When standard off-the-shelf general purpose servers cannot meet project-specific performance, space, interface, or environmental requirements, BEIJING YINTAI coordinates custom server ODM and Joint Design Manufacturing (JDM) programs through qualified engineering and manufacturing partners. The scope can begin at motherboard or barebone level and extend to complete systems, racks, or containerized delivery.

Partner capabilities cover high-speed interconnects, thermal design, power delivery, firmware, chassis engineering, and production validation. Whether the requirement is a specialized motherboard or barebone (L6), a custom chassis for edge deployment, or an integrated AI server rack (L9/L10), BEIJING YINTAI coordinates requirements, technical reviews, prototyping, validation, and the transition to production.

We act as an engineering liaison and project coordinator between the customer and partner manufacturing platforms. This model supports documented design iterations, component planning, review gates, and responsibility tracking without representing partner facilities or personnel as BEIJING YINTAI-owned assets.

At a Glance

Delivery Levels

L6 (Barebone), L9/L10 (Full System/Rack), L11 (Containerized)

Engineering Scope

Motherboard design, chassis, thermal, power, BIOS/BMC

Manufacturing

Partner-supported trial production and volume manufacturing, subject to project scope

Validation

Agreed EMI/EMC, thermal, signal-integrity, and memory-stability test plans

ODM

Collaboration

ODM (Original Design Manufacturing)

JDM

Engineering Models

JDM (Joint Design Manufacturing)

Delivery Level Matrix

Delivery LevelForm FactorScopeDeployment
L6Motherboard / BareboneCustom motherboard and chassis designIntegrated into partner systems
L9 / L10Full System / RackFully integrated AI servers and racksStandard data center or edge sites
L11ContainerizedPrefabricated modular compute facilityOutdoor or rapid deployment sites

Comprehensive ODM Capabilities

Applications or Use Cases

Custom AI Accelerators: Designing specialized motherboards and topologies to support proprietary AI chips or unique GPU configurations.

Edge Computing Appliances: Developing short-depth, ruggedized servers for telecommunications, manufacturing, or retail environments.

High-Density Cloud Infrastructure: Engineering custom rack-scale solutions optimized for specific cloud workloads, targeting compute density and power efficiency, which can be integrated through our AI compute center design and integration services.

Immersion Cooling Systems: Co-developing native immersion-ready servers and chassis to target low PUE in green data centers under suitable design conditions.

computing center2

Customer Input Checklist

Target Workload: Describe the primary application (e.g., AI training, edge inference, storage).

Hardware Requirements: Specify preferred CPU/GPU platforms, memory, storage, and networking interfaces.

Form Factor and Environment: Detail size constraints, cooling methods (air, liquid, immersion), and environmental conditions (temperature, humidity, vibration).

Volume and Timeline: Estimate the required production volume and target delivery dates for prototypes and mass production.

Delivery Level: Indicate whether you need L6 barebones, L10 full systems, or L11 containerized solutions.

Why BEIJING YINTAI

Full lifecycle coverage

Partner Engineering Access

Coordinated access to qualified partner capabilities in motherboard, system, rack, and container engineering.

Flexible Engagement

ODM and JDM models can accommodate different levels of customer design ownership and technical input.

International delivery model

Cross-Border Project Coordination

English-language technical communication, review records, supply planning, and logistics coordination can be scoped for international projects.

commitments

Lifecycle Coordination

The engagement can extend from feasibility and NRE quotation through validation, production, delivery, and post-deployment support.

Frequently Asked Questions

What is the difference between ODM and JDM?

In an ODM (Original Design Manufacturing) model, our partners design and manufacture the product based on your specifications. In a JDM (Joint Design Manufacturing) model, your engineering team collaborates closely with our partners throughout the design process, sharing development responsibilities.

What is the typical timeline for a custom server project?

Timelines vary significantly based on complexity. A minor modification to an existing design might take a few weeks, while a clean-sheet motherboard design can take several months from concept to mass production. We provide detailed schedules during the NRE quoting phase.

Do you support small-batch production?

Yes, our partner manufacturing platform supports flexible production runs, from small-batch trial production (EVT/DVT) to high-volume mass production (PVT).

How do you ensure the quality of custom designs?

Each project should define its validation plan and acceptance criteria. Depending on scope, partner capabilities can include EMI/EMC, thermal, signal-integrity, memory-stability, fixture, and production-line testing. The exact methods, reports, and applicable certifications are confirmed in the project documentation.

Can you help with global logistics and shipping?

Cross-border logistics and delivery can be evaluated for the destination and product. Final availability depends on export controls, import requirements, carrier capability, product classification, and the agreed delivery terms.