Code to Capital: How AI and FASB Are Rewriting the Balance Sheet

MCKINNEY, Texas – November 11, 2026 – For decades, the intersection of software engineering and corporate finance has been defined by a quiet, end-of-year friction. While developers operate in rapid sprints, continuous integration, and microservices, accounting departments have been forced to translate this fluid reality into rigid, linear spreadsheets. It is a process often characterized by retrospective guesswork, where engineers are asked in December to recall what they built in March.

Today, that fragile system is colliding with two massive disruptive forces: the proliferation of artificial intelligence in code generation, and a sweeping regulatory overhaul from the Financial Accounting Standards Board (FASB). In response, a new category of financial technology is emerging to bridge the gap. CodeROI, a compliance and capitalization platform, announced today that it has expanded its software to automatically track and capitalize AI development costs alongside human engineering labor.

The announcement serves as a bellwether for a broader shift in corporate reporting. As companies prepare for FASB’s new internal-use software rule, ASU 2025-06, the days of estimating capitalized labor through memory-based surveys are rapidly coming to a close. Finance leaders are now being pushed to provide deterministic, audit-ready evidence of how their software is actually built.

The Accounting Reality of AI

Artificial intelligence is no longer merely a productivity enhancement; it is a direct contributor to the corporate codebase. As engineering workflows increasingly rely on generative AI tools alongside human coders, standard accounting practices have struggled to accurately classify these development costs.

Historically, general-purpose software subscriptions were treated as overhead and immediately expensed. However, as AI becomes deeply embedded in the software development lifecycle, the accounting treatment is shifting. Under U.S. GAAP (ASC 350-40) and IFRS (IAS 38), the cost of AI tools used to build internal-use software can be capitalized if they are direct costs attributable to a specific, qualifying project.

The challenge lies in attribution. Auditor defensibility hinges on robust evidence that precisely links AI tool usage—such as token consumption or dedicated seat licenses—to specific capitalizable projects. Without reliable activity splitting, a coding session involving both capitalizable feature development and expensable routine maintenance might lead to the entire cost hitting the income statement.

CodeROI’s latest update tackles this directly by measuring AI costs next to employee pay on the exact same projects. By turning everyday engineering work into audit-ready numbers, the platform ensures that balance sheets reflect the modern reality of software construction. This is a critical evolution. As one corporate accounting specialist noted recently, treating AI purely as an operational expense artificially deflates EBITDA and misrepresents the true capital investment companies are making in their proprietary technology.

Retiring the Spreadsheet Guesswork

The urgency behind automated capitalization is being driven by regulatory changes. In September 2025, FASB issued ASU 2025-06, fundamentally modernizing the accounting guidance for internal-use software costs, which had largely remained unchanged since 1998.

The old rules were built for a waterfall development model, requiring companies to track costs through rigid, sequential stages: preliminary project, application development, and post-implementation. Modern agile methodologies, with their iterative sprints and constantly evolving backlogs, made these stages practically impossible to isolate.

FASB’s new standard eliminates these project stages entirely. Instead, a company starts capitalizing once a project is approved, funded, and likely to be finished. While this aligns better with agile development, fewer fixed steps mean more management judgment calls. The new rule also introduces the concept of "significant development uncertainty," mandating that capitalization be deferred if a project faces major technological hurdles.

These changes—which become mandatory for annual periods starting after December 15, 2027—mean that auditors will demand to see the granular facts behind every capitalization decision. Rebuilding the year from memory via spreadsheets will no longer pass muster.

"Software capitalization has always been an estimate dressed up as a number. Finance asks engineering in December what everyone worked on all year, someone builds a spreadsheet, and the auditor is asked to trust it," said Taylor Meadows, CPA, CEO of CodeROI. "The new FASB rule asks companies to show their judgment, not just their totals. CodeROI gives them actuals: what was built, on which project, by which person or AI tool, and what it cost, recorded while it happened."

Bridging the Gap Between Code and Capital

To satisfy auditors under the new regime, the facts finance needs must be extracted directly from engineering systems. However, asking developers to manually log hours against specific financial capitalization criteria is a surefire way to destroy productivity and morale.

The solution lies in deterministic data ingestion. The platform operates by connecting directly to where code lives—such as GitHub or GitLab repositories—and recording the work as it happens. It ties each engineer's pay and each AI tool's cost to the actual projects they contributed to, applying the appropriate U.S. GAAP or IFRS treatment.

Crucially, the McKinney-based firm emphasizes that its math is deterministic. It does not use AI in its calculations or data sourcing, ensuring that the same facts always produce the same answer. There is no generative guessing at what "probably" happened. By tracking activity prospectively from the day a company integrates the software, the system avoids the credibility pitfalls of reconstructing past years from memory.

This automated approach simultaneously links technical activity to other critical financial levers, including Section 174 amortization and R&D tax credits like the U.S. IRS Form 6765 and Canada's SR&ED program. Instead of maintaining separate, often conflicting spreadsheets for tax and accounting, organizations can operate from a single, immutable set of facts.

The New Financial Imperative for Tech Leaders

For Chief Financial Officers and corporate controllers, the transition away from manual capitalization is rapidly becoming a strategic imperative. The financial stakes are simply too high to rely on estimates.

Incorrect capitalization can significantly impact financial statements, altering reported asset values, distorting near-term operating expenses, and heavily influencing EBITDA. Furthermore, as the IAS 38 standard under IFRS sets a notoriously high bar for capitalization—requiring clear distinctions between research and development phases—multinational corporations need granular tracking to defend their asset recognition globally.

By automating the generation of monthly journal entries and amortization schedules, finance teams can achieve a faster, cleaner close. More importantly, they can provide a truer picture of the business, moving development costs from the income statement to the balance sheet and spreading them over the years the software actually generates value.

Ultimately, the evolution of software capitalization is about recognizing the true mechanics of success in the 21st century. As artificial intelligence writes a growing share of the code that powers modern enterprises, the financial infrastructure tracking that value must be equally sophisticated. The era of the year-end spreadsheet is ending, replaced by a continuous, data-driven alignment between the engineers building the future and the balance sheets financing it.