Digital Nomad Setup: Connectivity, Taxes, Hard Truths
By: Evgeny Padezhnov
The romantic version sells beaches and laptops. The working version is about uptime, redundancy, and tax residency. For a developer, going nomadic is an infrastructure problem before it is a lifestyle choice.
Most failures are predictable. A dropped video call during an incident. A missed tax deadline in two countries. A single point of failure that turns a great month into a lost client. The good news: every one of these is solvable with boring, testable systems. Below is what holds up under real travel, and what quietly breaks.
Connectivity Is the Job
Internet is not a convenience for a remote developer. It is the production environment. Treat it like one.
Key point: plan for two independent connections at all times. One primary, one failover. If both run on the same cell tower or the same ISP, they are not independent.
A practical stack looks like this:
- Primary: local fiber or the apartment's wired connection.
- Failover 1: mobile data over an eSIM.
- Failover 2: a second carrier eSIM on a different network.
eSIMs removed the old ritual of hunting for SIM cards at the airport. Services like Airalo sell regional and country data plans that activate by QR code in minutes. Buy the plan before landing. Test it on arrival, not during a call.
For automatic failover, a travel router matters. Devices in the GL.iNet line accept a SIM, bond Wi-Fi and cellular, and switch when the primary drops. Configure failover and verify it by pulling the main cable mid-ping. If the connection survives, the setup is real.
Tested in production: in regions with weak terrestrial internet, Starlink changed the equation. A flat or van with a clear sky view gets usable, low-latency satellite uplink. It is bulky and not cheap, but for long stays off the grid it removes the single biggest risk.
Common mistake: trusting café and hotel Wi-Fi for anything that matters. It is shared, throttled, and often hostile. Use it for reading, never for a deploy or a client call.
Measure before you commit
Numbers beat vibes. Before booking a place for a month, ask the host for a speed test screenshot. On arrival, run your own:
# Throughput and latency
speedtest-cli
ping -c 20 1.1.1.1
# Jitter and packet loss over time (better signal for calls)
mtr --report --report-cycles 100 1.1.1.1
For calls and SSH sessions, jitter and packet loss hurt more than raw bandwidth. A 50 Mbps line with 8% packet loss is worse than a stable 15 Mbps line. mtr exposes that fast.
Hardware: Redundancy Beats Power
A nomad laptop should be light, durable, and replaceable in its workflow — not irreplaceable in its data. The machine will get stolen, dropped, or seized at a border eventually. The setup survives only if losing the hardware costs a day, not a month.
Principles that hold up:
- Everything in version control or cloud sync. No file lives only on the laptop. Dotfiles in a git repo. Code pushed daily.
- Reproducible environment. A setup script or a Dev Container rebuilds the machine in an hour. Document it once; thank yourself later.
- Encrypted disk, always. FileVault on macOS, LUKS on Linux. A stolen unencrypted laptop is a client-data breach.
- A spare. A cheap second device — even a tablet with a keyboard — keeps work moving while a replacement ships. Shipping to a foreign address takes a week or more.
In plain terms: the question is not "how powerful is the laptop." The question is "how fast can work continue after the laptop is gone."
Power is the other quiet failure. Outlets differ by country. A universal adapter and a small power bank are non-negotiable. For a workspace, a 100W USB-C charger drives a laptop, phone, and router from one socket.
The Network Layer: Access and Security
Public networks are untrusted by definition. A developer carries credentials, SSH keys, and production access. That access cannot leak over hotel Wi-Fi.
A mesh VPN solves most of this cleanly. Tailscale builds an encrypted WireGuard network between devices. The laptop reaches the home server, the office machine, and any cloud box as if they were on one LAN — no port forwarding, no exposed services.
Concrete setup:
# On every device
tailscale up
# Route all traffic through a home exit node when on hostile Wi-Fi
tailscale up --exit-node=<home-node-ip>
The exit node trick matters. On sketchy networks, route traffic through a trusted node. The local network sees only encrypted WireGuard packets.
For services that must stay private — a database admin panel, an internal dashboard — keep them off the public internet entirely. Bind them to the Tailscale interface. They become unreachable to everyone outside the mesh. That is stronger than any password.
Common mistake: exposing SSH on port 22 to the whole internet "just while traveling." Bots find it in minutes. Use key-only auth, a VPN, and a service like Tailscale SSH or fail2ban at minimum.
Password and secret hygiene
A password manager with cross-device sync is mandatory: Bitwarden, 1Password, or self-hosted Vaultwarden. Two-factor backup codes go in encrypted storage, not on the phone that might get stolen with the laptop. Losing the only 2FA device in a country with no roaming is a real way to get locked out of your own bank.
Money Moves Differently Across Borders
Multi-currency income and spending break traditional banks. A nomad earns in one currency, pays rent in another, and books flights in a third. Conversion fees eat into that on every transaction.
Tools that work in practice:
- Wise for multi-currency accounts and real exchange-rate transfers. Local account details in several currencies. Receive USD, hold EUR, spend in local currency at near-mid-market rates.
- Revolut as a second card and backup. Different network, different failure modes.
- A home-country card kept in reserve. Some services reject foreign cards; some countries distrust fintech cards.
Carry at least two cards from two providers. A frozen card in a foreign country with no backup is a genuine emergency. Banks freeze cards on unexpected geography constantly. Notify them of travel where possible, and never rely on a single card.
Key point: keep a cash buffer in a hard currency. ATMs go down. Cards get declined. A few hundred dollars in cash has resolved more travel crises than any app.
Taxes and Visas: The Part That Actually Bites
The technical setup is the easy part. Tax residency and legal status are where nomads get hurt, often a year later when it is expensive to fix.
The core concept is the 183-day rule. Spend more than 183 days in a tax year in most countries, and you likely become a tax resident there — owing tax on income. The exact rules vary, and several countries use additional ties (a home, a family, a center of economic interest) to claim residency even under 183 days. Staying "always moving" does not automatically make tax liability disappear; some home countries tax based on citizenship or until residency is formally severed.
This is not optional reading. A developer earning a Western salary while living abroad can face liability in the home country, the host country, or both. Double-taxation treaties exist to prevent paying twice, but they require correct filing, not hope.
Dedicated digital nomad visas now exist in many countries and make staying legal straightforward:
- Portugal offers a remote-work visa (often called the D8) with a recurring-income requirement.
- Spain has a digital nomad visa tied to remote employment or freelance income.
- Estonia pioneered a formal digital nomad visa for location-independent workers.
Each sets a minimum monthly income, requires proof of remote work, and demands health insurance. The income thresholds are real gatekeepers — check the current official figure for each country before planning, because they change. Do not rely on a blog's number, including this one.
Key point: consult a cross-border tax professional before the first long stay, not after. The fee is trivial against a misfiled year. This article is not tax advice; it is a warning to get real advice.
Health insurance deserves the same seriousness. Travel insurance is not health insurance. Services aimed at nomads — SafetyWing and similar — cover long-term travel and medical needs across borders. Read the exclusions. Pre-existing conditions and "home country" coverage are common gaps.
Workflow: Time Zones and Async Discipline
A nomad ten time zones from the team cannot work synchronously. The fix is async-first habits, not heroic 3 a.m. calls.
What works:
- Write things down. Decisions, context, and status go in writing — pull requests, issue threads, a shared doc. The team reads on their schedule.
- Overlap windows. Agree on two or three hours of overlap for live discussion. Protect them. Schedule deep work outside them.
- Visible status. A shared calendar with the real working hours in the team's timezone removes guesswork.
Time-zone math is a frequent source of missed meetings. Tools like the World Clock in any calendar app, or tz-style CLI utilities, prevent the classic off-by-one error. Common mistake: scheduling across a daylight-saving change. Two countries shift on different dates. Always confirm with an explicit UTC time.
In practice, async work makes a remote developer more valuable, not less. Clear written communication is a skill the whole team benefits from. The constraint forces good habits.
The Honest Downsides
The lifestyle has real costs that the marketing skips.
- Isolation is constant. New cities mean no built-in social circle. It takes deliberate effort to avoid loneliness.
- Logistics never stop. Visas, apartments, SIM cards, and bank issues are a recurring tax on attention. That overhead competes with actual work.
- Burnout hides in plain sight. When the office is everywhere, work is everywhere. Boundaries must be built manually.
- Deep work suffers under churn. Changing cities every two weeks kills focus. Slower travel — a month or more per place — produces better work and lower stress.
If something does not work, the honest answer is to say so: constant relocation and demanding engineering work pull against each other. Many successful nomads slow down to one base per quarter. The label matters less than a setup that lets the work continue reliably.
What to Try Right Now
Pick one place you might live and run the test before committing: book a short stay, then on arrival run mtr --report 1.1.1.1 and a speedtest-cli at the hours you actually work. If latency and packet loss hold up for a real video call and an SSH session, the location passes. If it does not, you found out in three days instead of after a one-year lease.
Try it: the entire nomad setup reduces to one principle — redundancy in connectivity, hardware, money, and legal status. Build each layer so a single failure costs a day, not a month. If it works — it is correct.
Information is accurate as of the publication date. Terms, prices, and regulations may change — verify with relevant professionals.