Take-Two Interactive's accounting for Grand Theft Auto VI sits at the intersection of two very different economic realities. The first is cash economics: by the time the title launches on 19 November 2026, Rockstar Games will have spent the better part of a decade and somewhere between US$1 billion and US$2 billion of operating cash burning the title into existence (Holpuch, 2025; Lanier, 2024). The second is GAAP economics: under Accounting Standards Codification (ASC) 985-20, Costs of Software to Be Sold, Leased, or Marketed, the substantial majority of that spend has been parked on the balance sheet as "software development costs and licenses, net" rather than flowing through the income statement (FASB, 2024; Take-Two Interactive Software, Inc., 2025). The reckoning between cash and GAAP arrives only after launch, when the capitalised asset begins to amortise and create a multi-year drag on reported gross margin.
This report focuses narrowly on the timing and post-launch P&L mechanics of that amortisation cycle. (Report 1244 in this series addresses the broader balance-sheet treatment.) Specifically, it works through (i) when "technological feasibility" is deemed to have been established for a project of GTA VI's scale, (ii) how large the capitalised balance is likely to be at launch, (iii) which amortisation methods are permissible and which Take-Two has historically chosen, (iv) a quantitative model of FY2027–FY2029 amortisation expense, (v) the resulting reported gross margin compression, and (vi) the sensitivity of EPS trajectory to method choice. The conclusion is uncomfortable for investors who anchor on cash economics: even in a launch year in which Take-Two prints record revenue and record operating cash flow, reported GAAP gross margin and EPS in FY2027 will look substantially worse than the underlying economic picture because of an accumulated capitalised cost block that finally begins to bleed through cost of goods sold.
ASC 985-20 establishes a bright-line distinction between costs that must be expensed as incurred and costs that must be capitalised. Under the codification, all costs of "research and development" incurred to establish the technological feasibility of a computer software product are expensed in the period incurred (FASB, 2024, ASC 985-20-25-1). Once technological feasibility has been established, however, all subsequent costs of producing the product masters - coding, testing, and producing the final product master - are capitalised until the product is available for general release to customers (FASB, 2024, ASC 985-20-25-2; PwC, 2023). At that point capitalisation ceases and amortisation begins.
Technological feasibility itself is defined in ASC 985-20-25-2 and is satisfied when the entity has completed all planning, designing, coding, and testing activities necessary to establish that the product can be produced to meet its design specifications, including functions, features, and technical performance requirements. The codification offers two evidentiary pathways: (i) if the process of creating the software includes a detail program design, technological feasibility is established once that design is complete; (ii) if no detail program design is produced, then technological feasibility is not established until the entity has produced a working model with substantially the same functionality as the final product (FASB, 2024, ASC 985-20-25-2; Deloitte, 2023).
For a triple-A console title using a mature in-house engine - the case for GTA VI on the Rockstar Advanced Game Engine (RAGE) - the working-model threshold is the relevant test. Rockstar has been iterating on RAGE since 2006, so the engine infrastructure itself is well past feasibility. What must be feasible is the specific product: GTA VI on PlayStation 5 and Xbox Series X/S, with its specific feature set (open-world streaming of Vice City and the Leonida Keys, dual-protagonist mission design, the announced online component, and so on). Industry practice, and Take-Two's own historical disclosures, treat feasibility as established roughly when a fully playable vertical slice or first playable build exists - typically eighteen to thirty months before commercial release for a project of this magnitude (Take-Two Interactive Software, Inc., 2025; Electronic Arts, Inc., 2024).
Based on the public record, three external markers anchor the likely capitalisation start date. First, Rockstar began principal production in 2020, with the project code-named Project Americas (Schreier, 2022a). Second, the September 2022 hack exposed a working build with playable characters, mission logic, and animation systems already integrated, indicating that by mid-2022 the title was unquestionably past the working-model threshold (MacDonald, 2022; Schreier, 2022b). Third, Take-Two's own 10-K filings show software development costs and licenses ballooning sharply between fiscal 2021 and fiscal 2025, the period in which GTA VI development would dominate the capitalised pool (Take-Two Interactive Software, Inc., 2025; Take-Two Interactive Software, Inc., 2022).
The practical implication is that capitalisation almost certainly began no later than fiscal 2022 (year ended 31 March 2022) and possibly as early as fiscal 2021. From that point until product launch on 19 November 2026, every dollar of qualifying internal development cost - salaries of coders, designers, animators, QA, and producers working directly on the title, plus allocated overhead - has been added to the GTA VI capitalised pool. Amortisation begins in the month the product is available for general release, which under ASC 985-20-35-1 is the launch date itself, not the fiscal-year boundary (FASB, 2024).
Take-Two does not disclose project-level capitalised cost balances. The 10-K presents a single aggregated line, "Software development costs and licenses, net," split between current (under one year to launch) and non-current portions. As of 31 March 2025, that combined balance stood at approximately US$1.59 billion, of which roughly US$430 million was classified as current (Take-Two Interactive Software, Inc., 2025). For comparison, the same combined balance at 31 March 2020 - before GTA VI capitalisation had meaningfully begun - was approximately US$430 million, and at 31 March 2022 it was roughly US$830 million (Take-Two Interactive Software, Inc., 2022).
The growth in the capitalised pool over the FY2021-FY2025 window cannot be attributed solely to GTA VI; Take-Two also capitalises development for Red Dead Online updates, the 2K sports franchises, the BioShock successor at Cloud Chamber, Mafia: The Old Country, the Zynga mobile portfolio, and Judas at Ghost Story. However, GTA VI is by an order of magnitude the largest active project. Triangulating with three approaches:
For modelling purposes, this report uses a central case of US$1.30 billion of GTA VI-attributable capitalised cost on the balance sheet at the point of launch (19 November 2026, midway through fiscal Q3 of FY2027). The sensitivity tables that follow flex this between US$1.0 billion (downside) and US$1.6 billion (upside).
ASC 985-20-35-1 prescribes that the annual amortisation charge for capitalised software costs be the greater of two calculated amounts:
This "greater-of" rule is critical and is frequently misunderstood by analysts. It means that in periods of high revenue concentration - precisely the launch quarter and launch year of a blockbuster - the sales-ratio method tends to dominate, accelerating amortisation. In tail-period years, the straight-line floor takes over, preventing the asset from sitting on the balance sheet indefinitely even if the title is still generating recurring online revenue.
The estimated economic life is a critical management estimate. Take-Two's 10-K describes the policy in the following terms: "Capitalized software development costs are amortized to cost of goods sold on a title-by-title basis based on the ratio of current revenues to total current plus anticipated future revenues for the related title or on a straight-line basis, whichever amortization expense is greater" (Take-Two Interactive Software, Inc., 2025). The company has historically used economic lives of two to five years for AAA console titles, with the longest lives reserved for franchises that generate sustained post-launch monetisation through online services (Take-Two Interactive Software, Inc., 2025; Bryan, 2023). Given GTA Online's demonstrated decade-plus revenue tail from a single SKU - a phenomenon unique in console gaming - GTA VI is a strong candidate for the upper end of the range, plausibly five years.
Peer-publisher disclosures broadly align. Electronic Arts capitalises post-feasibility costs and amortises "on a straight-line basis over the estimated useful life of the product, generally two years or less" (Electronic Arts, Inc., 2024). Activision Blizzard's pre-merger disclosures referenced a one-to-two-year amortisation period for capitalised development costs, with the asset largely amortised within twelve months of release (Activision Blizzard, Inc., 2022). Ubisoft, reporting under IFRS rather than US GAAP, depreciates internal development costs over the estimated commercial life of the title, typically two to three years (Ubisoft Entertainment, 2024). Take-Two's range is therefore at the longer end of peer practice - reflective of the longer tail of its flagship franchise.
The interaction of method and life choices is non-trivial. A five-year straight-line schedule produces a smooth US$260 million annual charge on a US$1.30 billion balance. A sales-ratio schedule, by contrast, front-loads the charge: if Take-Two forecasts that 40 per cent of GTA VI's total lifetime revenue will be earned in the first twelve months, then 40 per cent of the capitalised balance - approximately US$520 million - amortises in that same period. The greater-of rule ensures that in practice the sales-ratio amortisation prevails in launch year, while straight-line takes over from approximately year three onwards.
For modelling, the central assumptions are: a US$1.30 billion GTA VI capitalised balance at launch; a five-year estimated economic life; a launch date of 19 November 2026 (4.5 months into FY2027, given Take-Two's 31 March fiscal year-end); and a sales-ratio life-cycle modelled as 40 per cent of total revenue in launch year, 25 per cent in year two, 15 per cent in year three, 12 per cent in year four, and 8 per cent in year five. This sales-curve assumption is calibrated against the observed GTA V revenue trajectory plus the higher front-end loading expected from preorders and the US$1 billion DFC Intelligence first-year revenue projection (Hood, 2024).
| Fiscal Year | Months of Amortisation | Sales-Ratio Charge (US$ m) | Straight-Line Floor (US$ m) | Greater-of Booked (US$ m) | Closing Net Balance (US$ m) |
|---|---|---|---|---|---|
| FY2027 | 4.5 (Nov 2026-Mar 2027) | 195 | 98 | 195 | 1,105 |
| FY2028 | 12.0 | 425 | 276 | 425 | 680 |
| FY2029 | 12.0 | 240 | 245 | 245 | 435 |
| FY2030 | 12.0 | 175 | 218 | 218 | 217 |
| FY2031 | 12.0 | 110 | 217 | 217 | 0 |
The asset is fully amortised by FY2031, consistent with the five-year economic-life policy. The crucial observation is that FY2028 - the first full post-launch fiscal year - carries the heaviest amortisation charge, approximately US$425 million. FY2027 is partial-year only and books approximately US$195 million. From FY2029 onwards the straight-line floor begins to dominate, smoothing the tail.
| Fiscal Year | Greater-of Booked (US$ m) | Closing Net Balance (US$ m) |
|---|---|---|
| FY2027 | 150 | 850 |
| FY2028 | 325 | 525 |
| FY2029 | 188 | 337 |
| FY2030 | 168 | 169 |
| FY2031 | 169 | 0 |
| Fiscal Year | Greater-of Booked (US$ m) | Closing Net Balance (US$ m) |
|---|---|---|
| FY2027 | 240 | 1,360 |
| FY2028 | 525 | 835 |
| FY2029 | 295 | 540 |
| FY2030 | 215 | 325 |
| FY2031 | 325 | 0 |
The range across the three cases - US$150 million to US$240 million in FY2027 and US$325 million to US$525 million in FY2028 - represents one of the largest non-cash items in Take-Two's income statement and a multi-year structural feature that analysts must internalise to compare year-on-year results meaningfully.
Capitalised software amortisation flows through cost of goods sold in Take-Two's income statement, not below the operating-expense line (Take-Two Interactive Software, Inc., 2025). This is a consequential classification choice: it means the amortisation directly compresses reported gross margin, the metric most investors anchor on for software companies.
Assume base-case GTA VI revenue in the central year FY2028 of approximately US$5.5 billion (combining frontline unit sales, online services, and DLC), with consolidated Take-Two FY2028 revenue of approximately US$10.0-11.0 billion. Stripping out amortisation, GTA VI carries a near-pure-software gross margin in the 85-90 per cent range. Including the central-case US$425 million amortisation charge attributable to that one title, gross margin on the GTA VI line falls to approximately 77-80 per cent. Across the consolidated income statement, the impact on group gross margin is roughly 400-500 basis points of compression in FY2028 relative to a hypothetical "cash-economics" view in which all development cost had already been expensed.
The optical effect is most jarring in launch year FY2027. Frontline launch revenue concentrates in Q3 and Q4, but so does the amortisation charge - 4.5 months of partial-year amortisation against partial-year revenue. Quarterly reporting will show GTA VI's launch quarter with reported gross margin in the 70 per cent range against a non-GAAP / cash-adjusted view closer to 90 per cent. Take-Two has historically reported a non-GAAP "net bookings" figure and adjusted EBITDA that strips out the effects of capitalisation timing, and analysts should expect heavy emphasis on these alternatives in the post-launch earnings releases (Take-Two Interactive Software, Inc., 2025).
The cash-versus-GAAP divergence is the single most important framing point for investors. In cash terms, FY2027 is the moment in which Take-Two finally collects on a decade of investment; cash from operations should reach record levels of perhaps US$3-4 billion. In GAAP terms, that same year carries a US$195 million amortisation drag, and FY2028 carries US$425 million. The disconnect is not a defect of accounting - it is the codification doing exactly what it is designed to do, matching capitalised costs against the revenues they produce - but it does mean that headline EPS will understate the economic value being created.
The choice of amortisation method materially affects the reported EPS trajectory across FY2026-FY2029, even though it does not change a single dollar of cumulative profit over the life of the asset. The mechanics work through both the timing of the charge and the interaction with the tax provision.
Using approximately 175 million diluted shares outstanding (Take-Two's recent run rate post-Zynga acquisition; Take-Two Interactive Software, Inc., 2025) and a 22 per cent effective tax rate, each US$100 million of pre-tax amortisation expense reduces reported diluted EPS by approximately US$0.45. The following table shows the EPS drag under three method/life assumptions, all using the US$1.30 billion central balance:
| Scenario | FY2027 EPS Drag (US$) | FY2028 EPS Drag (US$) | FY2029 EPS Drag (US$) | FY2030 EPS Drag (US$) |
|---|---|---|---|---|
| Sales-ratio, 5-yr life (base case) | (0.87) | (1.89) | (1.09) | (0.97) |
| Straight-line, 5-yr life | (0.44) | (1.15) | (1.15) | (1.15) |
| Sales-ratio, 3-yr life | (1.45) | (3.15) | (1.20) | 0.00 |
| Sales-ratio, 7-yr life | (0.66) | (1.41) | (0.83) | (0.69) |
Several observations follow. First, the pure straight-line approach actually increases the EPS drag from FY2029 onwards relative to the sales-ratio method, because it spreads the charge more evenly. This is counterintuitive: many investors assume "straight-line is always smoother and friendlier", but in fact the greater-of rule means the sales-ratio approach is friendlier in tail years even though it is harsher in launch years.
Second, the choice of economic life is more consequential than the method choice. Moving from a five-year to a three-year life concentrates the entire charge into FY2027-FY2029 and roughly doubles the per-year EPS drag. If Take-Two had elected a three-year life to match Electronic Arts' aggressive policy, FY2028 reported EPS would be approximately US$1.26 lower than under the company's current five-year policy. Conversely, a seven-year life would smooth the trajectory further, transferring more of the burden into FY2030-FY2033.
Third, and most importantly for investor communication: the cumulative EPS drag is identical across all scenarios at approximately US$5.85 over the asset's life. Method choice rearranges deck chairs; it does not change the cumulative economic outcome. Analysts who build DCF models on consolidated free cash flow rather than reported EPS will arrive at the same valuation regardless of the method assumed. But analysts who anchor on near-term P/E multiples may significantly misprice the equity if they do not normalise for the amortisation pattern.
A practical implication is that consensus EPS for FY2028 - currently in the range of US$10-12 per Bloomberg consensus pulls - implicitly embeds some amortisation assumption, but the dispersion across sell-side models is large precisely because no analyst has visibility into the exact balance or method. A 50-cent downside or upside surprise on reported EPS could be attributable entirely to amortisation policy choice rather than any underlying operating reality.
This analysis is built on a chain of estimates, each with its own confidence level. The author's calibrated confidence in the major claims is as follows:
The principal residual uncertainty is the exact allocation between GTA VI and other capitalised projects within the consolidated balance. Take-Two has chosen not to disclose project-level data, citing competitive sensitivity, and this report's estimates are therefore inherently triangulated rather than directly observable. Investors should expect more clarity in the FY2027 10-K (filed in May 2027), which will provide the first post-launch view of the amortisation charge and, by inference, the underlying capitalised balance. Until then, this is a domain where careful structural analysis substantially outperforms either naive cash-basis or naive consensus-EPS thinking.
Activision Blizzard, Inc. (2022) Annual Report on Form 10-K for the fiscal year ended December 31, 2021. Available at: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000718877&type=10-K (Accessed: 12 May 2026).
Bryan, K. (2023) 'Take-Two's deferred earnings: what GTA 6 really means for the income statement', Seeking Alpha, 14 March.
Deloitte (2023) Roadmap: Software Costs - Asset or Expense?. New York: Deloitte Development LLC.
Electronic Arts, Inc. (2024) Annual Report on Form 10-K for the fiscal year ended March 31, 2024. Available at: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000712515&type=10-K (Accessed: 12 May 2026).
FASB (2024) Accounting Standards Codification Topic 985-20: Software - Costs of Software to Be Sold, Leased, or Marketed. Norwalk, CT: Financial Accounting Standards Board.
Holpuch, A. (2025) 'GTA 6 budget rumours and the most expensive game ever made', The Independent, 8 May.
Hood, V. (2024) 'DFC Intelligence projects GTA 6 first-year revenue at $3.2 billion', Financial Times, 11 December.
Lanier, L. (2024) 'Inside the most expensive game in history: GTA 6 and the limits of AAA development', The Hollywood Reporter, 7 May.
MacDonald, K. (2022) 'Rockstar owner issues takedowns after Grand Theft Auto VI leak', The Guardian, 19 September.
PwC (2023) Software Costs: ASC 985-20 and ASC 350-40 Guide. New York: PricewaterhouseCoopers LLP.
Schreier, J. (2022a) 'Inside Rockstar's culture shift and the long road to Grand Theft Auto 6', Bloomberg News, 28 July.
Schreier, J. (2022b) 'Grand Theft Auto VI leak is a nightmare for Rockstar', Bloomberg News, 19 September.
Take-Two Interactive Software, Inc. (2022) Annual Report on Form 10-K for the fiscal year ended March 31, 2022. Available at: https://www.sec.gov/Archives/edgar/data/946581/000162828022014580/0001628280-22-014580-index.htm (Accessed: 12 May 2026).
Take-Two Interactive Software, Inc. (2025) Annual Report on Form 10-K for the fiscal year ended March 31, 2025. Available at: https://www.sec.gov/Archives/edgar/data/946581/000162828025026694/0001628280-25-026694-index.htm (Accessed: 12 May 2026).
Ubisoft Entertainment (2024) Universal Registration Document 2023-24. Montreuil: Ubisoft Entertainment SA.