Offshore Engineering and Transfer Pricing

You incorporated in Delaware. Your engineers live in Lahore, Bangalore, Cairo, Dubai, Ho Chi Minh, Manila, Mexico City, São Paulo, or Buenos Aires. The cap table belongs to a US C-corp. Payroll, in practice, does not. The structure that connects those two facts is the single most over-and-under-engineered piece of your company. This chapter is what real founders and real tax counsel actually use in 2025–26 — with the numbers, the code sections, and the failure modes.

The four structures

There are four — and only four — operationally common patterns.

Structure A — 1099 Independent Contractor. Delaware Inc pays the engineer directly. The engineer signs a W-8BEN-E (entity) or W-8BEN (individual). No 1099-NEC is required for non-US persons performing 100% of the work outside the US. Works for sub-5 truly-independent engineers. Breaks on (a) misclassification under local labor law (Pakistan, India, Egypt, Philippines all have de-facto employee tests — exclusivity, control, integration), (b) outbound dependent-agent permanent establishment for the US Inc in the engineer's country if the contractor concludes contracts on the Inc's behalf, and (c) zero clean equity path. Cost overhead ~0%; tax exposure medium-high; equity none; fundraising-friendly at seed, not at Series A.

Structure B — EOR / PEO. Deel, Remote, Rippling Global, Multiplier, OysterHR. Platform fees $400–699/month per employee. Real all-in cost runs 30–60% higher once you add employer payroll taxes, one-month deposits, 0.6–2% FX markup, and $50–150 country surcharges. Works for seed through early Series A. Limits: equity is workable but messy (engineer is technically employed by the EOR, not the Inc — Carta and Pulley both support a "non-employee service provider" template, but documentation gets ugly at exit), no transfer-pricing benefit, and at 20 engineers in one country at $599/month you're burning $143K/year in pure platform fees — a senior engineer, gone. Cost overhead 30–60% above platform fee; tax exposure low; equity workable-but-messy; fundraising-friendly through Series A.

Structure C — Wholly-Owned Offshore Subsidiary, cost-plus. Delaware Inc owns 100% of a local entity. Local sub employs the engineers. An intercompany services agreement (ISA) governs the relationship: sub renders software development services to parent at fully-loaded cost plus markup. This is the structure used by virtually every venture-backed company past Series A with 15+ engineers offshore. Cost overhead $15–40K/year ongoing (local audit, statutory filings, TP refresh) plus $5–15K setup. Tax exposure low if executed properly; medium-high if you skip the TP study. Equity is clean; IP clarity strong; highly fundraising-friendly.

Structure D — Founder's Pre-existing Local Entity as Vendor. Common for founders who ran a freelance/agency before raising US capital. The honest read: this is a related-party transaction in everything but legal form, and tax authorities treat it as such. If the founder owns >50% of both, you're inside §482 (US) and Section 92 (India) — full TP exposure with none of the diligence cleanliness. Investor fix is invariably: roll the offshore entity in (asset purchase, IP assignment, employee transition) before close. Budget $30–80K to clean it up. Use only as a 6–12 month transition while you stand up Structure C.

	Cost overhead	Tax exposure	Equity path	Fundraising friendliness
A — 1099 contractor	~0%	Medium-high	None clean	Seed only
B — EOR / PEO	30–60% over platform fee	Low	Workable, messy	Through Series A
C — Wholly-owned subsidiary	$15–40K/yr	Low if done right	Clean	Series A+ standard
D — Founder's pre-existing entity	Variable; cleanup $30–80K	High pre-cleanup	Blocked	Not fundraising-friendly

Transfer pricing 101

Under IRC §482 and the OECD Transfer Pricing Guidelines (2022 edition, Chapter VII on intra-group services), related-party transactions must be priced at "arm's length" — what unrelated parties would charge each other for the same service.

For a US Inc with a captive offshore engineering sub, the controlled transaction is "engineering services rendered by Sub to Parent." The defensible method is almost always Cost Plus or TNMM with operating cost as the profit-level indicator. Both produce the same answer when properly applied: cost-plus a markup in the 5%–10% range for routine, captive software development.

The "routine vs non-routine" distinction is everything. Routine services = cost-plus 5–10%. Non-routine, IP-creating services = profit split or residual profit methods, with materially higher economic returns. Your offshore sub must be characterized as routine if you want IP and the bulk of profit to live in the Delaware Inc.

Why IP must live in the Delaware Inc

The party that owns IP earns the residual profit. The textbook lesson is the Apple/Ireland State Aid case (CJEU C-465/20, September 2024, €14.1B recovery): Apple Sales International and Apple Operations Europe's Irish branches were allocated nearly all of Apple's non-US IP profit, but the CJEU found those branches lacked the substance — no employees doing IP development, only directors signing minutes — to justify the allocation.

The same lesson runs through the Microsoft $29B IRS dispute announced October 2023, Coca-Cola v. Commissioner (Tax Court 2020, $3.4B), and Medtronic v. Commissioner (8th Cir. 2018, remanded). For startups: IP is created and owned by the Delaware Inc; the offshore sub provides routine development services on a cost-plus basis under a written ISA with present-assignment language.

The TP study

Triggers: annual intercompany flow >~$1M, Series A close, or M&A diligence. 2025–26 pricing:

• Big 4 (PwC, EY, KPMG, Deloitte): $20–50K early-stage; $75K+ master/local file at Series C+.
• National (BDO, RSM, Grant Thornton, Plante Moran): $15–30K.
• Boutique TP (Valentiam, Quantera, Aibidia, NERA): $10–25K — best value at seed/Series A.
• Startup-CPA shops (Kruze, Pilot, Burkland): publish guidance and partner with TP firms; budget the partner separately.

Sub-$1M intercompany flow: a "best-efforts" memo using SEC EDGAR comparables under NAICS 541512, with a conservative 7–8% markup, is what most early-stage tax counsel will sign off on. As Glencoyne puts it: "a well-reasoned report showing your work is far better than a perfect but undocumented policy." Refresh every 3 years or on material business change.

The intercompany services agreement

A defensible ISA contains:

Supporting documentation: functional analysis (who does what, where, with what assets, bearing what risk), benchmarking study, annual one-page TP-refresh memo. Cooley GO's "Intercompany Agreements" guidance and Wilson Sonsini's emerging-companies practice both publish reference architectures.

Equity for offshore engineers

A current 409A valuation under IRC §409A and Treas. Reg. §1.409A-1(b)(5)(iv)(B) is required for any options granted to anyone — US or not. Annual, post-priced-round, or post-material-event. Cost: $1,500–5,000 from Carta, Pulley, Eqvista, Sofer, Andersen at early stage. An out-of-date or low-balled 409A creates immediate income inclusion, 20% federal penalty, plus interest under §409A(a)(1)(B) — for the optionee.

Instrument choice:

• NSOs. Default for non-US engineers (ISOs are limited to W-2 employees of the issuing company; offshore-sub employees of a US parent generally don't qualify). Tax: ordinary income at exercise on the spread, capital gains on subsequent sale. India treats NSO exercise as "perquisite" salary income; Pakistan as salary; Egypt as employment income. The local employer (offshore sub) typically owes withholding at exercise — so the sub needs to know when each engineer exercises.
• RSUs. Cleaner conceptually but India, Egypt, and Pakistan all tax at vest as salary, creating immediate cash withholding on illiquid stock. Bad for engineers without secondary liquidity.
• SARs. India: cash-settled = salary at exercise; equity-settled = salary at exercise plus capital gains at sale. FEMA implications for cross-border SARs in India are real.
• Phantom equity / cash bonus tied to value. No 409A issues if structured as a true short-term deferral; clean treatment as bonus in most countries. Downside: no real ownership; hard to retain talent.

Pragmatic 2025–26 default: NSOs of the US Inc to offshore-sub employees, with the sub committing in the ISA to handle local withholding at exercise, and double-trigger acceleration on change-of-control so engineers without near-term liquidity don't face dry-tax events. Carta's "International Equity Plans" module and Pulley's "Global Equity" both support this; Ledgy is the European-strong alternative.

Country safe harbours

The country-specific safe-harbour thresholds change the math:

• Pakistan: IT/ITES Final Tax Regime (Section 154A, ITO 2001) — 0.25% withholding on export proceeds for PSEB-registered companies, locked in until June 30, 2036 by Finance Act 2025.
• India: Safe-harbour 18% margin on operating cost for software dev services (CBDT Notification 21/2025, threshold raised to INR 3 billion / ~$36M intercompany volume). Proposed 2026 Rules consolidate to 15.5% safe-harbour margin.
• UAE: UAE Corporate Tax 9% above AED 375K. Free zones with Qualifying Free Zone Person (QFZP) status get 0% on Qualifying Income — but only if non-qualifying revenue stays under 5% of total revenue or AED 5M, whichever is lower, and audited financials are filed.
• Mexico: Mandatory maquiladora safe-harbour — greater of 6.5% of cost or 6.9% of asset value (Article 182 LISR). APAs eliminated for FY 2025.
• Vietnam: Decree 320/2025/ND-CP — 10% CIT for 15 years for qualifying tech/software (4-year exemption + 9-year 50% reduction). Software exports VAT 0%.
• Philippines: CREATE MORE Act (RA 12066, signed Nov 11, 2024) — RBE income tax from 25% to 20%, 50% WFH for PEZA, BPO local business tax exempt for 7–17 years.
• Egypt: ITIDA Export-IT Program — cash rebate 35% micro / 15% small / 10% medium on value-added export proceeds. Eligibility requires ≥50% Egyptian ownership.

Permanent establishment risk

Inbound (does the offshore sub create a US tax footprint for itself?): generally no, because the sub renders services to the US, not in the US. Risk arises only if (a) sub employees travel frequently to the US and conclude contracts, (b) the offshore CEO/CTO is also a US Inc officer working from the US, or (c) commission-based US sales activity exists. None typical for an engineering sub.

Outbound (does the US Inc create a tax footprint in the engineer's country?): yes — and this is the more common foot-gun under Structure A. A 1099 contractor in Pakistan habitually concluding contracts on the Inc's behalf, or working exclusively for the Inc and integrated like an employee, creates a dependent-agent PE under Pakistan FBR + OECD model treaty Article 5(5). Same theory under India's Section 9 + Article 5 of the relevant DTAA, Philippines NIRC Section 23, and so on. Mitigation: route through Structure B/C; ensure no contract authority sits with offshore staff.

Founder-on-both-cap-tables wrinkle. When the same founder owns and controls both entities, every transaction is "controlled" — the burden shifts to the founder to prove arm's length. Resolve by formally collapsing into parent-subsidiary with cost-plus before the priced round.

Audit triggers and penalties

Things that draw an audit:

• Intercompany flow >$10M annually with no contemporaneous TP documentation.
• Markup at extremes (0% or >25%) without justification.
• Persistent losses at the offshore sub (TPOs say: a routine service provider should not lose money).
• Persistent losses at the parent with profits piling at the sub.
• Material change in intercompany flow without contract amendment.
• Missing 5471/5472 filings. Penalties: $25,000 per missed form per year minimum.
• §6662(e) net-adjustment penalty: 20% of underpayment if a TP adjustment exceeds the lesser of $5M or 10% of gross receipts; 40% if it exceeds $20M or 20% of gross receipts.

The realistic year when this becomes a real problem is diligence year — Series C, late B, or M&A — not audit year. Acquirer counsel prices the gap into the deal as indemnity hold-back or price reduction. Deals do unwind over offshore-IP and offshore-employee mis-structuring.

State exposure: California ($800 minimum franchise tax + factor presence at $757,070 of CA-source sales or $75,707 of CA payroll for 2025); New York economic nexus; market-based sourcing pulling income to the customer's state regardless of where the engineer sits.

Recommended stack by stage

	Engineers	Structure	TP work	Equity
Pre-seed / seed	<5	EOR (Deel/Rippling/Multiplier)	None	NSOs, annual 409A
Seed / early Series A	5–15	EOR through close, then convert to Sub within 90 days post-close	Best-efforts memo, $5–15K	NSOs, annual 409A
Series A → B	15+ or >$1M flow	Wholly-owned Sub, ISA at cost-plus 7–8%	Formal study, $15–30K	NSOs, double-trigger acceleration
Series B+	25+, $10M ARR	Sub + master/local file in required jurisdictions	Big 4 study, annual refresh	Full international equity-plan

The two-sentence summary: EOR through Series A, wholly-owned offshore subsidiary thereafter, with a written intercompany services agreement at cost-plus 7–8% (or country safe-harbour if higher), IP owned by the Delaware Inc via present-assignment language in both the ISA and every individual employment contract, and a contemporaneous transfer-pricing memo refreshed every three years. Everything else — the EOR alphabet soup, the country-specific incentive program, the equity-plan exotica — is implementation detail around that one structural fact.

Notes & sources

1. OECD Transfer Pricing Guidelines 2022
2. Glencoyne — Benchmarking Transfer Pricing After Series A
3. BDO — IRS Clarifies SCM Exception to BEAT
4. Apple EU Tax Dispute (CJEU C-465/20)
5. Legal Dive — Microsoft $29B IRS TP Dispute
6. Cooley GO — Intercompany Agreements
7. KPMG — India Draft TP Safe Harbor Rules 2026
8. TPC Group — Mexico Maquiladora 2026
9. UAE FTA — Free Zone Persons Corporate Tax Guide
10. Vietnam Briefing — Decree 320/2025
11. Kruze — Transfer Pricing for Startups
12. IRS — IRC §6662(e) Penalty