Run an AI PhD stipend by country comparison the way most people do, and you get currency-converted numbers lined up as if a bigger figure means a better deal. It doesn't. A stipend only means something once you weigh it against what a room and a health plan cost in that city, and once you know whether it's a salary with benefits or a bare grant with none. Get that part right and the rankings often flip.
AI PhD stipend by country comparison: why the ratio beats the raw number
Two stipends that look identical on paper can fund very different lives. A modest-sounding figure in Montreal might cover rent and groceries with room to spare, while the same figure in Zurich or San Francisco barely covers rent alone. What predicts your day-to-day life is the ratio between the stipend and local cost of living, not the stipend on its own. Price offers against a rough monthly budget for the actual city, not the national average, since a stipend in Toronto and one in a smaller university town two hours away can buy noticeably different lives. Check what's bundled in too: a tuition waiver, health insurance, and a travel budget change an offer's real value more than a few hundred base dollars.
Europe treats a PhD like a job, not a favor
In much of Europe, a PhD candidate in AI or machine learning is a salaried employee, not a scholarship recipient. Germany, the Netherlands, the Nordic countries, and Switzerland typically put PhD students on a payroll with a real contract: pension contributions, paid holiday, sick leave, and national health coverage. Pay usually tracks a public sector scale rather than an individual advisor's grant, so students doing similar work land in a similar range. The EU-wide Marie Sklodowska-Curie Actions program is a good example of this: it publishes a fixed living and mobility allowance rather than leaving the number to an individual advisor's grant. Switzerland pays the highest nominal numbers here, and Zurich or Geneva rent eats into that advantage fast. Germany and the Netherlands often land in a more comfortable middle: lower pay than Switzerland, but lower rent too, plus a real safety net. That employee status also shapes your visa. Germany spells this out directly on its official researcher visa page: contracts tend to renew and extend through a documented process, while a scholarship position often doesn't carry the same protections, so ask directly which one you're being offered.
The US: it depends enormously on where you land
US PhD funding doesn't run on a national scale the way German or Dutch funding does. A stipend is set by the department and the advisor's grant money, so the same field at two universities twenty miles apart can differ by a third. Layer cost of living on top and the gap widens: a stipend that's comfortable in Pittsburgh or Madison can be tight in Boston, the Bay Area, or New York. Health insurance is sometimes fully covered, sometimes a payroll deduction you won't notice until your first pay stub. None of this means the US pays badly, since well-funded labs at strong departments are genuinely competitive. It means the country-level number is close to useless, and the department-and-city number is what's worth asking for.
Where the funding runs thin
Not every system funds PhD students generously. Some programs expect candidates to secure their own scholarship before enrolling, then treat the university's role as supervision rather than payment. Others offer a grant that hasn't been adjusted for inflation in years, so its real value quietly erodes while the number on the funding page stays the same. A lower cost of living helps, but it doesn't rescue a stipend that doesn't cover rent where the lab is based. Germany's DAAD is a useful cross-check here, since it publishes funded program listings you can compare against what a university's own page claims. If a funding page is vague about the amount or whether it renews, that vagueness is itself useful information.
Strong labs are worth chasing regardless of the country line item
Some of the best reasons to pick a place have less to do with funding scale and more to do with who you'd work alongside. Mila in Montreal, the Max Planck Institutes in Germany, the pan-European ELLIS network, INRIA in France, and ETH Zurich all have long track records in AI research and alumni networks that open doors later. A slightly lower stipend with an advisor who's genuinely active in the field tends to beat a higher one somewhere quieter. Weigh the lab and advisor at least as heavily as the funding line.
Visa and work authorization notes for research movers
Most countries route PhD students through a dedicated student or researcher visa, separate from the general skilled-worker categories, and some build in a stay-back window after you defend so you can look for postdoc or industry work. What varies a lot is what a spouse or partner can do while you study: some grant a dependent visa with full work rights from day one, others restrict it heavily, and that difference can matter more to a couple's finances than the stipend itself. If your partner needs to work, ask about dependent work rights before anything else on the visa side.
How to evaluate an actual offer
- Ask whether the figure is net or gross, and what gets deducted before it hits your account
- Ask if health insurance, tuition, and a conference or travel budget are included or extra
- Check rent for the specific neighborhood near the lab, not the national or city-wide average
- Ask how long the funding is guaranteed and what happens if it doesn't renew on schedule
- Ask about your spouse's or partner's work rights if that applies to your situation
- Ask two current students in the lab what they actually spend and save each month
The offer letter tells you what they will pay. Only a current student in that lab can tell you what that pay actually buys.
There's no single country that wins on stipend for every AI PhD applicant, because it depends on the city, the lab, your family situation, and what you value getting in return for taxes and contributions. Stipend rules, tax treatment, and visa terms shift by program and by year, so verify the actual figure and its conditions directly with the department before you commit to anything. This is informational, not legal, tax, or immigration advice.
If you haven't narrowed your shortlist yet, the best countries to study AI abroad in 2026 is a good place to start before getting this granular. For the fuller picture, the AI Relocation Guide includes a PhD and academic labs section with stipend-to-cost-of-living notes across all 21 countries, so you can compare all 21 countries instead of piecing it together from department websites. Once you've picked two or three finalists, the Scholarship and Funding Guide goes deeper on the fellowships that can top up a thin stipend.
