CNEX has secured early commitments and is actively allocating limited GB300 capacity among enterprise customers. Demand is validated, pipeline is structured, and revenue visibility is strong.
Committed ARR
Ancapex AI — reserved capacity
Qualified Customers
Across enterprise and AI-native verticals
Total Rack Demand
Validated pipeline requirements
Total Pipeline Ceiling
$470M–$680M gross range
Capacity Is Being Allocated
CNEX's GB300 systems are deploying sequentially, with each unit entering active allocation before the prior unit reaches full utilization. This structure ensures continuous revenue generation and maximizes commitment density across the pipeline.
GB300 System #1 — Reserved
Fully reserved by Ancapex AI at approximately $11.4M annual value. Contractual commitment in place. Infrastructure deployed and operational. This system establishes CNEX's per-rack revenue benchmark and serves as the reference point for all subsequent allocations.
GB300 System #2 — In Allocation
Multiple enterprise customers have submitted Letters of Intent and are currently under active review. Allocation will be prioritized based on commitment size, contract structure, and strategic fit. Customers include unicorn-stage platforms and well-funded AI-native companies.
GB300 System #3 — Pre-Allocation Signal
Early-stage demand has already been identified and mapped from within the existing qualified pipeline. The third system is expected to reach pre-commitment status prior to full deployment, consistent with the trajectory established by systems one and two.
Capacity Deployment & Allocation Timeline
Demand is forming ahead of infrastructure deployment — each GB300 system enters an allocation queue before it becomes available, ensuring minimal gap between capacity and revenue generation.
GB300 #1
Status: Reserved
Customer: Ancapex AI
Revenue: $11.4M ARR
Fully committed. Operational.
GB300 #2
Status: In Allocation
State: Multiple LOIs under review
Outcome: Allocation based on commitment strength and contract terms
GB300 #3
Status: Pre-Allocation Signal
State: Demand mapped from pipeline
Note: Expected pre-commitment prior to deployment
Enterprise Customers Competing for Capacity
The following organizations have submitted Letters of Intent for GB300 System #2 and are currently under active allocation review. Ranking reflects budget scale, strategic alignment, and contract structure suitability.
A Diversified, High-Quality Demand Base
CNEX's qualified pipeline spans eight distinct verticals, reducing concentration risk and validating broad enterprise adoption of dedicated AI infrastructure. Each segment represents customers with defined, immediate compute requirements.
AI-Native Model Builders
~11.75 racks
Foundation model developers and AI-first platforms requiring sustained high-throughput compute
Biotech & Drug Discovery
~11 racks
Pharmaceutical and life sciences firms running large-scale molecular simulation and genomics workloads
Media & Real-Time AI
~11–13 racks
Live-streaming platforms and generative media companies requiring ultra-low-latency inference
Financial Services
~4 racks
Risk modeling, algorithmic trading, and compliance-grade AI workloads in regulated environments
Academia & Healthcare
~3 racks
Research institutions and health systems with complex, long-horizon compute projects
Sovereign & Government
~5 racks
National AI initiatives and government agencies requiring sovereign, secure infrastructure
Industrial & Applied AI
~2.75 racks
Manufacturing, logistics, and industrial optimization use cases with consistent baseline demand
Consulting & Enterprise Multipliers
~2.2 racks
Systems integrators and managed service providers deploying AI on behalf of enterprise clients
Total demand: 52–58 racks across ~58–60 qualified customers
Demand Significantly Exceeds Near-Term Deployable Capacity
Demand Side
52–58 racks of validated requirements across 58–60 qualified customers
$470M–$680M gross pipeline value
$330M–$475M risk-adjusted pipeline
Demand is diversified across eight enterprise verticals with defined use cases and near-term deployment timelines.
Capacity Side
1–3 GB300 racks available near-term
Each system allocated sequentially with LOI-based prioritization
Additional capacity deployment tied to capital deployment and infrastructure buildout schedule
Limited immediate availability is a structural feature, not a constraint.
Pipeline Designed for Conversion
CNEX's pipeline is structured around access scarcity rather than sales velocity. Customers advance through stages by demonstrating commitment strength, not simply by expressing interest. This structure produces a high-quality, conversion-ready pipeline at every stage.
Per-Rack Revenue Model
Each GB300 rack functions as a cash-flowing infrastructure asset with predictable revenue, high utilization, and strong margin characteristics. Dedicated workload allocation ensures consistent consumption without the variability inherent in shared cloud environments.
Infrastructure as an Asset
At $9M–$12M per rack annually with ~50%+ gross margins and a 12–18 month payback period, each GB300 system delivers infrastructure-grade economics with software-grade margins. Long-duration contracts provide revenue predictability that most cloud infrastructure models cannot replicate.
Revenue Scales Directly with Deployed Capacity
CNEX's revenue model is straightforward: each rack deployed against a committed customer generates predictable, recurring revenue. The following metrics represent the total addressable revenue opportunity within the current qualified pipeline.
Pipeline Ceiling
Total gross pipeline value ($470M–$680M range)
Risk-Adjusted Pipeline
$330M–$475M after probability weighting
Max Rack Demand
52–58 racks of validated customer requirements
Peak Revenue Density
$9M–$12M per rack per year
Revenue visibility improves with each rack deployed. As capacity scales, CNEX converts qualified pipeline to committed ARR with minimal sales friction — access to infrastructure is the primary value driver.
Structural Shift in AI Infrastructure
The enterprise AI compute market is undergoing a fundamental reorientation. Organizations that once relied on shared public cloud GPU pools are encountering capacity constraints, pricing volatility, and compliance gaps that make on-demand infrastructure increasingly unsuitable for production AI workloads.
Shared Cloud Constraints
Public GPU cloud environments face persistent availability gaps, unpredictable pricing, and multi-tenant performance variability — making them unsuitable for enterprise SLA requirements and regulated AI workloads.
Dedicated Infrastructure Demand
Enterprises are increasingly requiring dedicated, isolated compute environments that guarantee performance, compliance, and data sovereignty — capabilities that shared infrastructure architecturally cannot provide.
Reserved Capacity as Strategy
AI infrastructure is shifting from a variable, on-demand cost center to a reserved, long-duration strategic asset — analogous to the evolution from cloud storage to enterprise data centers in prior infrastructure cycles.
AI infrastructure is shifting from on-demand usage to reserved capacity — from a commodity cost line to a strategic enterprise asset.
Infrastructure Aligned with Demand
CambridgeNexus is building AI infrastructure that is deliberately aligned with a diversified and validated demand base. Capacity is not speculative — it is allocated to customers with defined workloads, near-term deployment requirements, and the financial commitment to support long-duration contracts.
The CNEX model is built on three principles: allocate only to committed demand, deploy capacity ahead of need, and maintain pricing discipline through structured access. The result is a business with strong revenue visibility, high utilization, and compounding customer relationships across eight enterprise verticals.
Clarity
Revenue tied directly to committed capacity — no speculative build-out, no on-demand variability
Scalability
52–58 racks of qualified demand across 8 verticals provides a defined runway for disciplined expansion
Discipline
Allocation based on commitment strength ensures only the highest-quality customers access limited capacity
Designed for clarity, scalability, and disciplined growth.
©2026 CambridgeNexus, Inc. · [email protected] · GB300 NVL72 · AIFaaS · New England AI Infrastructure
AI Infrastructure Demand with Clear Revenue Visibility