Question 1

What does Leviathan Systems do?

Accepted Answer

Leviathan Systems provides GPU infrastructure services for AI data centers: rack assembly and integration, structured cabling and fiber optic installation, and network testing and commissioning. We deploy NVIDIA GPU infrastructure (H100, H200, GH200, GB200 NVL72, GB300 NVL72) at scale across the United States.

Question 2

Where does Leviathan Systems operate?

Accepted Answer

We provide services nationwide across the United States. Our team deploys to data center facilities in any state, with active projects currently in Texas and California.

Question 3

What GPU platforms does Leviathan have experience with?

Accepted Answer

We have hands-on deployment experience with NVIDIA H100, H200, GH200, GB200 NVL72, and GB300 NVL72 platforms. We are actively deploying GB300 (Blackwell Ultra) infrastructure, the latest NVIDIA GPU system in production.

Question 4

What size projects does Leviathan handle?

Accepted Answer

We deploy GPU infrastructure from single-rack installations to hyperscale facilities with thousands of racks. Our team of 10–30 technicians scales to match project scope, and we have assembled between 1,000 and 5,000 GPU racks to date.

Question 5

How quickly can Leviathan mobilize for a project?

Accepted Answer

We can typically mobilize a deployment team within one week of project kickoff, depending on scope and location.

Question 6

What does GPU rack assembly include?

Accepted Answer

Our rack assembly service covers the full lifecycle: rail installation and PDU mounting, server and switch placement per the customer’s runbook, power cabling (C13/C19, C19/C20), network cabling (DAC, AOC, fiber), NVLink and InfiniBand interconnect routing, cable management, and power-on self-test (POST) verification.

Question 7

Can you assemble GB200 and GB300 NVL72 racks?

Accepted Answer

Yes. We have direct deployment experience with both NVIDIA GB200 NVL72 and GB300 NVL72 systems. These rack-scale liquid-cooled units require specialized handling due to their 1,360 kg weight, 120 kW power draw, and mandatory liquid cooling integration.

Question 8

What is the difference between GPU rack assembly and traditional rack-and-stack?

Accepted Answer

Traditional rack-and-stack involves mounting servers and connecting basic power and Ethernet. GPU rack assembly adds high-density power distribution (40–120+ kW per rack), specialized interconnects (NVLink, InfiniBand), precision cable routing for signal integrity, liquid cooling integration, and comprehensive testing. The skill set, tooling, and process rigor are significantly different.

Question 9

What OEM server platforms do you work with?

Accepted Answer

We deploy GPU infrastructure from major OEMs including NVIDIA DGX systems, Dell PowerEdge with HGX, HPE ProLiant/Cray, SuperMicro GPU servers, and custom-built platforms. The assembly process adapts to each manufacturer’s specifications and runbooks.

Question 10

How do you handle GPU rack weight and floor loading?

Accepted Answer

GB200 and GB300 NVL72 racks weigh approximately 1,360 kg (3,000 lbs). We verify floor loading capacity (minimum 21 kN/m² for distributed load), plan the path of travel from loading dock to final position, and install seismic bracing where required. Our pre-deployment site survey identifies any structural constraints before equipment arrives.

Question 11

What types of cabling does Leviathan install?

Accepted Answer

We install OM4 and OM5 multimode fiber, OS2 single-mode fiber, MPO/MTP trunk cables, Cat6A copper, DAC (Direct Attach Copper), AOC (Active Optical Cable), and AEC (Active Electrical Cable). Cable type selection depends on distance, speed, and application requirements.

Question 12

How many cable connections can you deploy per project?

Accepted Answer

Our projects range from hundreds to tens of thousands of individual connections. A 1,000-GPU cluster typically requires 3,000–5,000+ cable connections across power, network, management, and storage fabrics.

Question 13

What standards do you follow for cabling installation?

Accepted Answer

All installations follow TIA-942 (data center infrastructure), TIA-568 (cabling performance), and BICSI standards. Our technicians use certified installation practices for proper bend radius, pull tension limits, connector termination, and labeling.

Question 14

Do you install cable tray systems?

Accepted Answer

Yes. We design and install overhead cable tray systems (ladder tray, wire basket, trough) as part of the structured cabling scope. In AI data centers, cable tray must accommodate both data cables and liquid cooling hoses, requiring dedicated separation and pathway planning.

Question 15

What is your approach to cable management and labeling?

Accepted Answer

Every cable receives a standardized label at both ends with a unique identifier that maps to our cable documentation. We use color coding to differentiate cable types (fiber, copper, power, management) and install horizontal and vertical cable managers throughout the deployment. This structured approach reduces troubleshooting time by enabling rapid identification and tracing of individual cables among thousands.

Question 16

How do you test fiber optic connections?

Accepted Answer

We use OTDR (Optical Time Domain Reflectometer) testing to map fiber characteristics, plus power meter and light source testing to verify insertion loss and return loss against TIA-568 link budgets. Every connection is tested and documented before handoff.

Question 17

What documentation do you provide after testing?

Accepted Answer

We deliver complete test reports for every connection, cable maps showing physical routing and port assignments, as-built drawings, and serial number / asset tracking documentation. This becomes the baseline reference for the facility’s operations team.

Question 18

What is your first-pass test success rate?

Accepted Answer

Our structured cabling installations consistently achieve high first-pass test success rates, with any failures identified and remediated before handoff. We attribute this to our rigorous installation practices, proper cable handling, and connector termination quality.

Question 19

Do you perform burn-in testing on GPU hardware?

Accepted Answer

Yes. We perform burn-in testing as part of our commissioning process, running GPU hardware at sustained high load for 24–72 hours to identify early failures. Burn-in stresses power delivery, cooling, memory, and interconnects simultaneously, ensuring components that enter production are statistically reliable.

Question 20

What is POST verification and when do you perform it?

Accepted Answer

POST (Power-On Self-Test) verification confirms that each server correctly initializes after deployment: BIOS loads, all GPUs are detected, memory is at the correct capacity, network interfaces are active, and management is accessible. POST is the first validation step after physical deployment, performed before burn-in testing.

Question 21

Does Leviathan install liquid cooling systems?

Accepted Answer

We integrate liquid cooling infrastructure as part of GPU rack deployments: connecting rack-level manifolds to facility CDUs, installing quick-disconnect fittings, performing leak testing, and verifying coolant flow rates. We coordinate with cooling system vendors and facility engineers on CDU placement and facility water loop connections.

Question 22

Is liquid cooling required for GB200 and GB300 systems?

Accepted Answer

Yes. NVIDIA GB200 NVL72 and GB300 NVL72 systems require 100% direct liquid cooling. There is no air-cooled option. Each rack draws approximately 120 kW, which exceeds what air cooling can physically remove. Any facility planning to deploy Blackwell-generation hardware must have liquid cooling infrastructure in place.

Question 23

What liquid cooling components do you install?

Accepted Answer

Our liquid cooling scope includes coolant distribution unit (CDU) placement and connection, rack-level manifold routing with quick-disconnect fittings, leak detection systems (both drip and zone-based), isolation valves for maintenance, and integration with facility chilled water loops or dry coolers. We pressure test every connection before coolant fill.

Question 24

How is Leviathan different from general data center contractors?

Accepted Answer

General contractors handle traditional server installations at 5–15 kW per rack. Leviathan specializes exclusively in GPU infrastructure at 40–120+ kW per rack. The power density, cooling requirements, interconnect complexity (NVLink, InfiniBand), and testing rigor are fundamentally different. Our team works on GPU deployments every day — this is all we do.

Question 25

Do you work as a subcontractor or prime contractor?

Accepted Answer

Both. We work as a prime contractor for organizations that need GPU infrastructure deployed in their facilities, and as a specialized subcontractor for general contractors and systems integrators who need GPU-specific expertise on larger data center projects.

Question 26

How do I get a quote for a project?

Accepted Answer

Contact us through our website form with your project details: GPU platform, number of racks, facility location, and desired timeline. We respond within 48 hours with an initial scope assessment.

Question 27

What information do you need to provide a quote?

Accepted Answer

Minimum: GPU platform (H100, GB200, GB300, etc.), approximate rack count, facility location, desired timeline, and which services you need (assembly, cabling, testing, or all three). We’ll follow up with detailed questions about facility readiness, power availability, and cooling infrastructure.

Question 28

What is your typical project timeline?

Accepted Answer

Timelines vary by scope. A single-rack deployment can be completed in days, while hyperscale projects with hundreds of racks run weeks to months. We provide a detailed timeline estimate during the scoping process, including mobilization, deployment phases, testing, and commissioning milestones.

Question 29

Do you provide ongoing support after deployment?

Accepted Answer

Our primary scope is deployment and commissioning. We deliver a fully documented, tested infrastructure baseline and hand off to the facility’s operations team. For clients who need ongoing support, we can discuss maintenance agreements and return engagements for expansion projects.

Question 30

Can you help with data center site selection or design?

Accepted Answer

We provide pre-deployment site surveys that assess power capacity, cooling infrastructure, floor loading, and cable pathway readiness for GPU deployments. While we don’t design data centers, our surveys identify critical infrastructure gaps before equipment arrives, preventing costly delays during deployment.

Frequently Asked Questions_

About Leviathan Systems

GPU Rack Assembly

Structured Cabling

Network Testing

Liquid Cooling

General

Still Have Questions? \u2014 Let\u2019s Talk_