PhD funding is where scholarship matching stops being a marketing claim and becomes testable. A stipend either covers rent in the city where the lab sits, or it does not. So we took the doctoral slice of our dataset and asked a blunt question: what could an AI scholarship finder actually match you on, given the data that programs publish?
The doctoral slice: 121 scholarships, half of them half-described
Our dataset holds 494 scholarships, and 121 of them are open at doctoral level. Here is what those 121 entries publish. Sixty-four state an award amount. Fifty publish a deadline a machine can read. Forty-eight state a GPA requirement. Thirty-seven are tagged fully funded. In other words, for the typical PhD scholarship, a finder has an amount to compare in only about half the cases, and a deadline to track in even fewer.
Among the 64 doctoral entries that do publish an amount, the median award is $10,000. Treat that number carefully: it mixes one-off grants with annual stipends, which is exactly the kind of quiet data problem a matcher has to solve before its ranking means anything. A $10,000 one-off award and a $10,000 per year stipend look identical in a spreadsheet column and completely different in a bank account.
What a real stipend looks like
The benchmarks worth memorising have not changed. DAAD doctoral funding in Germany pays around 1,300 euro a month plus insurance. The Australian RTP pays around AUD 32,000 a year, tax-free. Rhodes at Oxford covers full cost plus a stipend and requires institutional endorsement. And in the US, most top-50 PhD programs fund you through teaching or research assistant lines, roughly $25,000 to $40,000 a year: the scholarship is the admission itself. Any finder result that cannot be sanity checked against numbers like these is decoration.
The doctoral pool skews generous, and paper-heavy
Two more counts from the same 121 entries sharpen the picture. Thirty-seven of the doctoral scholarships are tagged fully funded, about 31 percent, against 19 percent fully funded across the whole dataset. Doctoral money, when it exists, is more likely to be complete money, because funders are buying four years of research labor, not subsidising a degree. That is the good news. The paperwork news is that 45 of the 121 doctoral entries specify recommendation letters as part of the application, and 48 state a GPA floor, so a matcher that only filters on country and field will hand you programs whose real gates it never showed you.
This is also why stipend hunting rewards patience over breadth. A masters applicant can spray twenty applications. A doctoral applicant is really courting a handful of labs, and the finder's job is to make sure the handful is eligible, funded and open this cycle, not to inflate the count.
Where the matcher helps, and where it cannot
Matching works on the fields that exist: country, degree level, field, nationality, and the 64 published amounts. It fails on the fields that decide real PhD funding. Supervisor fit is not in any dataset. Lab funding cycles are not in any dataset. Whether a professor is taking students this year lives in email threads, not databases. That is why the honest workflow is two-stage: let the finder build the eligible list, then do the human work of emailing professors before you apply, because for PhDs the stipend follows the supervisor, not the brochure. A funded lab with an average stipend beats an unfunded dream school every time. We stress-tested the same limits across the broader dataset in our guide to AI scholarship matcher apps and where they fall short.
Put the finder to work, then do the human part
ScholarshipMatch matches your profile against all 494 scholarships in this dataset, doctoral entries included, and hands you the eligible list with deadlines attached for $9.99 one-time per search. Use it to build the list. Then spend your real hours on supervisor emails and the merit funding landscape if a masters comes first. The finder narrows the field; it cannot walk into the lab for you.
One last habit separates funded doctoral students from the rest: they treat the stipend as a salary negotiation with reality, checking it against rent, insurance and visa income thresholds in the actual city, not against the number's size in isolation. A modest stipend in a cheap university town can leave you richer than a headline figure in a capital city.
Sources: EducationUSA, DAAD