Revenue Recognition for AI-as-a-Service (ASC 606 / IFRS 15)#
As AI startups pivot from R&D to commercialization, they face a massive regulatory hurdle: recognizing revenue under the strict 5-step model of ASC 606 (US GAAP) and IFRS 15.
Selling "AI-as-a-Service" (AIaaS) is significantly more complex than selling traditional SaaS software.
Identifying the Performance Obligations#
When an enterprise signs a $1 Million contract with an AI startup, what exactly are they buying? The contract usually contains multiple deliverables:
- Access to the foundational AI model via API.
- The initial "fine-tuning" or training of the model using the client's proprietary data.
- Ongoing support and model updates.
The Accounting Challenge: Are these distinct performance obligations? If the client cannot use the API without the initial fine-tuning, the fine-tuning is not "distinct." The startup must bundle the fine-tuning fee and the API access fee together, and recognize the revenue over the life of the contract, rather than recognizing the massive implementation fee on day one.
Variable Consideration (Usage-Based Pricing)#
Many AI startups charge based on "compute usage" or "API tokens consumed." This creates Variable Consideration. Under ASC 606, the startup cannot just guess how much the client will use. They must estimate the variable consideration and only recognize revenue to the extent it is highly probable that a significant reversal will not occur.
For AI CFOs, correctly structuring customer contracts to ensure favorable, upfront revenue recognition is a critical skill for maintaining a high valuation.