How Tech Companies Are Turning to Employer of Record Services

In 2025, tech shed more than 150,000 jobs across hundreds of companies, with TechCrunch's running 2025 layoff list and Crunchbase's tracker recording cuts at almost every large public tech firm. In the same year, top-tier AI companies pushed senior engineering compensation toward and past $700K-$1M+ per year at the upper bands, with Anthropic engineering levels running in the $560K-$760K range and OpenAI's L5-L6 reaching well into seven figures. Both facts are true. They describe the same labor market.

What used to be called the tech talent pipeline is now two pipelines: an AI-adjacent bidding war at the top, and a broader rebalancing underneath it. In both, the answer has turned out to be the same. Hire further from the HQ, faster than an entity can be opened, and pay properly in the worker's home country. That is the quiet operating pattern behind the growth of Employer of Record (EOR) services in tech, and it is no longer a niche model. It is how a large fraction of the industry now opens a new country.

This article is the tech-industry thesis: why EOR took hold in tech first, what the current wave actually looks like, and where it is leading by sub-vertical. The companion pieces linked throughout go deeper into the concerns specific to SaaS expansion, AI, fintech, and open-source companies.

Why tech adopted EOR before anyone else

Three structural features of the industry made it an early adopter.

Remote was already the operating model. By 2022, developer workflows were already asynchronous, documented, and tool-mediated. Moving employment across borders did not require reinventing the org; it required reinventing the contract. That is narrower work, and EOR providers fit into it cleanly. Stack Overflow's 2024 developer survey found that the majority of professional developers work fully remote or hybrid, and GitHub's developer-population data shows that the largest growth in active developers is now outside the United States.

Venture economics rewards time-to-hire over locality. Every quarter a Series-A company spends setting up a Dublin entity is a quarter its revenue plan is waiting. EOR compresses time-to-first-hire in a new country from the three to nine months an entity takes, plus banking and payroll provisioning, to roughly the length of a background check. For capital-efficient startups that is the only version of the trade that makes sense.

The talent pool was already distributed. Senior engineering, design, and ML talent sits in Tel Aviv, Warsaw, London, Bengaluru, São Paulo, Lagos, Seoul. Hiring at the frontier has always been international; the question was only how to do it legally. Tech's comfort with distributed work pre-committed it to a model where the employment wrapper is portable.

Healthcare, industrials, and regulated finance are now adopting EOR. They are following a playbook tech wrote. Our healthcare thesis piece traces the same pattern in a different industry.

What changed in 2025

Four shifts, all reinforcing one another.

The shortage conversation moved. The durable story is no longer "there aren't enough engineers." It is "there aren't enough of the right engineers at the right price." BLS occupational projections still show software development among the fastest-growing US occupations through 2033, but the hiring pyramid has thinned at the top, where research-level ML, infrastructure at scale, and security-cleared work all trade at premia the median US startup cannot fund from one country.

Layoffs concentrated in non-engineering work. The 2025 cut pattern is not a universal engineering retrenchment. Most reductions landed in recruiting, marketing operations, middle management, and discontinued products, with Computerworld's 2026 timeline and the Crunchbase tracker showing the same shape. Where engineering cuts landed, they were quickly offset by AI-focused rehiring, with Microsoft and Meta's restructurings explicitly redirecting headcount into AI infrastructure. Net shape: leaner in headquarters, more focused in hiring, and more willing to hire wherever the candidate lives.

AI forced the compensation conversation. Once frontier labs began paying in the $500K-$1M+ band for research and applied ML engineers, every AI company below them had to find leverage somewhere, and the cleanest lever is geography. A strong applied ML engineer in Paris, London, or Tel Aviv is not "cheaper" than one in San Francisco in any meaningful quality sense, but the offer structure and cost of living produce a compensation line that a seed-stage board will approve. EOR is what makes that offer executable without opening four entities.

The regulatory picture hardened. The US Department of Labor's 2024 contractor rule restored the multi-factor "economic realities" test and is now actively enforced. The EU's Digital Operational Resilience Act (DORA) went live in January 2025 for regulated ICT roles in financial services. The US BIS AI export-control rule, issued in early 2025, reshaped how AI companies can staff frontier-model work. The 1099-as-a-workaround era is over in the parts of tech that now matter most.

What an Employer of Record actually is (and isn't)

An Employer of Record is a third party that legally employs a worker on behalf of the company that directs the worker's day-to-day output. The EOR handles local employment contracts, payroll, tax withholding, statutory benefits, termination compliance, and in most countries the work-permit infrastructure.

It is not the same as:

• a staffing agency, which usually sources and places candidates
• a PEO, which co-employs within a country where the client already has an entity
• an AOR / agent of record, a different role used in benefits broking and contractor management
• "EOR software", which is the platform layer some providers sell separately from employment services

For a tech company specifically, two limits are worth knowing up front. EORs typically cannot hold US-state-regulated licenses on the client's behalf, and in some countries certain regulated roles, including financial-services controlled functions, must sit on the client's own licensed entity. The companion pieces below get into the specifics where they matter.

Why tech companies specifically are turning to EOR

Five pressures show up repeatedly across the conversations we have with tech buyers.

Speed. Opening a legal entity in a new country typically takes three to nine months once banking and payroll provisioning are factored in. A quarterly hiring plan, a new-market launch, or a runway-sensitive ramp does not wait. EOR compresses time-to-first-hire from months to days.

Reach into the right talent. A great senior engineer in Warsaw, a strong applied-ML scientist in Paris, a design lead in Lisbon, a DevRel hire who lives in Tokyo. Hiring "where the talent lives" is a recurring theme, and EOR is the mechanism. Major outsourcing analysts and academic labor-mobility research both show steady internationalization of high-skill tech employment over the last decade.

Misclassification risk. Using 1099 contractors for work that looks employment-like is now actively enforced under the DOL economic-realities test. California's AB-5 and the ABC test sit on top of that domestically. Outside the US, the picture is similar in the UK (IR35), Germany (Scheinselbstständigkeit), and Spain (the Riders Law). For the kind of long-running, exclusive, supervised work tech companies often need from a "contractor," the cleanest answer is to make them a properly contracted employee in their home country.

Capital efficiency. Pre-funding a multi-country payroll, the model at many legacy EORs, sits awkwardly against tight runways. An EOR that invoices in arrears, rather than holding salary deposits, keeps working capital on the balance sheet. Finance leaders at Series A-C tech companies flag this specifically when comparing providers.

Equity that actually works for global hires. ISOs are not available to non-US employees, period, and ISOs through an EOR are a non-starter even in principle. UK EMI options are not available to EOR-employed workers because EMI requires a direct employer-employee relationship with a qualifying UK company. Israeli Section 102 trustee options require the Israeli entity to be the grantor. The honest answer is NSOs everywhere, with per-country optimization where a real entity exists. Saying that on the offer letter is not a defect; it is a cap-table choice. Our stock options for EOR employees guide goes deeper.

Where EOR fits, and where it doesn't

Being honest about fit is the most useful thing a tech-focused EOR article can do.

Good fits. Most engineering, design, product, data, marketing, and commercial roles employed in a country where the work does not require a regulated license:

• software, platform, and data engineers
• applied ML and ML infrastructure engineers
• DevOps, SRE, and security engineers (non-cleared)
• product designers, design researchers, and engineering managers
• DevRel, technical writers, and developer-experience hires
• product managers, growth, and analytics
• marketing, content, and brand
• account executives, sales engineers, customer success, partnerships

Edge cases. Roles where a license, regulator approval, or specific local registration is needed but not always blocking. Examples include compliance and risk roles in fintech that may be eligible only at the non-controlled-function tier, security roles requiring background-check depth beyond standard EOR onboarding, and roles where the worker will hold significant decision authority that could trigger permanent establishment for the parent company.

Poor fits. Roles that require the worker to sit on a licensed entity's payroll for legal or regulatory reasons. The clearest examples in tech are UK SMCR Senior Manager Functions, Irish PCFs, and roles whose job description amounts to "running the licensed entity." Some defense-tech, government-cleared, and export-controlled work also has direct-employment requirements that an EOR cannot satisfy.

Any EOR promising to make these poor-fit situations disappear is worth a second look.

Roles tech companies typically hire through EOR

A non-exhaustive map of where international tech hiring is most concentrated. Each row names the geographies where Borderless and other EORs see the most volume.

Two notes from the table. First, the same role profile lives in different countries depending on stage and purpose: a Series-B SaaS opens UK / NL / DE for GTM, while a seed-stage AI company hires research engineers across France, the UK, and Israel. Second, many of the highest-volume roles, such as DevRel and ML research, sit at the edge of standard EOR onboarding and benefit from a provider with explicit experience in those patterns.

Countries tech companies hire most often via EOR

Twelve jurisdictions come up most. Treat as a starting map, not a recommendation.

• Poland. EU's deepest engineering bench at mid-tier cost. Strong on backend, data, and platform.
• Portugal. Lower-cost EU hire with English fluency. Common for SaaS engineering, design, and DevRel.
• Spain. Larger talent pool than Portugal; useful for design, data, and applied ML.
• United Kingdom. Post-Brexit hiring without an entity is now a routine EOR use case. Dense GTM, product, and senior-engineering bench.
• Netherlands. Common second EMEA country after the UK; stronger product and design pool than people expect.
• Germany. Higher cost, deeper specialist bench. Works-council and termination realities matter; see the SaaS EMEA hire guide.
• France. Critical for AI and design hiring. Inventor-remuneration and non-compete-indemnity rules are unique line items; see the AI playbook.
• Israel. Critical for senior ML and security. Section 102 trustee options and BIS export-control posture both apply.
• India. Largest by volume globally for engineering and operations. Strong English-language depth and well-developed EOR infrastructure.
• Brazil. LATAM's deepest engineering market. Tax and stock-options treatment have specific quirks.
• Mexico / Argentina / Colombia. Nearshore for US tech; timezone overlap with US is the main draw.
• Singapore / Australia / Japan. APAC anchors; common for first regional GTM hires.

For each, the practical EOR question is the same: does the provider hold an entity in the country, and have they hired this exact role profile before?

What "turning to EOR" actually looks like, by sub-vertical

Tech is not one industry. Four patterns have emerged, and each is its own operating conversation. Each has its own deep-dive companion piece.

US SaaS opening a first EMEA hub. A Series-B to Series-D US SaaS, coming off strong ARR, needs commercial hires in London, Amsterdam, Munich. This is the most common EOR use case in tech today. The concerns are practical: total landed cost, UK / NL / DE notice and severance mechanics, the EMI-eligibility gap for EOR-employed AEs, and the exit from EOR once the country has enough weight to justify an entity. Companion piece: How US SaaS companies build their first EMEA team with EOR.

AI and ML companies hiring research talent. A different conversation. The binding constraints are export-control posture for frontier-model work, IP-assignment clarity for model weights and fine-tunes, and equity mechanisms that actually work, because at senior ML levels equity is half the offer. Inventor-remuneration law in France, Germany, and Poland adds a TCO line that EOR calculators do not show. Companion piece: The AI company EOR playbook: export controls, model-IP, and equity.

Fintech pursuing licenses while hiring regulators. A fintech in the authorization window cannot put an SMCR Senior Manager Function holder, a Money Laundering Reporting Officer, or an Irish PCF on an EOR. Those roles have to sit on the licensed entity's paper. But the engineers, designers, risk analysts, and ops leads who do not hold a controlled function very often can. The EOR decision in fintech is a regulated-role matrix, not a yes/no. Companion piece: EOR for fintechs: regulated roles that can and cannot sit on EOR paper.

Open source and developer-tools companies converting contributors. The newest pattern and the messiest. A maintainer who has been sending PRs for two years is already effectively part of the company; formalizing that is not a standard hire. It requires a Prior Inventions schedule that actually enumerates pre-employment commits, a CLA-to-employment handoff that does not tangle the IP lineage, and honest conversations about 409A, NSO-only equity, and OFAC diligence in some countries. Companion piece: Contributor to employee: the EOR playbook for open-source companies.

Cybersecurity, gaming and live-ops, and Web3 / crypto have their own variants. Those pieces follow.

What to look for in an EOR partner for tech companies

Five criteria separate EORs that work for tech buyers from those that do not.

1. Owned entities in the countries that matter to your hiring plan, not "coverage through partners." For tech today, the practical minimum is usually the UK, Ireland, Netherlands, Germany, France, Poland, Spain, Portugal, Brazil, Mexico, Argentina, Colombia, India, Singapore, Japan, and Australia. Verify directly.
2. Owned entities versus third-party aggregation. Owned entities shorten the chain between the client and the worker's employment. Aggregator models introduce a second legal employer and can create lag on onboarding changes, terminations, and compliance questions. The market has had several public reminders of why this matters; treat aggregator-only providers as an operational risk, not just a margin choice.
3. Payroll funding model. Some EORs require the client to pre-fund the month's payroll in advance. Others invoice in arrears. For capital-constrained tech finance teams, the difference is material across a 12-month plan.
4. Equity mechanics. Ask explicitly how the provider handles stock-option grants, 409A considerations for US-based grantors, the NSO-only reality for EOR-employed grantees, and whether the provider has a workable phantom-stock or growth-shares pattern for jurisdictions where qualified equity is unavailable. Vague answers here mean vague execution later.
5. EOR-to-entity migration, written down. A provider should be able to show you the playbook for handing employees off to your own entity when you decide to open one. If that document does not exist, the end of the relationship will not be clean.

Borderless AI operates on an owned-entity model across 170+ countries, invoices payroll without requiring pre-funded salary deposits, and runs in-house support from North America. It is one of several providers a tech buyer should evaluate; the criteria above matter more than any single brand.

The first 90 days with an EOR

What actually changes operationally when a tech company starts using an EOR.

• Days 1-7. Provider verification (entities, certifications, SOC 2, data-handling). Country and role list finalized. Master service agreement signed. Onboarding portal access. HRIS / IT integrations scoped (Rippling, Workday, BambooHR; Okta / Google Workspace; GitHub / GitLab provisioning).
• Days 7-21. First country live. Local employment contract issued in the worker's language and jurisdiction. Background checks and right-to-work verifications completed. Benefits enrollment (statutory + supplemental). Stock-option grant documents issued and 409A treatment confirmed.
• Days 21-45. Payroll cycle one runs. HRIS sync confirmed. Expense, time-off, and equipment processes verified. Country-specific items addressed: UK PAYE reporting, Dutch 30% ruling where applicable, German social-security registration, Polish ZUS registration.
• Days 45-90. Second and third countries onboarded using the same template. Quarterly compliance review scheduled. Legal review of any country-specific anomalies (works councils in Germany, France, and the Netherlands; non-compete indemnity in France; inventor-remuneration in France, Germany, and Poland). EOR-to-entity migration plan drafted for any country headed past the threshold.

Buyers who treat the first 90 days as a structured rollout rather than a one-off hire get materially better results from any EOR.

Cost and timeline: entity vs EOR vs contractor

Numbers below are typical ranges, not quotes. Confirm with finance and legal for any specific country.

A useful heuristic for tech companies: EOR is usually cheaper below 8-15 employees per country; entity ownership is usually cheaper above that. The mid-zone is a finance call that depends on country (Germany, France, and Brazil push the threshold lower because of statutory employer load and complexity).

A practical decision frame

A simple way to decide.

The sequencing matters. Many tech companies use EOR as the first foothold in a country, then graduate specific hires into a local entity once the team reaches a size where owning the entity is cheaper than renting the employment relationship.

FAQs

What is an EOR for tech companies?

An Employer of Record is a third party that legally employs a worker in a country on behalf of a tech company that does not have a local entity. The EOR handles local employment contracts, payroll, taxes, statutory benefits, and compliance, while the client directs the work. Tech companies use EOR for engineering, design, product, GTM, and ML roles where opening a local entity is not yet justified.

Can I grant stock options to EOR employees?

Yes, but with limits. Non-qualified stock options (NSOs) are typically grantable to EOR-employed workers, with the 409Aconsiderations that apply to any foreign-optionee grant. ISOs are not available to non-US employees and are not workable through an EOR even for US-domiciled workers. UK EMI is not available because it requires a direct employment relationship with a qualifying UK company. Israeli §102 trustee options require the Israeli entity to be the grantor. Our EOR equity guidewalks through workable patterns.

Is EOR cheaper than opening an entity for a tech startup?

For the first several hires in a country, almost always yes. Entity setup runs from roughly $15K to $50K+ plus recurring legal, accounting, and payroll-provider costs, on top of the three-to-nine-month timeline. EOR pricing is typically a monthly fee per employee. The crossover point is usually 8-15 employees per country, faster in Germany, France, and Brazil because statutory employer load runs higher.

Can an EOR employ a senior software engineer in the EU?

Yes. EU senior-engineering hiring is one of the most common EOR use cases in tech, with the UK, Germany, the Netherlands, France, and Poland accounting for most of the volume. The provider's owned-entity coverage in the country, equity-grant handling, and termination expertise are the things to verify.

What countries do tech companies hire most via EOR?

Poland, Portugal, Spain, the UK, Germany, the Netherlands, France, Israel, India, Brazil, Mexico, Argentina, Colombia, Singapore, Japan, and Australia cover the large majority of volume. The exact mix depends on whether the hire is engineering (more central Europe and LATAM), GTM (more UK / NL / DE / Singapore), or ML research (more UK / France / Israel).

How is hiring through an EOR different for AI companies vs SaaS companies?

The roles overlap, but the binding constraints differ. SaaS expansion is mostly about commercial hires, EMI eligibility, and EOR-to-entity migration. AI hiring adds export-control posture, model-IP assignment, mandatory inventor-remuneration in some EU countries, and equity at the senior ML level where stock is half the offer. The SaaS first EMEA hire guide and the AI company EOR playbook treat each in detail.

Where this leaves tech

Tech's turn toward EOR is not "the future of work." It is an operating answer to a current math problem: the talent you want rarely lives where your entity is, and every quarter you wait to hire them, the AI-funded competition moves another notch up the offer scale. EOR is the mechanism that keeps the math workable in the period between "we need the hire" and "we are ready to own an entity here."

If you recognize your company in one of the four patterns above, the companion piece for your sub-vertical is the better next read. If you do not, the startup-focused guide is the general entry point. Either way, the conversation worth having is specific to your sub-vertical. Borderless AI's tech team is set up for that one.