Generic engineering scholarship lists mostly miss how AI and data science get funded

If you search for scholarships in computer science, data science or AI, you mostly get lists built for engineering in general, heavy on civil and mechanical programmes, with a line or two about “STEM” tacked on. That is not where the real money for this field sits. A large share of funding for graduate work in AI and machine learning specifically comes from technology companies running their own fellowship programmes, not from government scholarship bodies, and those programmes have their own separate application calendars and rules.

This guide covers the fellowship route specifically, plus the handful of funding mechanisms in this space that are structured differently from a normal scholarship.

The company-run PhD fellowships: Google, Meta, NVIDIA

Google runs a PhD Fellowship Program open to full-time PhD students working in areas like machine learning, natural language processing, security and privacy, and human-computer interaction, among the research areas it lists. Coverage varies by region: in Canada and the US it is worth up to $85,000 a year toward tuition and fees, living expenses, travel and equipment, for up to two years; in Europe and the Middle East it is a yearly bursary covering stipend, health care, tuition and conference travel, also for up to two years; in India, early-stage PhD students can get up to four years at $50,000 a year; in Africa the award is $15,000 a year for research costs and travel, for up to three years.

Universities nominate students rather than students applying directly, and you have to stay enrolled full-time for the life of the award or forfeit it. The 2026 cycle already closed, with the window running from March 5 to April 30, so budget for an annual cycle that opens in spring.

Meta runs a similar programme, the Meta Research PhD Fellowship, worth two years of paid tuition and fees plus a $42,000 annual stipend for living and conference costs, and it flies winners to a Fellowship Summit at Meta’s headquarters. It is open to PhD students in any year of study, not just incoming ones, but only in a defined list of research areas that includes AI systems, AR/VR, databases, networking, privacy and computational social science. Funding a PhD works differently at the master’s level, so check that guide if you are still deciding which degree stage to fund.

NVIDIA’s Graduate Fellowship Program pays up to $60,000, with the exact amount based on the student’s actual costs of stipend, tuition and mandatory fees, for PhD students in computer science, computer engineering, system architecture or electrical engineering who have already completed their first year of study. It comes with a catch most scholarship lists leave out: an in-person summer internship at an NVIDIA office in the US, Canada, Germany, Sweden, Finland, Switzerland, Israel or Taiwan is mandatory, not optional, and it happens before the fellowship year begins. Submissions for the current cycle are closed.

These are not scholarships in the usual sense

None of the three programmes above route through a university’s financial aid office the way a named scholarship does. They are closer to research grants tied to a company’s own priorities, and funding routes that are not scholarships covers other examples of this pattern, like assistantships and industry-sponsored studentships, if you want the fuller picture.

Because these are company programmes, the research-area lists matter more than they do for a general scholarship. A strong application in, say, computational biology will not be considered for the AI systems categories even if the work overlaps.

Master’s-level funding in this field is thinner and less centralised

Below PhD level, there is no equivalent industry-run master’s fellowship at this kind of scale. Master’s students in AI and data science generally rely on the same routes as any other field: university-specific merit awards, national scholarships like Fulbright or Chevening, or country-specific programmes covered elsewhere on this site. One clear exception exists for African students specifically.

AMMI: a fully funded machine learning master’s built for African students

The African Master’s in Machine Intelligence, run by the African Institute for Mathematical Sciences with sponsorship from Google and Meta (still listed under its earlier name, Facebook, on the programme’s own page), is described on its own site as a fully funded one-year intensive graduate programme in machine learning aimed at young Africans. The most recent cohort drew students from 18 African countries. The programme’s own materials do not spell out exactly what “fully funded” includes item by item, tuition, housing, stipend, so confirm the specifics directly with AIMS before you count on any one component. If you are weighing this against other funded routes for students from the region, funding for students from developing countries is a useful companion read.

Europe’s less obvious route: Marie Skłodowska-Curie doctoral networks

If you are aiming for a PhD in Europe, the EU’s Marie Skłodowska-Curie Actions fund Doctoral Networks that recruit researchers of any nationality, not just EU citizens, into multi-institution projects. Funded researchers get a living allowance and a mobility allowance, on top of research and training costs, and the money is not restricted to computer science, many networks are specifically built around AI and machine learning topics each call cycle. The catch is a mobility rule: you generally cannot have lived or worked in the country of the recruiting institution for more than 12 months in the three years before you are hired, which is designed to push researchers to move somewhere new.

Fellowship duration within a project runs from 3 to 36 months (up to 48 for Joint Doctorates), with full projects lasting up to four years (up to five for Joint Doctorates). The exact living allowance figure is set by an EU unit-cost table that is adjusted by country and updated with each work programme, so check the current rate rather than relying on a fixed number. Note that you do not apply to the EU directly for a PhD position: the consortium of universities and organisations applies for Doctoral Network funding first (the current call closes 24 November 2026), and only once a project is funded do individual PhD vacancies get advertised, on the EURAXESS portal.

Practical order of operations

Work out which degree stage you are funding first, since almost none of this overlaps. PhD applicants in AI-adjacent fields should check Google, Meta and NVIDIA’s current research-area lists before writing a research statement aimed at any one of them, since a mismatch there is a common and avoidable rejection reason. If Europe is on the table, search the EURAXESS portal for open PhD vacancies under funded Doctoral Network projects in your specific subfield rather than assuming a generic EU scholarship exists for CS. If you are an African student aiming at a machine learning master’s specifically, AMMI is worth checking before you spend time on more generic STEM scholarship lists that were not built with this field in mind.