Note 5. Revenue

REVENUE

The Company generates revenue from two service offerings: (i) integrated managed cloud services, and (ii) beginning in the second quarter of fiscal 2026, the Company’s AI Inference Services. Revenue from each offering is recognized in accordance with ASC 606, Revenue from Contracts with Customers. The Company’s revenue recognition policies for each offering are described below.

Managed Cloud Services

The Company generates revenue from integrated cloud computing and managed services provided under Enterprise Service Agreements and related Purchase Orders.

Performance Obligation

Each customer contract contains a single performance obligation — the Company’s integrated managed service — consisting of a series of distinct services that are substantially the same and have the same pattern of transfer. The services include cloud computing capacity, continuous management, monitoring and optimization, technical consultation, enterprise-level support, and 24/7 technical account management. These components are not sold separately, are highly interdependent, and together represent a single combined output.

Although individual components (e.g., cloud computing capacity, managed services) could theoretically be distinct in isolation, they are not separately identifiable within the context of the contract. The Company provides a significant service of integrating cloud capacity, management, monitoring, and support into a single combined output per ASC 606-10-25-21. Accordingly, the contract contains one performance obligation.

Transaction Price and Measure of Progress

The transaction price is determined based on usage-based fees specified in the applicable Purchase Order. Usage-based pricing qualifies for the right-to-invoice exception under ASC 606-10-55-18, as consideration earned in each period corresponds directly with the value transferred to the customer. Revenue is recognized monthly based on actual usage multiplied by contractual unit rates. Variable consideration (service credits) is constrained to amounts for which it is probable that a significant reversal of cumulative revenue recognized will not occur. Based on historical service availability consistently exceeding the 99.95% contractual threshold, no service credits have been issued and no constraint has been applied.

Principal vs. Agent

The Company acts as principal. The Company controls the integrated managed service before transfer to the customer, is primarily responsible for fulfilling all service commitments, bears performance and availability risk regardless of third-party provider performance, and sets pricing independently. Third-party cloud infrastructure is an input the Company transforms and integrates; it is not the specified service. Accordingly, revenue is recognized on a gross basis.

AI Inference Services

Beginning in the second quarter of fiscal 2026, the Company launched its AI Inference Services, under which it provides large language model (LLM) and cloud-related AI inference services, together with technical consultation, service optimization advisory, enterprise support, and dedicated technical account management, to enterprise customers. The AI Inference Services is offered through the Company’s AI portal and is governed by the Company’s standard Master Service Agreements and related Order Forms. Customers access the services through access keys issued by the Company that authenticate calls to the Company’s AI inference endpoints; the underlying compute and model-inference capacity is sourced from one or more third-party providers, with whom customers have no direct contractual relationship.

Performance Obligation

Each contract contains a single performance obligation — the Company’s integrated AI inference service — comprising AI inference capacity together with technical consultation, optimization advisory, enterprise support, and dedicated technical account management. These components are bundled, are not separately priced, and together represent a single combined output. The arrangement represents a series of distinct service periods that are substantially the same and have the same pattern of transfer. Pricing across model classes and service tiers is a billing mechanic and does not give rise to separate performance obligations.

Transaction Price and Measure of Progress

The transaction price is usage-based and is calculated each service month as actual metered usage of each model class multiplied by the contractual unit rate, which equals the Company’s published list price less customer-specific percentage discounts fixed at contract inception. Revenue is recognized over time because the customer simultaneously receives and consumes the benefits of the integrated service as it is performed, and is measured using the right-to-invoice practical expedient under ASC 606-10-55-18, as the Company’s right to consideration corresponds directly with the value transferred to the customer in each period.

Principal vs. Agent

The Company acts as principal in its AI Inference Services. The specified service promised to the customer is the integrated AI inference service delivered under the Company’s service-level commitments — not raw third-party model tokens or capacity. The Company controls that service before transfer to the customer: it designs, configures, and operates the inference platform; selects and may substitute third-party providers without customer consent and at its own cost; is solely responsible to the customer for fulfilling all service commitments; bears continuity, capacity, and performance risk; sets pricing independently of any third-party provider’s rates; and bears credit risk. Accordingly, revenue is recognized on a gross basis, and amounts payable to third-party providers for underlying compute and model-inference inputs are recognized as cost of revenue.