> [!FOUNDER]
> "For three years, paying $200,000 to software engineers meant a $20,000 Year 1 deduction and a phantom tax bill on income the company never received in cash. That era ended for domestic research: the One Big Beautiful Bill Act restored immediate expensing for tax years beginning after December 31, 2024. Foreign research is the exception — it still amortizes over fifteen years." — Enow A. Jovial, Founder & Chief Executive Officer
The Post-AI Gap in Section 174 Compliance
The Tax Cuts and Jobs Act amended Internal Revenue Code Section 174 to mandate 5-year amortization for domestic software development and 15-year for foreign development, and for tax years 2022 through 2024 that was the law. The One Big Beautiful Bill Act, Public Law 119-21, section 70302, ended it for domestic research: new section 174A allows domestic research or experimental expenditures to be deducted when paid or incurred, for tax years beginning after December 31, 2024, while section 174 now amortizes foreign research only, over 15 years. The companion section 174A guide runs the calculation in full. The ledgers below are the 2022–2024 record every founder with unamortized balances still has to unwind.
> [!WARNING]
> The abandonment trap survives — for foreign research only. Under IRC § 174(d), if a foreign research project is abandoned, you are forbidden from taking an immediate write-off; you must continue amortizing over the full 15-year schedule. Domestic research has no such trap under section 174A.
This guide provides explicit mathematical models, employee vs contractor allocation rules, and strategies to offset liabilities using the Section 41 R&D Credit.
!Tax Planning and Accounting Spreadsheets for Founders
Domestic vs. Foreign Software Development Amortization
From 2025, domestic research is deductible when paid or incurred under IRC § 174A(a), with an election to capitalize and amortize over not less than 60 months under IRC § 174A(c). Section 174 now amortizes foreign research only, over 15 years with a mid-year convention under IRC § 174(a). The formulas below are the 2022–2024 schedule, and the foreign schedule that still applies:
$\text{Year 1 Deduction (US Dev, 2022-2024)} = \frac{\text{SRE Expenses}}{5} \times 50\% = 10\% \text{ of Total Expenditure}$
$\text{Year 1 Deduction (Foreign Dev)} = \frac{\text{SRE Expenses}}{15} \times 50\% = 3.33\% \text{ of Total Expenditure}$
Comparative Cash Impact Model ($500k Revenue, $400k Dev Costs, Then vs Now)
Assume a software company with $500,000 of revenue, $400,000 of US engineering labor, and $50,000 of operating expenses:
| Metric | Pre-TCJA (Immediate Deduction) | 2022–2024 Section 174 (5-Yr Amortization) | 2025+ Domestic (Section 174A) |
| :--- | :--- | :--- | :--- |
| Gross Revenue | $500,000 | $500,000 | $500,000 |
| Operating Expenses | -$50,000 | -$50,000 | -$50,000 |
| Allowed Engineering Deduction | -$400,000 (100%) | -$40,000 (10% Year 1) | -$400,000 (100%) |
| Taxable Income (IRS) | $50,000 | $410,000 | $50,000 |
| Federal Tax Liability (21% Corporate Rate) | $10,500 | $86,100 | $10,500 |
| Actual Cash Balance | $50,000 | $50,000 | $50,000 |
| Net Cash After Taxes | +$39,500 | -$36,100 (Insolvent) | +$39,500 |
Under the 2022–2024 rules the company was cash-flow negative after tax despite breaking even operationally; under section 174A it is cash-positive again.
Distinguishing Section 174 SRE vs. Section 162 Maintenance
From 2025, this split no longer decides whether domestic spend is deductible — domestic research is deductible under section 174A either way — but it still decides the foreign schedule and the section 41 credit base. Bookkeepers must categorize developer activities carefully:
#### Capitalized under Section 174:
#### Deductible under Section 162:
Section 41 Payroll Tax Credit Offset Strategy
Early-stage startups with under $5,000,000 in gross receipts and no gross receipts prior to the 5-taxable-year period can elect under IRC § 41(h) to apply up to $500,000 annually of the Research Credit against their FICA payroll tax liability, for immediate cash liquidity even without income-tax liability.
Real-World Case Study: SaaS Startup Amortization Reclassification
This case is a 2022–2024 record: under the law then in force, the $650,000 spend was capitalized and amortized. From 2025, the same domestic spend is deductible when paid or incurred.
In an enterprise scenario, a cloud compliance startup generating $1,900,000 in ARR, $480,000 in EBITDA, and $180,000 in COGS.
The company spent $650,000 on engineering, initially categorizing all developer salaries as immediate operational expenses on their draft profit and loss statement. Under a Section 174 audit review by a specialist CPA:
-
4-Step Section 174 Software Capitalization Protocol
1. Bifurcate Developer Payroll by jurisdiction first, then by activity: Segment Jira sprint issues and Git commit branches into domestic vs. foreign work, and within each, Section 174 novel feature development vs. Section 162 routine bug fixes.
2. Audit Contractor Jurisdictions: Verify physical developer work locations to prevent accidental classification into the 15-year foreign schedule.
3. Execute Section 41 Payroll Offsets: File IRS Form 6765 and Form 8974 to convert eligible engineering hours into quarterly employer FICA tax credits.
4. File IRS Form 3115 When Required: Submit an Application for Change in Accounting Method when formalizing capitalization procedures across historical returns.
The Five-Year Amortization Ledger (2022–2024 Record)
The mid-year convention in the article's own formula fixed the whole schedule for 2022–2024, and still fixes it for foreign spend today: Year 1 takes half a year's amortization, the middle years take a full year each, and a stub year closes the schedule. Applied to the $400,000 US engineering spend from the cash-impact model:
| Year | US deduction (5-year, 2022–2024) | Cumulative | Foreign deduction (15-year, still live) | Cumulative |
| :--- | :--- | :--- | :--- | :--- |
| Year 1 | $40,000 (10%) | $40,000 | $13,333 (3.33%) | $13,333 |
| Year 2 | $80,000 (20%) | $120,000 | $26,667 (6.67%) | $40,000 |
| Year 3 | $80,000 (20%) | $200,000 | $26,667 (6.67%) | $66,667 |
| Year 4 | $80,000 (20%) | $280,000 | $26,667 (6.67%) | $93,334 |
| Year 5 | $80,000 (20%) | $360,000 | $26,667 (6.67%) | $120,001 |
| Year 6 | $40,000 (10%) | $400,000 | $26,667 (6.67%) | $146,668 |
| Years 7-15 | — | $400,000 | $26,667 per year | $386,671 |
| Year 16 | — | $400,000 | $13,333 (3.33%) | $400,004 |
Foreign figures are rounded to the nearest dollar, which is why the foreign column closes at $400,004. The founder's $200,000 example follows the same shape at half scale. The timing gap was the whole story for 2022–2024: after two full years, a US project had deducted only 30% of its cost; a foreign project only 10%.
The phantom profit in the cash-impact model is exactly the undeducted balance. The formula below is derived arithmetic from the model's own numbers, not a new rule:
$ ext{Phantom Taxable Income} = ext{SRE Spend} - ext{Year 1 Deduction} = 90% ext{ of SRE Spend}$
For the $400,000 model: $400,000 - $40,000 = $360,000 of phantom income, and $360,000 x 21% = $75,600 of extra tax — exactly the gap between the model's two tax bills ($86,100 - $10,500), the internal check that the table and the formula agree.
> [!KEY TAKEAWAY]
> The Section 174 cost was a timing cost, and its size was always 90% of the first year's US engineering spend sitting in taxable income. From 2025, that timing cost lives only in the foreign column: a foreign project still deducts 3.33% in Year 1. A founder who can state that number in dollars before the year starts can budget the quarterly estimated payments; a founder who discovers it at filing time is the one who goes cash-negative.
The Cash-Flow Gap Ledger
The same 90% rule scales the model's 2022–2024 result to any engineering budget, using the 21% corporate rate from the model. For tax years beginning after December 31, 2024, the extra Year-1 tax on domestic spend is $0 — the table below is the historical record, and the only column that still runs is the foreign one:
| Annual US engineering spend | Phantom income (90%) | Extra Year-1 tax at 21% |
| :--- | :--- | :--- |
| $100,000 | $90,000 | $18,900 |
| $400,000 (model) | $360,000 | $75,600 |
| $1,000,000 | $900,000 | $189,000 |
The swing in the model's net cash — from +$39,500 under immediate deduction to -$36,100 under amortization — is $75,600, the same extra-tax figure: the entire cash-flow reversal is the phantom-profit tax. A company with $1,000,000 of US engineering spend and the model's $50,000 cash cushion would face a $189,000 Year-1 tax bill on $900,000 of income it never received in cash. That is the number to carry into every hiring plan: each additional $100,000 of US engineering hires adds $18,900 of Year-1 tax on top of the salary itself.
The Section 41 payroll offset in this guide is a separate cash channel, not a reduction of that income-tax bill: the election applies up to $500,000 of Research Credit per year against FICA payroll tax. For the model company, the $500,000 annual cap is more than five times its $86,100 Year-1 income-tax bill — but it offsets payroll tax, so it relieves cash only to the extent the company actually owes payroll tax. Eligibility is narrow, which is why the offset belongs in the plan before the hiring, not after the tax bill arrives.
The Reclassification Ledger
The case study's numbers show why the Section 174 / Section 162 split is the highest-leverage bookkeeping decision in the guide:
| Treatment | Year-1 deduction |
| :--- | :--- |
| All $650,000 as Section 174 SRE (no split) | $65,000 (10%) |
| $420,000 SRE + $230,000 reclassified as Section 162 | $42,000 + $230,000 = $272,000 |
| Difference from the split | +$207,000 of Year-1 deduction |
The $420,000 SRE portion then follows its own schedule: $42,000 in Year 1, $84,000 in each of Years 2-5, and a final $42,000 in Year 6, totalling $420,000. From 2025, the split no longer decides domestic deductibility — domestic research is deductible under section 174A either way — but it still decides the foreign schedule and the section 41 credit base. The case study reports $39,480 of unexpected paper tax eliminated by the reclassification — that figure is the study's own outcome, stated here as the guide records it. The mechanism is the $207,000 of deduction pulled forward into Year 1 by the $230,000 reclassification, defensible only because the engineering team's contemporaneous Git logs separated bug maintenance from novel feature work at the time the work was done.
> [!TIP]
> Tag every Jira ticket and Git branch as Section 174 or Section 162 at creation, not at year end. The reclassification in the case study survived because the evidence was contemporaneous; a year-end reconstruction from memory is the first thing an examiner discounts.
Decision Table
| Operator shape | The Section 174 move that matters |
| :--- | :--- |
| Bootstrapped solo founder, $100,000 of US dev | Budget $0 of Year-1 phantom tax on domestic spend from 2025; the 90% rule now applies only to foreign spend. |
| Seed-stage SaaS, $400,000 of US dev | Run the five-year ledger above only for foreign spend; for domestic spend the $75,600 of extra Year-1 tax is gone from 2025. |
| Offshore-heavy team, mixed US/foreign dev | Split the ledger by jurisdiction: the foreign 15-year column deducts only $13,333 of each $400,000 in Year 1. |
| Profitable scale-up with payroll tax | File Form 6765 and Form 8974 early; the $500,000 annual Section 41 payroll offset is a cash channel, not an income-tax cut. |
| Pre-revenue R&D lab, project may be abandoned | Price the Section 174(d) trap before starting: an abandoned foreign project keeps amortizing, so a dead foreign build still produces phantom income for years. A domestic build has no trap under section 174A. |