Customer concentration
A few customers account for a very large share of receivables, increasing dependence on individual counterparties.
- Scope
- Three customers accounted for 86% of accounts receivable at March 31, 2026.
- Materiality
- high
Cerebras Systems Inc.
Cerebras Systems Inc. designs and sells AI computing systems built around its wafer-scale chip architecture, along with cloud-based access to its compute infrastructure. The company serves customers that need large-scale training and inference capacity, combining on-premises hardware, hosted cloud capacity, and related software and support services.
| % | |
|---|---|
| Hardware solutions | 45% Sales of Cerebras AI systems and related equipment for customer deployment on-premises. |
| Dedicated Capacity cloud services | 30% Take-or-pay cloud compute capacity contracts for dedicated AI workloads. |
| On-Demand cloud services | 15% Consumption-based inference and training access sold as tokens or time-based usage. |
| Support and managed services | 10% Software support, cluster operations, and related service agreements. |
Cerebras sells to organizations that need very large AI training and inference capacity, including frontier model...
- Frontier model developersprimary
Buy dedicated AI compute and hosted inference for training and serving large models.
- Cloud platform partnersprimary
Buy or resell Cerebras capacity to offer fast inference and AI services to end users.
- Enterprise AI customerssecondary
Buy on-premises systems or cloud access for internal AI workloads and deployment flexibility.
- Managed service customerssecondary
Buy software support and cluster management for systems located in their data centers.
Cerebras is headquartered in the United States and operates from a U.S. base, with customer relationships that extend...
- •Headquartered in the United States
- •Customer demand spans cloud and on-premises AI deployments
- •Data center operations are central to cloud service delivery
- •Third-party foundries support chip manufacturing
- •No country revenue split disclosed in the excerpts
Cerebras is focused on scaling its wafer-scale AI architecture across both hardware and cloud delivery models, which...
01
Expand cloud capacityshort-term
Cloud delivery lets customers adopt Cerebras without owning infrastructure and can scale with workload demand.
02
Win large strategic customersshort-term
Large customers can anchor utilization, validate the platform, and create follow-on demand.
03
Maintain hardware differentiationmedium-term
The wafer-scale architecture is the core technical basis for performance and market positioning.
Cerebras faces concentration risk because a small number of customers can represent a very large share of receivables...
Quarterly revenue volatility
Large individual sales and variable close timing make results difficult to forecast.
- Scope
- Sales may occur later than expected or not at all within a quarter.
- Materiality
- high
Cybersecurity and IT disruption
Cloud services depend on internal systems, data centers, and third-party infrastructure.
- Scope
- Hosted inference, customer data, and internal controls are all exposed.
- Materiality
- high
Growth execution and scaling risk
The company must scale operations, capacity, and support while maintaining product performance.
- Scope
- Cloud capacity build-out and large customer deployments.
- Materiality
- high
Revenue recognition timing
Can shift revenue between quarters and change comparability
Pass-through data center costs
Inflates reported revenue without reflecting core technology economics
Customer warrant asset amortization
Can suppress sequential revenue growth trends
Warranty and inventory estimates
Affects cost of revenue and gross profit
Lease accounting
Affects operating expenses, cash commitments, and leverage-like obligations
: 17.7.2026