What's Happening?
U.S. organizations are grappling with how to properly account for the costs associated with developing Artificial Intelligence (AI) solutions. The primary challenge lies in determining whether these costs fall under ASC 350, which applies to software
developed for internal use, or ASC 985, which covers software intended for sale, lease, or external marketing. Most current AI initiatives, such as AI accounting assistants, employee support agents, and procurement automation, are generally categorized under ASC 350 because they aim to improve internal operations. However, the specific application depends on the intended audience, use case, and development workflow. The accounting framework dictates when capitalization of costs can begin, with significant differences between the two standards. For instance, under ASC 350, capitalization typically starts during the application development stage after the preliminary project stage is complete, provided management has authorized and committed to funding the project and its completion is probable. Conversely, ASC 985 requires technological feasibility to be established before capitalization can begin, often leading to a later start for capitalizing costs.
Why It's Important?
The correct accounting treatment for AI development costs has significant implications for U.S. businesses, impacting reported earnings, asset balances, and financial transparency. Misclassifying AI development costs can lead to accounting surprises and misrepresentation of a company's financial health. The distinction between internal-use software (ASC 350) and software for external sale (ASC 985) is crucial because their capitalization thresholds differ substantially. This affects when development expenses can be recognized as assets on the balance sheet versus being expensed immediately, directly influencing profitability metrics. Furthermore, the rapid evolution of AI technology introduces complexities in determining useful life, amortization methods, and impairment considerations for capitalized AI assets. Organizations must also navigate the intricacies of AI licensing, especially when relying on third-party large language models or cloud-based AI services, as this impacts intellectual property rights, commercialization strategies, and long-term sustainability. Clear accounting practices are essential for investors and stakeholders to accurately assess the return on AI investments and the overall financial performance of companies heavily investing in AI.
What's Next?
The accounting landscape for AI development is set to change with ASU 2025-06, which will eliminate the current project-stage framework for internal-use software and replace it with a principles-based capitalization threshold. This new guidance, effective for annual reporting periods beginning after December 15, 2027 (with early adoption permitted), will require capitalization of eligible costs when management has authorized and committed to funding the project, and it is probable that the project will be completed and the software will perform its intended function. This shift emphasizes the resolution of significant development uncertainty before capitalization can begin, which is particularly relevant for novel or unproven AI functionalities. Organizations will need to carefully define their software projects, especially in iterative AI development, to apply the capitalization criteria correctly. Additionally, ongoing evaluation of licensing agreements, third-party dependencies, and intellectual property will remain critical as AI solutions continue to evolve and integrate into core business operations.
Beyond the Headlines
Beyond the immediate financial reporting implications, the evolving accounting standards for AI reflect a broader challenge in valuing and managing intangible assets in the digital economy. The rapid pace of AI innovation means that traditional accounting frameworks, designed for more tangible assets, are being stretched to accommodate new forms of value creation. This raises ethical considerations regarding transparency in financial reporting, particularly for public companies whose operations become heavily dependent on external AI providers. The need for robust governance around AI development, testing, and validation is not just an operational requirement but also an accounting imperative, ensuring that costs are appropriately tracked and justified. The emphasis on defining the 'software project' and assessing 'development uncertainty' highlights the increasing complexity of managing cutting-edge technology within established financial guidelines, potentially influencing how companies structure their AI development teams and projects to align with capitalization criteria.
