Quanta Computer

廣達 / 廣達電腦股份有限公司

Taiwan's largest AI-server-focused original design manufacturer builds and integrates rack-scale systems around Nvidia and AMD accelerators for hyperscale cloud customers.

TWSE: 2382
Profile as of September 13, 2026
#9 in the AsiaAI Tech Index
How the ranking works →

Quanta is a direct beneficiary of hyperscaler AI infrastructure spending, converting Nvidia accelerator supply into deployable data-center capacity. Its main strength is execution at scale across customer-qualified AI-server programs, but its growth remains dependent on accelerator allocation, hyperscaler orders, and percentage-margin discipline.

Key figures

Revenue FY2025: NT$2.12 trillion
Revenue growth FY2025: 50.5%
Operating margin FY2025: 4.1%
Net income FY2025: NT$74.99 billion
R&D spend FY2025: Not disclosed
Capital expenditure FY2025: Not disclosed
Market capitalization 2026-09-11: NT$1.38 trillion Estimate
AI servers share of server revenue H1 2025: More than 60%
Worldwide server-assembly contracting volume Estimated: Approximately 17%
FY2025 capital expenditure budget Q1 2025: NT$20 billion
Thailand SMT production lines October 2025: About 80

Overview

Quanta Computer was founded in 1988 and has grown into Taiwan's largest original design manufacturer for AI servers by revenue scale. While it maintains a substantial legacy business manufacturing notebook computers, its structural position has shifted toward data-center infrastructure. As an ODM, Quanta supplies complete AI-server systems and rack-scale infrastructure, including compute-server motherboards, thermal designs, power integration, and networking integration. It operates manufacturing facilities in Taiwan, the United States, Thailand, and Mexico.

Quanta does not fabricate its own silicon, GPUs, or advanced packages. Instead, it translates complex accelerator reference architectures into high-volume, hyperscaler-qualified systems. This integration burden requires coordinated delivery of GPUs, CPUs, networking, memory, trays, cables, cooling, and power. Its installed engineering processes and customer qualification history reduce execution risk for cloud providers, making it a critical partner for deploying data-center capacity at scale.

The AI angle

AI demand reaches Quanta directly through its server business rather than through an indirect component exposure. The company engineers and manufactures rack-scale systems built around Nvidia and AMD accelerators for hyperscale cloud customers. Its relevant products include Nvidia GB200 and GB300 systems, AI-server racks, and liquid-cooling configurations. Quanta began shipments of Nvidia GB200 systems at the end of March 2025, with GB300 systems ramping during the second half of the year.

AI servers accounted for more than 60% of server revenue in the first half of 2025, and management indicated a path to about 70% for the full year. This AI-driven demand caused AI-server revenue to double in FY2025, lifting consolidated group revenue to NT$2.12 trillion.

Quanta captures value in the AI Server Hardware and Power & Infrastructure layers by converting constrained accelerator platforms into qualified, rack-scale systems that hyperscalers can deploy in volume. However, because it captures manufacturing and integration value rather than scarce silicon value, its revenue remains highly dependent on Nvidia's platform supply, component allocation, and hyperscaler purchase orders.

Technology and moat

Quanta's competitive advantage is its ability to translate complex accelerator reference architectures into high-volume, hyperscaler-qualified systems across several manufacturing regions. The difficult elements are not GPU design, but rather board and rack engineering, thermal integration, liquid-cooling implementation, power delivery, manufacturing test automation, supply-chain allocation, and customer-specific validation.

Nvidia's GB200 and GB300 rack-scale platforms increase this integration burden, requiring coordinated delivery of multiple hardware and software components. Quanta's installed engineering processes, customer qualification history, and manufacturing operations in Taiwan and the United States reduce execution risk for customers. While this does not create a permanent technological monopoly, it establishes high barriers to entry. A competitor must replicate customer-specific design validation, supply-chain allocations, thermal and power integration, and rack-scale delivery reliability to displace Quanta on a major program.

Five-pillar assessment

Scale and market position: Taiwan's largest AI-server-focused ODM by revenue scale, holding an estimated 17% of worldwide server-assembly contracting volume and generating NT$2.12 trillion in FY2025 revenue.
Technology and R&D: Expertise in system engineering, thermal integration, liquid-cooling implementation, and rack-scale deployment for advanced platforms like Nvidia's GB200 and GB300.
Supply-chain centrality: Critical manufacturing partner for Microsoft, Amazon, Google, and Meta, while heavily dependent on Nvidia and AMD for accelerator supply.
Financial momentum: FY2025 revenue grew 50.5% as AI-server revenue doubled, though gross margins face pressure from high component costs and foreign exchange effects.
Governance and quality: Independent public company led by its founders, with a seven-member board including three independent directors, managing significant geopolitical and geographic concentration risks.

Supply chain and relationships

Quanta's upstream dependencies include Nvidia GPUs and Grace CPUs, AMD processors, server CPUs, HBM-equipped accelerator modules, printed circuit boards, power supplies, liquid-cooling assemblies, memory, storage, and switches. The company depends especially on Nvidia's platform supply and component allocation, which can constrain Quanta's revenue even when end demand remains strong.

On the downstream side, Quanta's end customers are not formally named in its financial statements, but reporting identifies its main AI-server projects as serving Microsoft, Amazon, Google, and Meta. This gives a small number of hyperscale buyers substantial influence over volume, platform allocation, and pricing. Quanta relies on its own subsidiary, Quanta Cloud Technology, for cloud-data-center and server-system capabilities.

Customers

MicrosoftReported AI-server projects for cloud infrastructure
AmazonReported AI-server projects for cloud infrastructure
GoogleReported AI-server projects for cloud infrastructure
Meta PlatformsReported AI-server projects for infrastructure

Suppliers

NVIDIAAccelerator platforms including GB200 and GB300
AMDProcessors where customers specify them

Partners

Quanta Cloud TechnologySubsidiary supplying cloud-data-center capabilities

Competitors

FoxconnInferredCompetes in AI servers and rack-scale hardware
WistronInferredCompetes for AI-server and general server ODM programs
WiwynnInferredTargets hyperscale cloud-server and AI-infrastructure programs
InventecInferredTaiwanese ODM supplier competing in server manufacturing
PegatronInferredCompetes for selected server and system programs
SupermicroInferredSupplies AI server systems, often branded directly
Dell TechnologiesInferredCompetes for branded AI-server deployments
Hewlett Packard EnterpriseInferredCompetes for branded AI-server programs

Geopolitics and risk

Quanta retains material engineering and manufacturing exposure in Taiwan, exposing it to cross-strait conflict, blockades, logistics disruptions, or energy shortages that would interrupt production and component flow. Furthermore, AI-server systems containing advanced US-origin accelerators face export-control and customer-compliance constraints, limiting addressable configurations and destinations.

To mitigate geographic concentration, Quanta is expanding its manufacturing capacity overseas. It operates facilities in the United States, Mexico, and Thailand, planning to exceed 100 surface-mount-technology production lines in Thailand by the second half of 2026. While this expansion reduces sole dependence on Taiwan and improves delivery lead times for North American hyperscalers, it introduces labor, supplier-localization, permitting, quality-control, and ramp risks that can pressure margins.

Risk matrix
Risk	Severity	Why it matters
Nvidia accelerator allocation	High	Delayed accelerator allocation can defer high-value system revenue.
US-China technology controls	High	Export-control constraints limit addressable configurations and destinations.
Taiwan Strait disruption	High	Conflict or blockades would interrupt production and component flow.
Hyperscaler demand concentration	High	A small number of buyers have substantial influence over volume and pricing.
AI-platform transition risk	High	Revenue depends on orderly transitions to GB200 and GB300 systems.
Gross-margin pressure	Medium	Expensive components and foreign exchange effects can compress percentage margins.
Overseas capacity execution	Medium	Expansion introduces labor, supplier-localization, and ramp risks.
Notebook-cycle exposure	Medium	Group revenue remains partly exposed to mature PC demand cycles.

Governance and ownership

Quanta is an independent, publicly listed Taiwanese corporation. It is not state-owned, and there is no disclosed evidence of state influence. Founder and Chairman Barry Lam is central to governance and group leadership, alongside Vice Chairman and President C.C. Leung.

The board consists of seven directors, including three independent directors, all of whom are natural persons with no institutional-shareholder directors. Regulatory posture is shaped principally by Taiwan Stock Exchange disclosure obligations, Taiwan corporate-governance rules, export-control compliance, environmental and labor regulation across production sites, and customer standards for data-center hardware supply chains.

What to watch

Whether AI servers sustain approximately 70% of Quanta's server revenue as next-generation Nvidia systems replace earlier platforms.
Whether GB300 shipment volume remains above GB200 volume through subsequent platform ramps.
Whether Quanta's US expansion and Mexico capacity additions reduce delivery lead times without eroding margins.
Whether the Thailand SMT-line count exceeds 100 in the second half of the stated target year.
Whether FY2026 operating margin holds near or exceeds the FY2025 level of approximately 4.1%.
Whether export-control changes alter permitted configurations or customer destinations.

Recent developments

May 8, 2026Quanta's FY2025 annual report states that AI-server revenue doubles and group revenue reaches NT$2.12 trillion.
March 3, 2026Quanta reports FY2025 net income of NT$74.99 billion, compared with NT$59.70 billion in FY2024.
October 22, 2025Quanta says GB300 AI-server shipments increase steadily and expects GB300 volume to exceed GB200 volume during the quarter.
August 12, 2025Quanta's FY2025 second-quarter materials report record revenue and profit, with AI servers exceeding 60% of server sales in the first half.
June 13, 2025Quanta elects its 15th board of directors for a term ending 2028-06-12.
May 22, 2025Quanta says shipments of Nvidia GB200 servers begin at the end of March 2025, while GB300 shipments are expected to start in the second half of 2025.

Coverage on AsiaAI.FYI

Can Anthropic Save Japan’s Three Largest Tech Giants From AI Obsolescence? August 14, 2026

Guides that cover Quanta Computer

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. Audited FY2025 revenue, net income, leadership, and board composition are well supported, but R&D expense, actual capex, and named customer revenue shares are not disclosed.

Main sources: Quanta annual reports and investor-relations materials; Quanta board and executive disclosures; Taiwan Stock Exchange-related company disclosures; Quarterly earnings-call reporting; Market-data sources; Industry and financial press.

All 69 sources

Position in the AI stack

An original design manufacturer that engineers and builds rack-scale AI servers and infrastructure for hyperscale cloud providers.

Legal name: Quanta Computer Inc.
Headquarters: Taoyuan City, Taiwan
Founded: 1988
Founders: Barry Lam / 林百里
C.C. Leung / 梁次震
Leadership: Barry Lam / 林百里, Founder and Chairman
C.C. Leung / 梁次震, Vice Chairman and President
Ownership: Publicly listed
Listings: TWSE: 2382
Business: AI server and notebook original design manufacturing
Sector: Consumer Devices & IoT
Servers, PCs & ODMs
Website: quantatw.com