for branded customers. Its production footprint spans Taiwan, mainland China, the Czech Republic, Mexico, Malaysia, Vietnam, and Thailand.

The company operates primarily as an assembly and integration partner, converting components into finished systems for enterprise OEMs and cloud deployments. It does not sell its own branded AI accelerators or operate cloud platforms. Instead, it captures value from engineering, motherboard and system integration, manufacturing, validation, thermal design, procurement, and fulfillment.

Inventec's advantage lies in execution rather than proprietary semiconductor technology. It combines ODM engineering, global sourcing, system validation, rack integration, and a broad manufacturing footprint to meet the complex requirements of modern AI server programs.

The AI angle

Inventec benefits directly from AI infrastructure spending by assembling servers and racks for branded OEMs and cloud deployments. Its relevant products include the Artemis II rack system based on NVIDIA GB300 NVL72 and NVIDIA MGX, P9000-series GPU servers using NVIDIA Blackwell and Blackwell Ultra accelerators, and P8500G6 systems supporting AMD Instinct MI300 through MI350 accelerators.

AI demand reaches Inventec through its Enterprise Business Group. Management stated that AI-server shipments increased 40% year over year in 2025, and AI servers represented 40% to 50% of Inventec's server revenue. The business depends heavily on NVIDIA GPUs and platform specifications, AMD Instinct accelerators, Intel processors, and advanced networking and liquid-cooling components.

While Inventec can scale AI-server revenue rapidly, a material share of the high-value content is retained by chip suppliers like NVIDIA and by the branded OEM customers that control final customer relationships.

Technology and moat

Inventec's moat is built on execution, scale, and system-integration capabilities rather than proprietary semiconductor intellectual property. It combines ODM engineering, global sourcing, system validation, rack integration, and a manufacturing footprint across Asia, Europe, and North America.

AI server programs require validated combinations of accelerator modules, CPUs, networking, memory, liquid cooling, mechanicals, firmware, and customer-specific manufacturing processes. Inventec's Artemis II system, developed in collaboration with NVIDIA, illustrates this capability. Such designs require close alignment to NVIDIA's reference architecture, power, and cooling requirements. While this requires substantial qualification work and customer-specific operational knowledge, the underlying GPU, networking, and software intellectual property belongs to partners, limiting Inventec's pricing power and resulting in thin operating margins.

Five-pillar assessment

Scale and market position: A major global ODM with a broad manufacturing footprint across Asia, Europe, and North America, assembling high-volume servers and notebooks for leading technology brands.
Technology and R&D: Strong capabilities in server and rack engineering, thermal integration, and platform validation, though it relies on partners for core semiconductor and software intellectual property.
Supply-chain centrality: Critical assembly partner for Dell, HP, and Lenovo, and closely aligned with NVIDIA, AMD, and Intel for AI server platforms.
Financial momentum: Experiencing rapid growth in AI-server shipments and expanding global manufacturing capacity, though overall operating margins remain constrained at 1.8%.
Governance and quality: Independent public company with significant founding-family influence over management and the board, operating with standard Taiwanese corporate transparency.

Supply chain and relationships

Inventec relies on a concentrated group of technology suppliers. It integrates NVIDIA GPUs and networking components, AMD Instinct accelerators and EPYC processors, and Intel Xeon processors into its server platforms. It also sources chassis, liquid-cooling components, printed circuit boards, and power systems from various suppliers.

On the customer side, Inventec has publicly identified Dell Technologies, HP Inc., and Lenovo Group as major assembly customers for high-end server configurations. The company is an assembly partner for Dell, HP, and Lenovo level 6 and level 10 servers. Inventec does not disclose revenue by named customer, but this concentration means the loss of a major OEM program or a shift in sourcing could materially impact its server business.

Customers

Dell TechnologiesAssembles level 6 and level 10 servers
HP Inc.Assembles high-level server products
Lenovo GroupAssembles high-level server products

Suppliers

NVIDIASupplies GPUs, networking, and MGX platform technology
AMDSupplies Instinct GPUs and EPYC processors
IntelSupplies Xeon processors

Partners

Solomon TechnologyCollaborates on AI-driven smart manufacturing

Competitors

FoxconnInferredCompetes for AI-server and rack-integration programs
Quanta ComputerInferredCompetes for AI-server ODM programs
WistronInferredCompetes for OEM and cloud AI-server manufacturing
WiwynnInferredCompetes in hyperscale server manufacturing
Mitac ComputingInferredCompetes in server ODM capabilities
PegatronInferredCompetes in server ODM capabilities
CelesticaInferredCompetes in selected programs

Geopolitics and risk

Inventec's most significant geopolitical risk stems from its manufacturing footprint and reliance on US-origin technology. The company operates production campuses in Shanghai, Chongqing, and Nanjing, leaving part of its manufacturing and supply chain exposed to cross-strait disruption, trade restrictions, and China-specific regulation.

Additionally, AI-server designs using advanced NVIDIA or AMD accelerators are exposed to changing US export restrictions affecting shipments to China. To mitigate these risks and meet AI-server demand, Inventec is diversifying its global capacity. In March 2026, it announced planned investments exceeding US$300 million in the United States, Mexico, and Thailand, though new capacity requires labor, process qualification, supplier localization, and customer approval.

Risk matrix
Risk	Severity	Why it matters
NVIDIA accelerator allocation	High	Supply constraints for Blackwell and GB300 platforms can defer shipment and revenue.
US export controls	High	Restrictions affect shipments of advanced AI servers to China.
China manufacturing exposure	High	Operations in Shanghai, Chongqing, and Nanjing face cross-strait and regulatory risks.
Customer concentration	High	Reliance on Dell, HP, and Lenovo limits visibility if a program is lost.
Low-margin ODM economics	High	A 1.8% operating margin leaves little protection against inflation or price pressure.
Global capacity transfer	Medium	New capacity in the US, Mexico, and Thailand requires qualification and approval.
Competitor substitution	Medium	Foxconn, Quanta, and Wistron can compete for mature server designs.
Platform transition risk	Medium	Must successfully qualify changing NVIDIA, AMD, and Intel roadmaps.

Governance and ownership

Inventec is an independent Taiwan-listed company. Its governance is shaped by founding-family ownership and board representation rather than state control. Founder Yeh Kuo-i remains a director and held a 4.92% direct stake as of July 2026, while his son, Chairman Yeh Li-Cheng, held 3.27% directly.

The Yeh family exercises controlling influence in practice through family shareholdings, management succession, and board participation, though a complete consolidated family ownership percentage is not disclosed. The board includes independent directors, and the company reports under Taiwan exchange rules.

What to watch

Whether FY2026 server revenue grows faster than overall revenue and AI-server mix increases.
Whether AI-server shipments exceed the 2025 growth rate of 40% as new NVIDIA and AMD systems scale.
Whether the planned US$1 billion of FY2026 capital expenditure is deployed on schedule.
Whether the Artemis II and NVIDIA MGX designs convert into disclosed customer production programs.
Whether operating margin improves above the FY2025 level of 1.8%.
Changes in US export controls affecting China-linked customers or production.

Recent developments

May 27, 2026Re-elected Yeh Li-Cheng as chairman following the expiration of the prior board term.
March 10, 2026Reported FY2025 net income of NT$8.696 billion and planned over US$300 million of investment in the US, Mexico, and Thailand.
January 9, 2026Reported record FY2025 revenue of NT$691.169 billion, up 6.94% year over year.
October 29, 2025Announced collaboration with NVIDIA and Solomon on AI-driven smart manufacturing for servers.
October 3, 2025Announced P9000-series AI rack solutions combining NVIDIA Blackwell GPUs with AMD and Intel processors.
May 19, 2025Presented the Artemis II rack-scale solution developed with NVIDIA at COMPUTEX 2025.

Coverage on AsiaAI.FYI

We haven't written about Inventec yet. Subscribe below to get it when we do.

Guides that cover Inventec

About this profile

Compiled with AI-assisted research from company filings, market data, and published reporting as of September 13, 2026, then reviewed by AsiaAI.FYI. Figures marked Estimate are not company-reported. Check primary filings before relying on any number.

Confidence: B. Financial results are company-reported, while market capitalization uses market quotation and reported shares outstanding; customer-level revenue shares, segment revenue, and R&D expense are not publicly presented.

Main sources: Inventec investor-relations financial-report listings; Taiwan exchange-style corporate disclosures; Company product and news releases; Taiwan financial press.

All 93 sources

Position in the AI stack

An execution and system-integration layer that converts accelerators and CPUs into qualified server and rack deployments for OEM customers.

Legal name: Inventec Corporation
Headquarters: Taipei, Taiwan
Founded: 1975
Founders: Yeh Kuo-i / 葉國一
Leadership: Yeh Li-Cheng / 葉力誠, Chairman
Tsai Chih-an / 蔡枝安, President and General Manager
Employees: More than 38,000
Ownership: Publicly listed
Listings: TWSE: 2356
Business: Original-design manufacturer of servers, notebooks, and smart devices
Sector: Consumer Devices & IoT
Servers, PCs & ODMs
Website: inventec.com